Collins 30S-1 HF Amplifier - User contributions [en]

How and Why I Changed It

2025-05-04T22:46:57Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-point down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>
 
[[File:30s1_blower_pressure_rpm.png|500px|thumb|center|Blower-Pressure versus rpm, along with control-voltage]] 
 

 

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T22:44:31Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>
 
[[File:30s1_blower_pressure_rpm.png|500px|thumb|center|Blower-Pressure versus rpm, along with control-voltage]] 
 

 

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T22:43:33Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>
 
[[File:30s1_blower_pressure_rpm.png|500px|thumb|left|Blower-Pressure versus rpm, along with control-voltage]] 
 

 
== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T22:43:20Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>
 
[[File:30s1_blower_pressure_rpm.png|500px|thumb|left|Blower-Pressure versus rpm, along with control-voltage]] 
 

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T22:42:56Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

[[File:30s1_blower_pressure_rpm.png|500px|thumb|left|Blower-Pressure versus rpm, along with control-voltage]]

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T22:41:56Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

[[File:30s1_blower_pressure_rpm.png

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T22:41:34Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

30s1_blower_pressure_rpm.png

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

File:30s1 blower pressure rpm.png

2025-05-04T22:40:52Z

Gordonp: Data and graph of 30S-1 amp blower, pressure versus rpm (along with control-voltage to produce these points)

== Summary ==
Data and graph of 30S-1 amp blower, pressure versus rpm (along with control-voltage to produce these points)

How and Why I Changed It

2025-05-04T22:36:31Z

Gordonp: /* Comment */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power (1kW input), let it stabilize, then tweak R106 until it just shuts down. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T21:11:05Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest switching range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed. This brings the switch-poing down from the max. 0.3" w.c. 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T21:10:16Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring (to get us into the lowest range, about 0.1" w.c. - 0.3" w.c.) 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T21:09:05Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is 1100W - 1200W continuous dissipation, when the blower is spinning 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T21:07:22Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" around 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" about 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is that this could dissipate 1100W - 1200W continuously, when operating at 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T21:06:22Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial blower-speed to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" at 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is that this could dissipate 1100W - 1200W continuously, when operating at 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T21:06:05Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial fan to 2500rpm / 0.3" w.c. 
<li>Pressure Switch turns "on" at 2000rpm / 0.212" w.c. 
<li>Pressure Switch turns "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is that this could dissipate 1100W - 1200W continuously, when operating at 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T20:05:08Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial fan to 2500rpm / 0.3" w.c. 
<li>Press Sw "on" at 2000rpm / 0.212" w.c. 
<li>Press Sw "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>
It's challenging to estimate dissipation, because Eimac spec-sheet used 50degC inlet air-temp, and clearly specifies sea-level (which I am). But a reasonable estimate is that this could dissipate 1100W - 1200W continuously, when operating at 2500rpm.

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T20:02:19Z

Gordonp: /* Operating Point */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
<ul><li>dial fan to 2500rpm / 0.3" w.c. 
<li>Press Sw "on" at 2000rpm / 0.212" w.c. 
<li>Press Sw "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN! 
</ul>

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T20:01:19Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
<ul><li>Use weakest spring 
<li>Turn out adjustment hex approx 3/4-turn from gently bottomed 
<li>Hysteresis of switch found to be about 0.075 water-column 
</ul>

== Operating Point ==
dial fan to 2500rpm / 0.3" w.c.
Press Sw "on" at 2000rpm / 0.212" w.c.
Press Sw "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN!

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T20:00:28Z

Gordonp: /* Pressure Switch Adjustments */

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
Use weakest spring 
Turn out adjustment hex approx 3/4-turn from gently bottomed 
Hysteresis of switch found to be about 0.075 water-column 

== Operating Point ==
dial fan to 2500rpm / 0.3" w.c.
Press Sw "on" at 2000rpm / 0.212" w.c.
Press Sw "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN!

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

How and Why I Changed It

2025-05-04T19:59:49Z

Gordonp:

The original Collins thermal sensor is kind of neat: It's a normally-closed thermal switch with a heater; the heater biases the thermal-switch up toward nearly-opening... at this point, a delicate dance ensues: heat-calories from the tube try to open the switch, while heat-calories are removed by the blower-airflow.

BUT - it does not need to be such a delicate dance - the tube anode-seals require <250degC, and should be operated <225degC; in my prudence I think <200C should be safer. Infrared/laser remote temperature measurements suggest the tube may only rise xxxdegC above ambient, providing a very large margin, and a large window between "operation" and "danger". My solution will be a thermal-switch, which will open the 12V and K203 when it senses 160degC - ample safety for the tube, but not being a nagging nanny to the operator :-)

This alone will ensure tube-safety, but let's go even farther - let's use both belts, and suspenders :-) Modelling my Dayton-motored blower suggests the blower will produce somewhere around 0.8 inches water-column (wc) with the 4CX1500B. The published requirements for the 4CX1000A pressure is 0.2"wc at a full 1kW dissipation; the 4CX1500B requires even less at 0.18"wc for 1kW dissipation! And normal operation will duty-cycle / time-average the dissipation down.

So, we'll sense the air-pressure at the base of the tube. Again, we have a wide window to allow full operation, and also maintain total tube safety. 0.1"wc should be "sufficient" for normal operation; 0.3"wc should allow for "no-time-limit" 1kW dissipation (maybe my RTTY!).

Thanks to induced-draft furnaces and hot-water-heaters, the HVAC industry has a wide selection of suitable temperature- and pressure-sensors. On the other hand, the Collins sensor is UnObtainium, and mine doesn't appear to be working correctly. We can now have deterministic, sustainable, reproducible and improved tube safety!

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:160degC_nc_thermal_switch.jpeg|Honeywell 2455RC-90820388 SPST NC 160degC Auto-Reset Thermo-Switch
File:adjustable_pressure_switch.jpeg|All-Temp NS2-0000-03 Universal Air-Pressure Switch. RobertShaw 2374-510 looks similar
</gallery>

We'll use these temperature and pressure switches to cut the High voltage and screen supply, disabling the main tube - same as Collins did.

== Pressure Switch Adjustments ==
Use weakest spring
Turn out adjustment hex approx 3/4-turn from gently bottomed
Hysteresis of switch found to be about 0.075 water-column

== Operating Point ==
dial fan to 2500rpm / 0.3" w.c.
Press Sw "on" at 2000rpm / 0.212" w.c.
Press Sw "off" at 1729 rpm / 0.15" w.c. <- THIS IS SLIGHTLY ABOVE STOCK FAN!

==Comment==
Anecdotally, the adjustment procedure for the K102 thermal switch is to run the amp at just above full-rated-power, let it stabilize, then tweak R106 until it just shuts down. But following this means somewhat over 1kW DC input to the tube, and perhaps somewhere around 750W dissipation. At this dissipation level, with the stock blower's airflow, the anode temperature rises by xxxdegC.... far below damaging levels. I conclude that this sensor is not factory-set to save the tube, but rather to prevent the operator from exceeding the designed power-level. Saving the tube is a side-effect. I have re-prioritized, and made tube-safety paramount.

[[File:Thermal_overload_switch_k102.jpeg|500px|thumb|left|removed, original Collins thermal-overload switch K102]]

Collins' thermal sensor is quite elegant: it senses the heat-calories coming in (from the anode) and heat-calories
going out (from the airflow). And it doesn't noticeably impact the airflow.

When it's closed, this switch allows the 15VDC Control current to flow, which will then permit K203 to close and power the High Voltage transformer (if all other safety switches are also closed).

When this switch opens, the 15VDC applied to K203 is removed. K203 opens. High voltage and screen voltage cease, and the tube stops operating. 

<gallery caption="Mounting a New Temperature Sensor" widths="400px" heights="400px" perrow="3">
File:Coffeecan_snips_pliers_bracket_making.jpeg|Costco coffee-can, tin-snips and pliers for making a new sensor-bracket
File:Sensor_bracket_prototype_evolution.jpeg|a series of thermal-sensor bracket-prototypes. These feature folded coffee-can metal, which is then soldered for additional thickness and rigidity.
File:thermal_sensor_mount_testfitting.jpeg|test fitting the final iteration of the thermal-sensor mount
</gallery>

Updates and Changes

2025-01-17T01:53:05Z

Gordonp: /* Dial Lights "ON" when Amp is Ready */

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

I sourced this "ready-to-go" supply from the panel-mounted S208 "OFF" pushbutton. The outermost terminal loses it's 15VDC momentarily when S208 is pushed, so I used the innermost terminal (next to the red "ON" light) which remains live always (well, after 3min warm-up).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

===LEDs===
So, when I mentioned above that the dial-bulbs are now operating from ~15VDC ...
I replaced the dial-bulbs with some ubiquitous "pinball" bulbs, meant to replace #44 and #47 incandescent bulbs. My particular LED bulbs have a diode-bridge in the base and are DC-polarity insensitive. However, they were designed to operate on 6.3VAC, not my ~8VDC :-O At first, they were great and very bright. After a while, they began to "stutter" - not really "flicker" and certainly in no relation to any supply-ripple, but a slower, cyclical blink-stutter, sometimes opposite to each other. Hmmm...

I dug in: checked all connections, cleaned the bulb-socket contacts, bypassed all safety and thermal-timer connections and otherwise made sure the 30S-1 itself didn't generate the flicker at any point in the 12VAC circuitry. Then, I got out the oscilloscope, to check things: the mid-voltage between the LEDs was erratic and crazy. From this, I conjectured the LED bulbs were failing internally, possibly due to overheating. Replacing both the high- and low-side bulbs brought immediate relief and proper operation, but I knew it would not last ...

So - from my always-HOT side of S208 (which supplies the dial-LED bulbs) I added a 5.1V zener diode, effectively knocking the voltage down something like 5.5V actual. This results in the pair of LEDs receiving ~10V total, or ~5V across each. This reduction from 6.3VAc to 5VDC resulted in a less-bright and more-yellowish light... which I find AWESOME!! And I hope they'll last a very long time.

NOTE: LEDs are "fast", and the DC I now use for the dial-bulbs isn't well-filtered. There is a subtle 120Hz flicker when I TX and the relays load this lightly-filtered DC supply. Again, the slight dimming and slight flicker in TX seems a beautiful symbolism for my 30S-1 "picking up and shouldering the load". Appropriate for the era!

Updates and Changes

2025-01-17T01:48:03Z

Gordonp: /* LEDs */

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

===LEDs===
So, when I mentioned above that the dial-bulbs are now operating from ~15VDC ...
I replaced the dial-bulbs with some ubiquitous "pinball" bulbs, meant to replace #44 and #47 incandescent bulbs. My particular LED bulbs have a diode-bridge in the base and are DC-polarity insensitive. However, they were designed to operate on 6.3VAC, not my ~8VDC :-O At first, they were great and very bright. After a while, they began to "stutter" - not really "flicker" and certainly in no relation to any supply-ripple, but a slower, cyclical blink-stutter, sometimes opposite to each other. Hmmm...

I dug in: checked all connections, cleaned the bulb-socket contacts, bypassed all safety and thermal-timer connections and otherwise made sure the 30S-1 itself didn't generate the flicker at any point in the 12VAC circuitry. Then, I got out the oscilloscope, to check things: the mid-voltage between the LEDs was erratic and crazy. From this, I conjectured the LED bulbs were failing internally, possibly due to overheating. Replacing both the high- and low-side bulbs brought immediate relief and proper operation, but I knew it would not last ...

So - from my always-HOT side of S208 (which supplies the dial-LED bulbs) I added a 5.1V zener diode, effectively knocking the voltage down something like 5.5V actual. This results in the pair of LEDs receiving ~10V total, or ~5V across each. This reduction from 6.3VAc to 5VDC resulted in a less-bright and more-yellowish light... which I find AWESOME!! And I hope they'll last a very long time.

NOTE: LEDs are "fast", and the DC I now use for the dial-bulbs isn't well-filtered. There is a subtle 120Hz flicker when I TX and the relays load this lightly-filtered DC supply. Again, the slight dimming and slight flicker in TX seems a beautiful symbolism for my 30S-1 "picking up and shouldering the load". Appropriate for the era!

Updates and Changes

2025-01-17T01:47:30Z

Gordonp: /* LEDs */

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

===LEDs===
So, when I mentioned above that the dial-bulbs are now operating from ~15VDC ...
I replaced the dial-bulbs with some ubiquitous "pinball" bulbs, meant to replace #44 and #47 incandescent bulbs. My particular LED bulbs have a diode-bridge in the base and are DC-polarity insensitive. However, they were designed to operate on 6.3VAC, not my ~8VDC :-O At first, they were great and very bright. After a while, they began to "stutter" - not really "flicker" and certainly in no relation to any supply-ripple, but a slower, cyclical blink-stutter, sometimes opposite to each other. Hmmm...

I dug in: checked all connections, cleaned the bulb-socket contacts, bypassed all safety and thermal-timer connections and otherwise made sure the 30S-1 itself didn't generate the flicker at any point in the 12VAC circuitry. Then, I got out the oscilloscope, to check things: the mid-voltage between the LEDs was erratic and crazy. From this, I conjectured the LED bulbs were failing internally, possibly due to overheating. Replacing both the high- and low-side bulbs brought immediate relief and proper operation, but I knew it would last ...

So - from my always-HOT side of S208 (which supplies the dial-LED bulbs) I added a 5.1V zener diode, effectively knocking the voltage down something like 5.5V actual. This results in the pair of LEDs receiving ~10V total, or ~5V across each. This reduction from 6.3VAc to 5VDC resulted in a less-bright and more-yellowish light... which I find AWESOME!! And I hope they'll last a very long time.

NOTE: LEDs are "fast", and the DC I now use for the dial-bulbs isn't well-filtered. There is a subtle 120Hz flicker when I TX and the relays load this lightly-filtered DC supply. Again, the slight dimming and slight flicker in TX seems a beautiful symbolism for my 30S-1 "picking up and shouldering the load". Appropriate for the era!

Updates and Changes

2025-01-17T01:46:24Z

Gordonp: /* LEDs */

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

===LEDs===
So, when I mentioned above that the dial-bulbs are now operating from ~15VDC ...
I replaced the dial-bulbs with some ubiquitous "pinball" bulbs, meant to replace #44 and #47 incandescent bulbs. My particular LED bulbs have a diode-bridge in the base and are DC-polarity insensitive. However, they were designed to operate on 6.3VAC, not my ~8VDC :-O At first, they were great and very bright. After a while, they began to "stutter" - not really "flicker" and certainly in no relation to the ripple, but a slower, cyclical, sometimes opposite to each other. Hmmm...

I dug in: checked all connections, cleaned the bulb-socket contacts, bypassed all safety and thermal-timer connections and otherwise made sure the 30S-1 itself didn't generate the flicker at any point in the 12VAC circuitry. Then, I got out the oscilloscope, to check things: the mid-voltage between the LEDs was erratic and crazy. From this, I conjectured the LED bulbs were failing internally, possibly due to overheating. Replacing both the high- and low-side bulbs brought immediate relief and proper operation, but I knew it would last ...

So - from my always-HOT side of S208 (which supplies the dial-LED bulbs) I added a 5.1V zener diode, effectively knocking the voltage down something like 5.5V actual. This results in the pair of LEDs receiving ~10V total, or ~5V across each. This reduction from 6.3VAc to 5VDC resulted in a less-bright and more-yellowish light... which I find AWESOME!! And I hope they'll last a very long time.

NOTE: LEDs are "fast", and the DC I now use for the dial-bulbs isn't well-filtered. There is a subtle 120Hz flicker when I TX and the relays load this lightly-filtered DC supply. Again, the slight dimming and slight flicker in TX seems a beautiful symbolism for my 30S-1 "picking up and shouldering the load". Appropriate for the era!

Updates and Changes

2025-01-17T01:46:01Z

Gordonp: /* Dial Lights "ON" when Amp is Ready */

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

===LEDs===
So, when I mentioned above that the dial-bulbs are now operating from ~115VDC ...
I replaced the dial-bulbs with some ubiquitous "pinball" bulbs, meant to replace #44 and #47 incandescent bulbs. My particular LED bulbs have a diode-bridge in the base and are DC-polarity insensitive. However, they were designed to operate on 6.3VAC, not my ~8VDC :-O At first, they were great and very bright. After a while, they began to "stutter" - not really "flicker" and certainly in no relation to the ripple, but a slower, cyclical, sometimes opposite to each other. Hmmm...

I dug in: checked all connections, cleaned the bulb-socket contacts, bypassed all safety and thermal-timer connections and otherwise made sure the 30S-1 itself didn't generate the flicker at any point in the 12VAC circuitry. Then, I got out the oscilloscope, to check things: the mid-voltage between the LEDs was erratic and crazy. From this, I conjectured the LED bulbs were failing internally, possibly due to overheating. Replacing both the high- and low-side bulbs brought immediate relief and proper operation, but I knew it would last ...

So - from my always-HOT side of S208 (which supplies the dial-LED bulbs) I added a 5.1V zener diode, effectively knocking the voltage down something like 5.5V actual. This results in the pair of LEDs receiving ~10V total, or ~5V across each. This reduction from 6.3VAc to 5VDC resulted in a less-bright and more-yellowish light... which I find AWESOME!! And I hope they'll last a very long time.

NOTE: LEDs are "fast", and the DC I now use for the dial-bulbs isn't well-filtered. There is a subtle 120Hz flicker when I TX and the relays load this lightly-filtered DC supply. Again, the slight dimming and slight flicker in TX seems a beautiful symbolism for my 30S-1 "picking up and shouldering the load". Appropriate for the era!

Updates and Changes

2025-01-17T01:45:47Z

Gordonp: /* Dial Lights "ON" when Amp is Ready */

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

==LEDs==
So, when I mentioned above that the dial-bulbs are now operating from ~115VDC ...
I replaced the dial-bulbs with some ubiquitous "pinball" bulbs, meant to replace #44 and #47 incandescent bulbs. My particular LED bulbs have a diode-bridge in the base and are DC-polarity insensitive. However, they were designed to operate on 6.3VAC, not my ~8VDC :-O At first, they were great and very bright. After a while, they began to "stutter" - not really "flicker" and certainly in no relation to the ripple, but a slower, cyclical, sometimes opposite to each other. Hmmm...

I dug in: checked all connections, cleaned the bulb-socket contacts, bypassed all safety and thermal-timer connections and otherwise made sure the 30S-1 itself didn't generate the flicker at any point in the 12VAC circuitry. Then, I got out the oscilloscope, to check things: the mid-voltage between the LEDs was erratic and crazy. From this, I conjectured the LED bulbs were failing internally, possibly due to overheating. Replacing both the high- and low-side bulbs brought immediate relief and proper operation, but I knew it would last ...

So - from my always-HOT side of S208 (which supplies the dial-LED bulbs) I added a 5.1V zener diode, effectively knocking the voltage down something like 5.5V actual. This results in the pair of LEDs receiving ~10V total, or ~5V across each. This reduction from 6.3VAc to 5VDC resulted in a less-bright and more-yellowish light... which I find AWESOME!! And I hope they'll last a very long time.

NOTE: LEDs are "fast", and the DC I now use for the dial-bulbs isn't well-filtered. There is a subtle 120Hz flicker when I TX and the relays load this lightly-filtered DC supply. Again, the slight dimming and slight flicker in TX seems a beautiful symbolism for my 30S-1 "picking up and shouldering the load". Appropriate for the era!

Main Page

2021-06-30T21:52:13Z

Gordonp:

{| border="1" cellpadding="6" cellspacing="0" style="border: black solid 1px; border-collapse: collapse; text-align: left; width: 100%; background: #f0f0ff; "
| colspan="3" style="text-align: center; font-weight: bold; font-size: 1.4em;" |
Information, Repairs, Upgrades and Notes for Collins 30S-1 HF Amplifier
|-
| width="33%" valign="top" | '''Gord - VA7GP''' [mailto:gordon.pritchard+30s1@gmail.com email me!]
[https://tigerfire.ca/radio/collins_30s1_amp.html (Original, historical single-page site)]
 
* [[Acquisition]]
**[[Acquisition#A Bit of History|A Bit of History]]
**[[Acquisition#Preparation|Preparation]]
**[[Acquisition#First Power-Up|First Power-Up]]
 
* [[Repair]]
**[[Fixing Damaged and Abused R218|Fixing Damaged and Abused R218]]
**[[Dealing with the Overheated and Cracked R232|Dealing with the Overheated and Cracked R232]]
**[[Dealing with the RF Output|Dealing with the RF Output]]
**[[BIAS Transformer T203, and Associated Circuitry|BIAS Transformer T203, and Associated Circuitry]]
**[[Screen Supply Overhaul|Screen Supply Overhaul]]
**[[12V Safety and Control|12V Safety and Control]]
**[[CR216 and BIAS Multi-Metering|CR216 and BIAS Multi-Metering]]
**[[Primary AC Wiring|Primary AC Wiring]]
**[[Dials|Dials]]
**[[Power Supply & Lower Compartment|Power Supply & Lower Compartment]]
**[[Tuning & Loading Meter Circuit|Tuning & Loading Meter Circuit]]
 
* [[Testing]]
**[[Testing#120V Initial Testing|120V Initial Testing]]
**[[Testing#240V Testing|240V Testing]]
 
* [[Updates and Changes]]
**[[Updates and Changes#Boosting Screen Voltage|Boosting Screen Voltage]]
**[[Updates and Changes#Raising CW Plate- and Screen-Voltages|Raising CW Plate- and Screen-Voltages]]
**[[Updates and Changes#Installing the 4CX1500B|Installing the 4CX1500B]]
**[[Updates and Changes#Cooling|Cooling]]
**[[Updates and Changes#Thermal Overload Switch K102|Thermal Overload Switch K102]]
**[[Updates and Changes#Dial Lights "ON" when Amp is Ready|Dial Lights "ON" when Amp is Ready]]
 
* [[Operation]]
**[[Operation#Dayton Blower Sound Level|Dayton Blower Sound Level]]
**[[Operation#Dial Lights After Warm-Up|Dial Lights After Warm-Up]]
**[[Operation#Spares and Servicing|Spares and Servicing]]
**[[Operation#Contests: SSB and CW|Contests: SSB and CW]]
**[[Operation#Tuning|Tuning - How Often and How?]]

| width="33%" valign="top" | '''Chris - KB3BF'''
* [[ 160m Mods]]

| width="33%" valign="top" | '''Chet - VE3CFK'''
* [[The Acqusition]]
* [[The Game Plan]]
* [[Plan Execution]]
* [[Moment(s) of Truth]]
* [[Modifications]]
* [[Completed Collins 30S-1]]

|}

[[File:30s1_frontview.jpeg|500px|thumb|center]]

Updates and Changes

2021-06-30T21:47:18Z

Gordonp:

__FORCETOC__
=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

Updates and Changes

2021-06-29T22:11:38Z

Gordonp: /* Boosting Screen Voltage */

__FORCETOC__
=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.



[[File:Screen_boost_installation.jpg|500px|thumb|left|Completed Screen Boost Installation]]

The connections to the barrier-strip are clearly labelled; the wires are neatly laced - pretty much the way Collins might have done it :-)


 

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

File:Screen boost installation.jpg

2021-06-29T22:11:25Z

Gordonp: Gord - image showing installation of screen-voltage-boost transformer

== Summary ==
Gord - image showing installation of screen-voltage-boost transformer

Updates and Changes

2021-06-29T22:08:24Z

Gordonp: /* Boosting Screen Voltage */

__FORCETOC__
=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

<gallery caption="Measuring Screen Boost Voltages" widths="400px" heights="400px" perrow="2">
File:Screen_boost_ssb.jpg|SSB Screen Boost - about 100VDC
File:Screen_boost_cw.jpg|CW Screen Boost - about 40VDC
</gallery>

Again, three meters show the boost for SSB, and CW. The CW Screen Voltage reading appears lower-than-expected, because these measurements are made prior to raising the Plate + Screen voltage; thus, the CW primary-winding voltage is only about 2/3 that obtained in SSB mode.

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

File:Screen boost cw.jpg

2021-06-29T22:08:13Z

Gordonp: Gord - screen voltage boost in CW mode

== Summary ==
Gord - screen voltage boost in CW mode

File:Screen boost ssb.jpg

2021-06-29T22:07:10Z

Gordonp: Gord - screen voltage boost in SSB mode

== Summary ==
Gord - screen voltage boost in SSB mode

Updates and Changes

2021-06-29T22:02:12Z

Gordonp: /* Boosting Screen Voltage */

__FORCETOC__
=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Boosting Screen Voltage=

I bought a 100VA 120V transformer, with the intention of boosting the screen voltage. This Triad FD8-120 features a pair of primary windings - I connected these windings in series for 240VAC operation, and wired them to T201 SSB terminals, with the result:
<ul>
<li>When the user alters the AC input-jumper, this boost will follow and operate from either 240VAC (as now), or 120VAC, just like T201</li>
<li>This connection should give 100V screen-supply boost in SSB, and a lower 65V boost in CW, IF I HAD NOT ALTERED THE SSB/CW 240VAC PRIMARY CONNECTIONS. But in my particular case, the Screen Boost will always be the SSB value.</li>

The boost will only come alive once the front-panel ON HV button is pressed.
Repeating the same (no-load) measurements as above:

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

Updates and Changes

2021-06-29T21:57:43Z

Gordonp:

__FORCETOC__
=Raising CW Plate- and Screen-Voltages=
Collins' reduction in High Voltage and Screen Voltage when using CW mode held back the 4CX1500B gain, and total-power. I was almost at the point of using the SSB setting for CW (as many do), when Chet VE3CFK pointed out...

That feeding both the primary wires from the front-panel-switch to the SSB connections on the transformer would give me a constant 3kV plate-voltage, and the elevated screen-voltage, no matter the mode. In fact, doing this means the only difference between SSB and CW is the negative grid-bias, and the resulting Class of operation.

The realization of Chet's suggestion: simply moving the CW-primary-power wire from Term #1 to Term #2, and similarly moving Term #6 to Term #5:
[[File:HV_screen_always_at_SSB.jpeg|500px|thumb|center]]
Now I have good gain and total power! Thanks, Chet!
Here's a view of comfortable operating conditions now, for CW:
[[File:cw_operation_boosted_HV_screen.jpeg|500px|thumb|center]]

Raising the Screen supply meant my C204 Voltage rating was marginal... Here, I did a mod to a mod ;-) 500V / 105C rating are improvments over the 450V/85C new capacitor I recently installed.
[[File:upratedC204.jpeg|500px|thumb|center]]

=Installing the 4CX1500B=



[[File:4cx1500b_ready_to_install.jpg|Ready...|500px|thumb|left]]



 



[[File:4cx1500b_installed.jpg|AIM...|500px|thumb|right]]



 

Date-code of (mid-December) 1984. 12 hours of filament-only to getter the tube, after such long storage.



[[File:4cx1500b_making_power.jpg|FIRE!|500px|thumb|center]]

This photo shows 1kW into my dummy-load, making it sweat a bit :-)
But my calculation of the DC input power, now compared with the RF output-power, leads me to question my Plate Current meter accuracy. Another detour...



 
=Cooling=
Collins' original blower is very quiet. Nicely quiet. This is pretty much a result of laminar-airflow over the tube-fins, with a (nominal) 1800rpm motor.

The fin-design of the 4CX1500B includes offset-punched sections of each fin, intentionally to cause turbulence and remove more heat. I made that much worse :-) by swapping blower-motors - I swapped in the popular 3000rpm Dayton 4M093E. Some correspondance and digging through Collins' maillist archives led me to choose a 50-ohm series-resistor, primarily to keep the motor from overheat-tripping (this motor will overheat without some sort of slow-down, because it's intended to have airflow over the motor, and in this 30S-1 application it does not have cooling airflow over it).



[[File:Preparing_blower_motor_change.jpg|500px|thumb|left]]

The line-up of prime-suspects for the blower-motor change :-)


 



[[File:Dayton_motor_installed.jpg|500px|thumb|right]]

Motor changed, resistor mounted to housing, blower re-assembled. Home at long last :-) The lower mounting bolt took me 4hrs just to get it installed. I later read Mr. Carn's article in "The Signal" which says it should be a stud - they say the best advice comes just after the job is completed :-O The only one good thing about my (re-)using a bolt: it was easy to adjust the front mount, so that the weight of the motor is hanging equally from the top-mounts - just look at the "angle of the dangle" for this bolt, in the power-supply compartment. It's neutral and balanced when the bottom bolt hangs perfectly vertical.
The 50-ohm power resistor is screw-mounted on the lip of the blower intake, where it will receive some cooling, but not noticeably obstruct airflow.



 

My cut-off tool got a workout: the Dayton motor has front and rear mounting-bolts; I didn't like the knuckle-slicing appearance of the rear ones so I cut them off and added some protective heatshrink. I trimmed the front bolts down, to allow more lattitude in adjusting the squirrel-cage. And my 1/4" nut-driver was too long to fit into the blower compartment, so I cut that too :-)

For this 50-ohm resistor, I measured:
<ul><li>0.980 amps without resistor nor any nozzle-restriction (121VAC)</li>
<li>0.76 amps with resistor in steady-state (10-12s after start)</li>
<li>93VAC across motor, with resistor, in steady-state</li></ul>
Pretty much, this shifts the original motor-dissipation from 120W without a resistor, down to 70W in the motor and 32W in the resistor. Subjectively, the airflow seems the same, but the motor-temperature will be 'way down.

Test driving the new blower made my beach-towel wave in the wind, from 3 feet away :-) Too much is just about right :-)



[[File:Blower_performance_estimate.png|500px|thumb|left]]

An idea of how this sytem will work, when the forces of Hot and Cold do battle! Looks like the forces of Cold will win! This graph was generated with 1500W dissipated in the 4CX1500B plus 20% margin... I won't operate at this point, so I'll have even more margin when I operate at more-modest levels.
To push the tube up against the cooling-limits would require something like 3kW DC input!!! Even bleary-eyed, at the end of a long contest, starved for food and water, I am incapable of damaging my 30S-1 :-)



 

Predicted performance was reassuring, but I felt the need to measure and determine actual performance.
Tachometer: 3214rpm
Anode Temperature (750W CW Output): 
Manometer Pressure: 

<gallery caption="Measuring blower-rotational-speed" widths="400px" heights="400px" perrow="2">
File:30s-1_blower_tach_reflector.jpeg|Reflective Tape Applied to Blower-Squirrel-Cage Perimeter
File:30s-1_dayton_rpm.jpeg|30S-1 Blower Speed (rpm) Measurement
</gallery>

And Then...

I bought a Variable-Voltage / Variable Frequency controller! This V/F motor drive is ~perfect~ for the Dayton shaded-pole motor... My Plan: slow the motor when possible and enjoy quiet-ness, then speed it up when cooling is required. This V/F drive can be controlled with a 0-5V signal, so ... :-)

=Boosting Screen Voltage=

=Thermal Overload Switch K102=
My thermal sensor kicked my amp offline a couple of times, at only modest power-levels. Perhaps decades of time have taken their toll on the sensor; perhaps tube-changes have un-calibrated it; certainly my blower-motor-change will have altered it's response. The purpose of the thermal sensor: to open the HV-enable 12V circuit and K203 (primary AC power) if the tube gets too hot. The tube can get too hot with either excess dissipation, or lack of cooling-airflow.

I changed this protection to accurately focus on protecting the tube against overtemperature and also against loss of cooling airflow.

[[How and Why I Changed It]]

<gallery caption="Thermal and Pressure switches to replace Collins Overtemp K102" widths="400px" heights="400px" perrow="2">
File:rear_view_pressure_temp_sensors.jpeg|Rear View of the RF Compartment, Showing New Thermal-Safety
File:Top_view_pressure_temp_sensors.jpeg|Top View of the RF Compartment, Again Showing Thermal-Safety
</gallery>

<gallery widths="400px" heights="400px" perrow="2">
File:Air_sense_hose_under_boot.jpeg|Air-Pressure sensing tube simply stuck up between blower nozzle, and flexible coupling-boot
</gallery>

=Dial Lights "ON" when Amp is Ready=

I re-wired the dial-lights so they will illuminate only after the warm-up delay has passed and K202 closes. Not only does this indicate visually that the amp has completed the 3min warm-up and ready to hit the "ON" push-button, but it also confirms 12V is available to energize K203 - maybe this will help troubleshooting one day.
(Technical note: this "delay completed" lighting moves the dial bulbs from 12VAC to ~15VDC operation).

Final touch for the dial-light mod: hand-lacing!

[[File:Bulb_mod_cable_lacing.jpg|500px|thumb|center]]

Main Page

2021-06-29T21:56:55Z

Gordonp:

{| border="1" cellpadding="6" cellspacing="0" style="border: black solid 1px; border-collapse: collapse; text-align: left; width: 100%; background: #f0f0ff; "
| colspan="3" style="text-align: center; font-weight: bold; font-size: 1.4em;" |
Information, Repairs, Upgrades and Notes for Collins 30S-1 HF Amplifier
|-
| width="33%" valign="top" | '''Gord - VA7GP''' [mailto:gordon.pritchard+30s1@gmail.com email me!]
[https://tigerfire.ca/radio/collins_30s1_amp.html (Original, historical single-page site)]
 
* [[Acquisition]]
**[[Acquisition#A Bit of History|A Bit of History]]
**[[Acquisition#Preparation|Preparation]]
**[[Acquisition#First Power-Up|First Power-Up]]
 
* [[Repair]]
**[[Fixing Damaged and Abused R218|Fixing Damaged and Abused R218]]
**[[Dealing with the Overheated and Cracked R232|Dealing with the Overheated and Cracked R232]]
**[[Dealing with the RF Output|Dealing with the RF Output]]
**[[BIAS Transformer T203, and Associated Circuitry|BIAS Transformer T203, and Associated Circuitry]]
**[[Screen Supply Overhaul|Screen Supply Overhaul]]
**[[12V Safety and Control|12V Safety and Control]]
**[[CR216 and BIAS Multi-Metering|CR216 and BIAS Multi-Metering]]
**[[Primary AC Wiring|Primary AC Wiring]]
**[[Dials|Dials]]
**[[Power Supply & Lower Compartment|Power Supply & Lower Compartment]]
**[[Tuning & Loading Meter Circuit|Tuning & Loading Meter Circuit]]
 
* [[Testing]]
**[[Testing#120V Initial Testing|120V Initial Testing]]
**[[Testing#240V Testing|240V Testing]]
 
* [[Updates and Changes]]
**[[Updates and Changes#Raising CW Plate- and Screen-Voltages|Raising CW Plate- and Screen-Voltages]]
**[[Updates and Changes#Installing the 4CX1500B|Installing the 4CX1500B]]
**[[Updates and Changes#Cooling|Cooling]]
**[[Updates and Changes#Thermal Overload Switch K102|Thermal Overload Switch K102]]
**[[Updates and Changes#Boosting Screen Voltage|Boosting Screen Voltage]]
**[[Updates and Changes#Dial Lights "ON" when Amp is Ready|Dial Lights "ON" when Amp is Ready]]
 
* [[Operation]]
**[[Operation#Dayton Blower Sound Level|Dayton Blower Sound Level]]
**[[Operation#Dial Lights After Warm-Up|Dial Lights After Warm-Up]]
**[[Operation#Spares and Servicing|Spares and Servicing]]
**[[Operation#Contests: SSB and CW|Contests: SSB and CW]]
**[[Operation#Tuning|Tuning - How Often and How?]]

| width="33%" valign="top" | '''Chris - KB3BF'''
* [[ 160m Mods]]

| width="33%" valign="top" | '''Chet - VE3CFK'''
* [[The Acqusition]]
* [[The Game Plan]]
* [[Plan Execution]]
* [[Moment(s) of Truth]]
* [[Modifications]]
* [[Completed Collins 30S-1]]

|}

[[File:30s1_frontview.jpeg|500px|thumb|center]]

Testing

2021-06-29T21:54:10Z

Gordonp: /* 240V Testing */

=120V Initial Testing=
<ul>
<li>I installed the original 4CX1000A tube which came in my amplifier. Only the 3.2A fuse was installed, and I applied power!</li>
<li>The meters illuminated; after the time-delay relay closed, the dials also illuminated. The "ON / OFF" switch functioned.</li>
<li>Opening either the top-lid or the front-panel removed the dial-illumination, and prevented operation of the "ON / OFF" switch.</li>
<li>I tested the Thermal Overload switch K102 by moving it out of, and back into the airstream - it works, and disables the "ON / OFF" switch when there is no airflow over the sensor.</li>
</ul>
The Safety and Control seem to work. The Bias and Filament adjustments work.



[[File:Filament_voltage.jpg|500px|thumb|left|Reasonable correlation between panel-reading and at-the-socket measurement]]

I clipped some test-leads right onto the tube-socket, in order to measure and adjust filament-voltage accurately. Eimac specifies 6.0 +/-5%, so that's what I gave it.
The panel-meter actually reads the incoming 115VAC supply, using that as a proxy for the actual filament voltage - you can see this does a pretty good job.


 

With an eye on the amp, I ran it this way for 8 hours, to getter the tube. And to verify that the blower doesn't go into thermal-shutdown. And to get the various smells flushed out (especially that new Dayton motor - it runs quite hot).

=240V Testing=

After warm-up, I pressed "ON", and got a nice "clunk" from relay K203. But no High Voltage :-( I metered and monitored the primary AC windings of Transformer T201 - nothing! That pretty much points to K203 as the culprit. Out came the Relay Shelf, for a focussed look. Removal notes:

<ul>
<li>Remove the bakelite AC power cover, and disconnect all the power leads</li>
<li>Remove the 5 front / 2 rear screws, and shift the Shelf over - to allow removal of the HV shorting wire (hidden behind edge of compartment)</li></ul>
Servicing Note:
<ul>
<li>use a paint-stir-stick, behind the fuse-panel, to wedge and hold open the HV short, and keep closed the 12VDC interlock switch. Eyes Open!</li></ul>



[[File:Relay_contact_adjust.jpg|500px|thumb|left|Contact adjustment tools: shaving mirror, burnisher, contact-bender]]

With the Relay Shelf flopped onto it's side, I determined that the rearmost contact of K203 was failing to close. I burnished the contacts, then adjusted both such that they close when the relay armature has moved about 2/3 of it's travel.


 

<gallery caption="Measuring Per-Mode Plate Voltages" widths="400px" heights="400px" perrow="2">
File:HV_cw.jpg|CW Plate Voltage: 2050VDC
File:HV_ssb.jpg|SSB Plate Voltage: 2750V
</gallery>

Quick test: I couldn't get any plate-idle-current. Recall, I had pre-set the BIAS Voltage to -60V, to be on the conservative side and start out with a reduced idle-current. But now, I had no idle-current. A further quick test showed infinite input-SWR for the exciter. Two problems to look into...

So, next step: let's thoroughly check screen supply circuitry.

I got my 12VDC bench-supply, and verified the switching of K205, and also that K101 would also trigger K205. Both were good. Basic ohmmeter checking showed connectivity from the screen supply to the cathode-chock L103. Hmmm... let's thoroughly check the AC and DC values related to the Screen supply.



[[File:Screen_supply_check.jpg|500px|thumb|left|Instrumenting the AC input, AC output and DC output]]

I got three meters, and carefully measured:
<ul>
<li>AC input voltage on primary of T201, after fuses, switches, relays and wiring: 240VAC</li>
<li>AC output voltage from Screen secondary of T201: 260VAC</li>
<li>DC output of the Screen supply, across C204: 235.2VDC</li></ul>


 

We have connectivity, we have all voltages. Hmmm... maybe I just have to crank harder on the BIAS Adjustment. So, I did. I managed to achieve the prescribed 200mA idle-current in SSB mode, but only when the BIAS was down to -36V! I wonder if my tube is nearly-shot?

Now... RF in! But wait - I have infinite input SWR?!?! :-(

I metered every cable and connection along the RF input path, and found that the culprit was K205 at the back of the power compartment. It appears that my burning T203 deposited a tarry film on the open contacts. After cleaning, I tested K205 only - now good! Along the path of debugging, I removed and cleaned the input band-switch (cut strips of bond paper, spray contact-cleaner on the strips, use tweezers to sneak the paper between the contacts and drag it through).

RF in now brings RF out, but it isn't very impressive with this particular tube. Roughly, I drove it with 50W CW input, 900W DC to the final, and 400W out. As I increased the drive, I began to note slight movement toward -0.2mA grid-current... this tube simply wasn't going to give anything more. Time to get Contestant #2 - a 1979 used 4CX1000A which looked a lot less darkened and still has Eimac paint on it.

A day later...

Tube #2 is much better - I obtained a gain of about 7.3: 75W RF input drive, 900W DC plate input, 550W RF output to the dummy-load. I'll set this aside as a tested spare.

File:Screen supply check.jpg

2021-06-29T21:53:49Z

Gordonp: Gord - checking 30S-1 screen supply voltage

== Summary ==
Gord - checking 30S-1 screen supply voltage

Testing

2021-06-29T21:50:57Z

Gordonp: /* 240V Testing */

Testing

2021-06-29T21:50:31Z

Gordonp: /* 240V Testing */

File:HV ssb.jpg

2021-06-29T21:50:26Z

Gordonp: Gord - measuring High Voltage in SSB Mode

== Summary ==
Gord - measuring High Voltage in SSB Mode

File:HV cw.jpg

2021-06-29T21:48:58Z

Gordonp: Gord - High Voltage in CW Mode

== Summary ==
Gord - High Voltage in CW Mode

Testing

2021-06-29T21:44:23Z

Gordonp:

File:Relay contact adjust.jpg

2021-06-29T21:43:02Z

Gordonp: Gord - 30s-1 Relay-contact-adjustment tools

== Summary ==
Gord - 30s-1 Relay-contact-adjustment tools

Testing

2021-06-29T21:41:58Z

Gordonp: /* 240V Testing */

Testing

2021-06-29T21:40:26Z

Gordonp:

Testing

2021-06-29T21:39:50Z

Gordonp: /* 120V Initial Testing */

Testing

2021-06-29T21:39:23Z

Gordonp: /* 120V Initial Testing */

Testing

2021-06-29T21:38:24Z

Gordonp: /* 120V Initial Testing */

File:Filament voltage.jpg

2021-06-29T21:37:35Z

Gordonp: Gord - 30S-1 filament voltage at tube-socket, comparing built-in metering to bench-meter reading

== Summary ==
Gord - 30S-1 filament voltage at tube-socket, comparing built-in metering to bench-meter reading