RF Triode Filament Monitoring Circuit

[K1JOS]

I have a ham radio RF amp (Alpha 87a) running two 3CX800A7 ceramic triodes (all else solid state) with a lot of protective circuitry.  I have been getting intermittent faults on the filament monitor for the 3CX800A7’s showing the filament current is low  - suggesting that one of the tubes has an open filament.  I know this is not the case as it is intermittent fault and when not faulted I can get full output power so I think it’s an issue with the filament monitoring circuitry.  I understand the basic concepts of current monitoring using a parallel resistive shunt load and feeding that voltage drop into an op amp to provide a corresponding proportional voltage that could be read by a microprocessor analog input.  My amp does not require precision, it is designed to shut down the amp if the filament current is too low (<2 A) or too high (>4 A).

Attached are the two schematics that show the filament monitoring circuit.

The transformer’s 13.6 vac go via lines labeled ‘FIL’ and “FIL RETURN” through the Low Voltage Board, these are passed through to the Tube deck along with a 3rd line called  FIL_SENSE GND.    On the Tube Deck,  FIL is connected to one side of the filaments and the other side of the filaments are grounded.  FIL_SENSE GND is connected directly to ground  and FIL_RETURN is connected to ground through a series 0.3 ohm, 5W voltage dropping resistor.  At first I figured the filament monitoring circuit is sensing the current through this 0.3 ohm dropping resistor but both sides of the sensing points (FIL_SENSE_GND and FIL_MON are at or near ground and nothing is shunting current from the FIL and FIL_RETURN lines.  I cannot figure out how the monitoring circuit is working and would appreciate some advice.  I hope someone can volunteer to help me on this issue.

Many thanks in advance

Jerry NY2KW

[SeanB]

Simple enough, it gets the AC voltage across the sense resistor, around 0.9VAC or so, and this is passed via C27 to a simple voltage multiplier comprised of D19,20 and C25,26,27 so that a DC voltage roughly proportional to filament current is present across R35. U8A and D26 does some gain and clips it so that you only get a DC voltage applied to the filament monitoring pin which is clipped so it will be 5V if the current is ok and 0V if not.

At a guess you should first change C26,27 with new ones, as these are likely getting elderly, and low value electrolytics are always suspect. As well you can monitor U8 pin 1 and it will show you an analogue value that should be reasonably steady at around 4-8V, with no noise showing.

Of course you have checked for dry joints on the filament connectors and that the wiring and interconnects are all fine, and no dry joints on the sense resistor.

[K1JOS]

Thanks SeanB but I don't think it's that simple.  The sense circuit input is between FIL_RETURN and FIL_SENSE_GND.  One is tied directly to GND and the other is tied to GND via R7.  Where is the potential difference coming from between these two points?

Jerry NY2KW

[Fank1]

The voltage is applied from the transformer to the filament return ground and the filament.
One of the filament pins is directly connected to ground and then to filament return ground thru the resistor.
In order to get a complete path to the filament transformer the current must flow thru the resistor.
I agree change the caps.
WA8HTO

[AF6LJ]

It is as simple as SeanB explained it.

And yes change those caps....

[K1JOS]

Yes... duhhh I was working on a simulation and drew the schematic wrong.  The AC voltage drop across R7 of about 1v is being fed into the diode and op-amp.  Boy, i stared at this for such a long time and now its like the nose on my face.  Sorry and thanks !!

[K1JOS]

One more question please for those who have looked at the schematics.  What is the function of L1 (15uH) in the sensor circuit which is grounded on both sides (one side on the LV board) and the other side on the Tube Deck via FIL_SENSE_GND?    The tube deck and LV Board are hard bonded to the amp chassis. 

tnx
jerry

[AF6LJ]

One more question please for those who have looked at the schematics.  What is the function of L1 (15uH) in the sensor circuit which is grounded on both sides (one side on the LV board) and the other side on the Tube Deck via FIL_SENSE_GND?    The tube deck and LV Board are hard bonded to the amp chassis. 

tnx
jerry

It looks like the choke in question provides a DC return for the the filament sense circuit while providing a high impedance to RF.

[K1JOS]

Hi Sue,

Please bear with me -- if you look at the schematics the connector FIL_SENSE_GND is grounded on the Tube deck so L1 is immediately grounded on both of its sides - one side to tube deck and other side on LV Board but the LV board is bonded to the tube deck via the chassis bolts/screws.  Wouldn't the ground on both sides of L1 defeat any RF choke function?  This is where my basic knowledge has a huge gap with practical circuits.

jerry 

[AF6LJ]

RF tends to get places you don't want it, even when you have a board and a sub chassis that are connected with a common ground. The ground wire can pick up RF internal to the amplifier and that can be problematic.

Without seeing the internals that is my best educated guess.

[K1JOS]

Thank you for the clear explanation.  I am digging further into the schematics to see how the designers monitor and create protection faults for out of range plate HV, plate current, etc.  Looking at this in a real working amplifier is the best way for me to learn new things otherwise theory becomes too abstract.

Jerry NY2KW

[AF6LJ]

Just be careful there are voltages in there that can kill you dead'er than a hammer...

[dave_k]

RF tends to get places you don't want it, even when you have a board and a sub chassis that are connected with a common ground. The ground wire can pick up RF internal to the amplifier and that can be problematic.

Without seeing the internals that is my best educated guess.

There are 2 different ground symbols on the schematic, so there might be 2 different grounds used in the equipment.

[AF6LJ]

Indeed.