Vacuum Tube Power supply fault

[scsi]

Hi all! I need advice on troubleshooting tube power supply. Problem - negative voltage on 0 - 600V channel output. All the components seems to be in tolerance, but I can't find a single clue what is wrong with it. I measured voltages and added a schematics and already see inproper bias for tubes. At this point all the pass tubes were taken out. How should I proceed?


Updated schematics with component values and measured voltages are here:
https://www.eevblog.com/forum/repair/vacuum-tubes-power-supply-fault/msg3062998/#msg3062998

[trobbins]

I can suggest inserting one rectifier group at a time, using ss diode dummy valves, and use a variac to bring up each raw DC rail and confirm no faults for the basic rectifier sections and filter caps.  Make sure there is a discharge resistor added across each raw filter cap for safe discharging when valves are out.

Then focus on each regulator section one at a time, using the dummy ss diodes and variac, to bring up the DC raw rails slowly and confirm the regulation circuitry is working as expected. 

Faultfinding the top section of parallel pass tubes is likely to only need one pass tube in to check functionality.

I would strongly suggest adding in sufficient series 1N4007 in series with each valve diode anode, as a way to reduce PIV stress on the valve diodes, and avoid internal arcing when a valve diode ages sufficiently to start failing.  That may also save some collateral damage to the power transformer should a valve diode start arcing.

I'd also do a suite of insulation resistance tests on all the various secondary windings, to confirm no unexpected leakage paths at at least 500VDC and preferably 1kVDC - perhaps best done at the start of testing.

You really do need to know what you are doing here, and have suitable test tools, as this equipment may bite if you don't have a competent awareness of safety processes and risk mitigation techniques and a thorough understanding of the circuitry.

[scsi]

In a meantime powering up with only one rectifier (L14) (All other tubes out) with variac already gives negative voltage on the output. Does that make any sense?

[trobbins]

Yup - try a dummy ss diode instead of the valve diode - always handy to have ss diode dummy's to plug in for testing.

[scsi]

Put some ss in place of T14 and have the same negative voltage on ouput. All the tubes were out at this point. If I do the same thing on 0-400V (working channel) with ss no negative on ouptup whatsoer.

[Chris56000]

Hi!

Have you got the full manual with the component values or a link to where you got that circuit from so I can make a clearer and better diagram please, and then go through it in more detail please?

Chris Williams

[scsi]

here you go

https://drive.google.com/open?id=1LFl_gpEtUgUKDxGCXbSIMoCO86UujIi_

very much appreciated!

P.S. It's the last time I try to fix something with a lack of appropriate knowledge. I'll go study first

[ArthurDent]

Using the negative terminal as common and a properly rated DMM, carefully check the following:
1)The voltage on C1 which probably has a 1Kv rating and will have around +800 VDC on it.
2)The plate (anode) voltage on all 8 pass tubes (5 to 12). This will also be around +800 VDC.
3) Tube 13 is the difference amplifier that drives the pass tubes. It will have a grid #1 voltage of low plus to minus VDC from potentiometer R32, the fine voltage adjust.
3)Tube 13’s cathode is the reference voltage and potentiometer 32 will vary the grid #1 voltage either side of what the cathode voltage is to vary tube 13’s  plate voltage from about zero to plus some high positive voltage to vary the conduction of the pass tubes.
4)If there is a low negative voltage on the positive 0-600 VDC output terminal then it is probably coming from the negative supply through the path from the low side of the potentiometer and there is an open in the positive high voltage side like an open mA meter.

Trying to diagnose a problem like this by posts is like having your doctor diagnose your ailments over the phone – not going to be too reliable. Please be careful with these high voltages and consider my advice to troubleshoot this schematic, in what is to me an unknown language, for what you payed for it. I do hope my comments will help though. Below is a much simpler power supply I made years ago and if you drew a block diagram of mine and yours you would see all the same types of functional pieces. This might help you as well.

[schmitt trigger]

With the unit disconnected and all of the capacitors completely discharged (double check this step!!) and utilizing the continuity function in your DMM, rotate  the range-changing switch to make sure that it is making proper contact in all poles and all positions.

Could you share some photos of the unit?

[JohnPen]

Just a thought does the Negative voltage vary when you operate the V range selector particularly when you only had T14 present.  Also as suggested earlier check the mA meter is working, not o/c, for continuity.  No power on for latter test.  

[scsi]

Scary amouts of voltages

1)The voltage on C1 which probably has a 1Kv rating and will have around +800 VDC on it.

1010V

2)The plate (anode) voltage on all 8 pass tubes (5 to 12). This will also be around +800 VDC.

1008V +- 1V

3) Tube 13 is the difference amplifier that drives the pass tubes. It will have a grid #1 voltage of low plus to minus VDC from potentiometer R32, the fine voltage adjust.

Voltage on T13 1st grid was -342V

also T5-12 have -52V on cathodes

Trying to diagnose a problem like this by posts is like having your doctor diagnose your ailments over the phone – not going to be too reliable

I do understand that, and appreciate that you are trying I'll do my part as good as I can. This unit was actually for learning and experiment with tube circuits and it supposed to work, but oh well. Now I am learning the hard way.

[trobbins]

Are the measurements in previous post when all valves are now back in place?  What switch (T2) setting is being used?  Are there any loads on the output terminals?  Are the voltage readings relative to the 0V terminal of the 60-600V output?

Assuming all valves are in, and measurements are relative to 0V terminal of 60-600V output, and no external loads applied, and highest secondary voltage is fed from T2 switch:

C1 = 1kV would imply using the max voltage range switch setting, and no loading on that raw rail.

All plate voltages show the same as C1 voltage because that is the same voltage rail, so no new information (unless it was the plate to cathode voltage being measured).

T13 = -342V indicates the negative voltage provided by the second rectifier/regulator is possibly not being divided properly by R27-29 and R30, so I would check those resistors are correct valued and connected to their respective end circuit points.

T5-T12 cathodes with -52V.  This indicates no current through T5-T12, and the common cathode line is pulled to the negative rail provided by the secondary rectifier through the string of resistors R34-43, R32//R33, and R44-49.

The original red-lined schematic in post #1 appears to show some voltage readings, but it is ambiguous as to whether the stand-along voltages are relative to the 0V terminal of the 60-600V output, or the negative rail of the second regulator section, or ...?

[scsi]

All measurements were done with all tubes in place using negative terminal of 20-600V channels as common. No load were on output. Selector switch - centered.

Summary of all bias voltages for tubes:

T5-T12 : Anode 1010V, Cathode -52V

T13 : Anode -345V, 1st grid -348V, Screen -342V

T15 : Anode 58V

T16: Anode -1.9V, 1st grid -315V, Screen -335V, Cathode -360V

T13 = -342V indicates the negative voltage provided by the second rectifier/regulator is possibly not being divided properly by R27-29 and R30, so I would check those resistors are correct valued and connected to their respective end circuit points.

Rechecked R27-29 300K resistors in series and R30 28k seems to be fine and connected to their respective end circuit points.

Addition: when I am measuring 20-600V channels output between + and - and moving selector switch from lowest to largest voltage scale I get:

(position: voltage)
1: -128V
2: -88V
3: -56V
4: -30V
5: -4V

[trobbins]

That implies R18 (2Meg) has 1. 35kv across it?
And valve 16 has input grid sitting at +45V, which can't be ok condition.

[scsi]

That implies R18 (2Meg) has 1. 35kv across it?

R18 is two 1M resistor in series acros one of them 620V

[ArthurDent]

I have redrawn the schematic to simplify it. The positive voltage from the pass tubes goes through the mA meter then selected resistors to the top of the voltage adjust potentiometer. A negative voltage goes through other selected resistors to the bottom of the voltage adjust potentiometer. The wiper on the potentiometer should therefore have a voltage that varies from some negative voltage to some positive voltage. This voltage is connected to G1 of T13 and the anode should vary more and control the pass tubes and the output voltage. The cathode of T13 is the reference voltage that is compared to the output sampled voltage that goes to G1 and the G1 voltage should be close to the cathode voltage no matter what the position of the potentiometer wiper is. It is the Anode voltage that should vary from the lowest marked voltage on the range switch to the highest marked voltage for that range.

If the meter is okay and some of the resistors connected between it and the top of the potentiometer are open, there will be no positive voltage on the top of the potentiometer and there will only be a negative voltage on the wiper that comes from the negative supply. There should not be only a high negative voltage going to T13’s G1 if everything is okay. A high negative voltage compared to the cathode voltage will force T13 to turn off and the output to remain at zero. This is the area you should be checking to see if you get the expected changes. The details I’ve posted here without having the power supply in front of me may not be totally correct but the principles should be sound.

[trobbins]

It's worthwhile confirming the voltage across C6 (should be circa 500V), and with the -360V at the top of the regulator valve 17, that will confirm regulator valve is operating and at about 140V across it.  The regulator current appears to be circa 360V/30k + 420V/402k + 60V/3M3 = 12 + 1 = 13mA, which is fine.

Do you have a spare valve to swap in for T16 ? 

T16 should be conducting current to pull its cathode voltage up from -360V, and hence increasing the anode-cathode voltage of T15 which would raise the voltage on the bottom of the R44-49 string and hence the grid voltage of T13, in sympathy with T13 cathode voltage rising from -360V.  That should then bring T13 out from saturation and allow the pass valve grid voltages to rise out of deep cut-off and start raising the output terminal voltage.

That assumes all resistors are ok, with nominal values, and the mA meter is ok with a fairly low resistance (checking that may need some caution, as putting a DVM directly across it may be enough to pin the needle to an end stop, so perhaps measure the meter resistance along with one of the resistors on either side of it).  Also it assumes that none of the regulating circuitry caps are leaky or shorting - can you measure all those caps for capacitance and do you have any kind of leakage tool (eg. a 100VDC supply and use a DVM voltmeter to connect the cap to the supply - if the cap has at least a 100VDC rating).

How old is the equipment.  The electrolytics may well need replacing.

[scsi]

Did a lot of checking. Things I can verify that are OK and working as expected:

1. mA - is OK
2. Every resistor R33 - R44 is OK (I am measuring resistance through mA and varying selector swich - getting all the results as expected)
3. Potentiometer R32 not open
4. Selector switch is OK

Things are not OK:

Measuring voltages of T13 with respect to cathode (Common on T13 cathode)

Anode to Cathode -     7.4V
G1 to Cathode -              0V (If I rotate R32 potentiometer nothing changes)
Screen to Cathode -   14.8V

Voltage on potentiometer R32 between outer pin and center: varies (if rotated) from 0 to 13V

can you measure all those caps for capacitance and do you have any kind of leakage tool (eg. a 100VDC supply and use a DVM voltmeter to connect the cap to the supply - if the cap has at least a 100VDC rating).

How old is the equipment.  The electrolytics may well need replacing.

I measured all the capacitors, tested ESR and Leakage up to 600V - everything is in spec. Theres is not a single electrolytic in this equipment everything is paper in oil sealed military grade capacitors and one silver Mica.

It's worthwhile confirming the voltage across C6 (should be circa 500V)

Voltage across C6 - 530V

after I done voltage measurments resistor R19 (100 ohm) poped. That can't be good

[trobbins]

R19-R26 appear to just have a current sharing function between the pass valves.  I can't think of a simple scenario for R19 to fail - assuming it is 1W, it would need 100mA continuous just to get to its dissipation rating - which only leaves C4, C5 or an unknown shorting path (eg. around the voltmeter).

So the reference valve is circa 170V type.

T13 could be positive biased with some input grid current flowing (measuring Vgk may confirm that), and depending on the valve model that could allow Vak to be close to 0V (ie. 7V).  So imho its uncertain whether T13 is ok or not.

I'm guessing you don't know what the valve models/types are?  Nor do you have spares?  Can you tell if T5, 13, 15, 16, 19, 20 are the same, or some of them are the same?  If T19 or 20 are the same as T16, and you can confirm that the lower regulator works ok, then you could do some tube swapping.

[scsi]

unknown shorting path (eg. around the voltmeter).

When it poped two aligator clips were placed on C6 to measure voltage.

I'm guessing you don't know what the valve models/types are?  Nor do you have spares?  Can you tell if T5, 13, 15, 16, 19, 20 are the same, or some of them are the same?  If T19 or 20 are the same as T16, and you can confirm that the lower regulator works ok, then you could do some tube swapping.

I do know the valve type:
T13,16,20 - the same 12J1L
T15,19 - the same 4P1L

What information would be useful? Anode characteristics? I could try to find it.

[scsi]

some tube swapping.

swaped T13 tube from working 0-400V channel and I have some action! Positive voltage on 20-600V output! Still need some testing just letting you know.

OMG IT'S WORKING! Such a simple thing. You guys are supermans in my eyes. Thank you!!

Could it be that something is damaging the tube in circuit?

Yet I have few more questions: is it possible to change R32 (6.9k) and R68 (40k) weird value potentiometers (because they are old, rusty and scratchy and wirewound) to different say 10k and 47k and tune R30 and R66 appropriately or it is not going to work?

[trobbins]

Tubes die for many and varied reasons.  Perhaps try to acquire as much info now about the tube types, and if there are any equivalents - that may come in handy in the future!

I'd recommend you get a spray can of electrical/electronic contact cleaner/lubricant (with a hollow plastic squirt pipe),and use it on the pot wiper and ww contact surfaces.  It is good experience to try and restore parts like pots, even if it requires them to be carefully taken apart and then cinched back together.  Of course some pots can get beyond repair, but if the rust is just aesthetic, and the scratchiness subsides, then you can probably live with it being old.

The switch contacts may also welcome a clean.  If you do any more equipment servicing, the spray can will likely come in handy - many an instrument or audio gear or ... can be brought back to full function.

I'd still strongly recommend you take the time to add a discharge resistor across the raw caps for future safety, and retrofit ss diodes in series with the valve rectifier anodes as a form of insurance for when the valve diodes die.

[scsi]

Working perfectly. Tubes are just megical for me.

Thanks to all!

[schmitt trigger]

Beautiful photo!