Marconi Valve Voltmeter TF1041C - repair thread

[turretslug]

Not saying that it's definitely the cause, but with circuits like this, it's worth checking resistor values to see that they aren't outrageously out of spec.- old carbon composition resistors are prone to drifting well off value (notoriously the types in a white ceramic tube), particularly if they have a high potential across them as most of the ones in this circuit do.

The OA2 seems to be behaving about right, the 12BH7 could be very tired and upsetting the circuit action- but the resistors are easy to check and are at least some way up on the suspects list,

Colin

[Cathode Ray]

Thanks Colin,
yes I did check the resistors and found them to be fairly close.
The 150k read 175k, it's a silver band resistor.
the 470k read about 485k.

So not enough to cause an issue I think.

My next step is to see if I can measure the regulated current being drawn at 220v.

[turretslug]

Sorry, I'd read post #19, but missed the resistor check in #17!

[Cathode Ray]

No worries Colin.

So with current measured at 11mA here is what I have.

Output voltage is 220v when loaded at 11mA
It rises to 360v with no load.

Adjustment control has little effect, voltage o/p remains between 213v and 218V, target is 230v according to the manual.

Input voltage is 276v when loaded, rising to 500v when not loaded.

The 12BH7 gets fairly warm when in operation, the glass surface reaches 85 degrees C with the 11mA load.

What say the jury?

[buggies]

Just some thoughts...
I have not compared the data but do you have a 12AU7 (ECC82) you could try for a test? You could use a variable resistor load for say 5mA to 11mA instead of the rest of the device.
Are the reservoir caps up to scratch? Ripple voltage on the 276V?

[Cathode Ray]

Reservoir caps were changed out earlier tonight, the original ones measured OK but I soldered some new ones in just in case, but no change.

I have an ECC85 but its probably not a match.

I can source cheaper 12BH7A valves, these seem to be a direct drop in, but with faster heater warm up times??

[buggies]

I would pass on the ECC85 anyway

EDIT: - Just deleted a silly statement - too late for me - brain will resume normal service in the morning...

[Cathode Ray]

There is one other thing, there is a HT- of 120v and I didn't disconnect it when measuring the non-loaded HT+
So perhaps I should revisit this and disconnect both HT lines before passing judgement.

[Cathode Ray]

Ok so some interesting results.
When I disconnected both HT+ and HT- from the meter I got mad readings wrt ground, bouncing around 50v and zero v
But then I realised there was no reference to the common ground any more.
So I measured across from HT- to HT+ and read 350V which is exactly -120 +230V, this implies the regulator is good but I was referencing the common ground.

However when reconnecting the HT lines, the HT+ plus HT- reading fell back to about 340V , which is 10V shy of the target.

I also measured HT- current draw at 12mA (HT+ was connected to the meter)

Now I wondering if I'm chasing my tail, the meter works very well on all AC ranges, all Ohms ranges and only has a 600mV offset on the lowest DC ranges.

I have a mate who can test with an ECC82 on Sunday morning, that should be the test.

[Cathode Ray]

Having thought about , the 360v I measured when HT+ was disconnected was actually the potential across HT+ and HT- , bit it was 10v above target because I had the trim pot turned up.

So the valve regulates perfectly well off load and drops 5v when in load.

I'm thinking this isn't a big issue. Rabbit down a hole and all that !

[buggies]

I thijnk you are right. The confusing part (for me) was the lack of a stable ground reference point - depends on the rest of the circuit. The regulator does not set the 230 volts it controls the total 350.
I did a very rough simulation with "ideal" components in an old student version of Electronic Workbench (I am not up to driving spice).
Attached shows voltage and current values with two different loads - the preset left in central position.

[Cathode Ray]

That looks exactly right George, many thanks for taking the time to do that for me. Much appreciated.
The key for me was realising that the regulator set the combined 350v. Your results look very close to real world readings, nicely done.

So now learned a lot about valve regulators but I'm back to the main meter to see where this 600mV DC offset is coming from.

The offset seems to be arising at the DC probe tip, at least I can measure it there.
So the actual bridge circuit is reporting correctly, its showing the offset is present.
Possible causes I'm considering for this are:

• A resistor gone out of tolerance.
• A grounding issue.
• An imbalance in the valves which were once matched using two SIC (R64 and R65) resistors at the Marconi factory, I think that means Select In Circuit, we used to call them SOT Select On Test.
• An issue with the replacement Thermistor & 100 Ohm resistor which is in series with the meter movement, perhaps it necessary to fine tune the SIC resistors once this is replaced.

BTW, The version of TF1041C I have is a bit different to the diagram provided by wpm. I have the correct diagram but it won't attach here, too big I expect.
I have a link to it here;
https://lh3.googleusercontent.com/lr...w2308-h1299-no

I can tack high value resistors across the SIC resistors and see what effect this has, I can also try the same with the Thermistor. This might lead to something.

[turretslug]

Originally Posted by Cathode Ray

Having thought about , the 360v I measured when HT+ was disconnected was actually the potential across HT+ and HT- , bit it was 10v above target because I had the trim pot turned up.

So the valve regulates perfectly well off load and drops 5v when in load.

I'm thinking this isn't a big issue. Rabbit down a hole and all that !

No sweat Ray- surely many of us have been caught out by the "what is with reference to what" question when measuring/checking multi-rail PSU stuff- I know I have! Particularly with valve stuff, where there's incentive to save active devices and cunning tricks can get played to keep heat and expenditure down. Some more complex valve oscilloscopes have lots of plus and minus rails, referenced to each other in various ways to minimise DC-coupling drift and not unusually involving the EHT section too. Perhaps all supplies taking a cue from a single reference device- it makes me think of half-a-dozen plate spinners all standing on the shoulders of a one-legged man who's standing on a roller skate.....,

Colin

[buggies]

I can't follow the operation of the rest of the circuit - too many switch positions and ranges/functions.
You said a few posts back-

Originally Posted by Cathode Ray

Ok progress made, offset was reduced by about 50% by replacing a 0.03uf disc cap, ESR was 85 Ohms.

Puzzling that replacing a cap should reduce a dc offset? Which cap? Is it really ac noise or something?
I would try swapping the two similar valves over to see if any change and also look for any high impedance leakage paths. I have had trouble in the past with leaky pcb material and board contamination in high impedance DVM input circuits.

[Cathode Ray]

Hi George and Colin,
Thanks for the replies.

The 50% reduction seems to have been an error on my part.
After more study I can see the offset rises gradually over 20 to 30 minutes of operation.

From cold the offset is about 50% of it's final value , that's the mistake I made . Not waiting for the reading to settle...

I try swapping the valves, good suggestion.

Cheers
Ray

[turretslug]

Originally Posted by buggies

Puzzling that replacing a cap should reduce a dc offset? Which cap? Is it really ac noise or something?
I would try swapping the two similar valves over to see if any change and also look for any high impedance leakage paths. I have had trouble in the past with leaky pcb material and board contamination in high impedance DVM input circuits.

I've found that early ceramic discs can go very leaky sometimes, particularly low voltage types- I don't know if they absorb water vapour or chemically decompose in some way but they can certainly degrade. I once had a bag of NOS 100nF 30V discs sitting around for a while, having encountered the problem quite often in '60s/'70s kit I thought I'd check them on a (1.5V!) ohm-meter- every one was useless.

I also spent a long time distrusting the tiny multi-layer ceramics as I often found that one amongst maybe dozens of parallelled specimens on logic or signal cards would go very low resistance short, but as they abound in reliable stuff now and have done for two or three decades, I can only assume that they have improved somewhat.

[Cathode Ray]

Hi Turtle slug
I think ive replaced all the caps at this stage. The meter it has about 6 caps to begin with.
The source of the offset is very high impedance.

My 10 megohm DMM knocks it back by 90%.
Actually there is a direct correlation with the DMM reading and the actual offset.
The DMM reads 6mv at the prove tip, one tenth of what the valve meter shows.
Considering the valve meter has a 100 megohm input impedance, it's easy to see why the DMM reads 1/10th .. it's acting as a divider.

This suggests that the bridge is not in balance , but it must be adjustable somehow.

More digging needed I think...

[Cathode Ray]

Quick update on the valve swap... the offset is 600mV with valves in their original slots but only about 60mV with valves reversed... so what does that tell me?

Valves are 6BS7's

[buggies]

At that stage I would be reaching into the stock cupboard for half dozen valves to select a good un.
Sticking my neck out - my guess would be grid current. The 6BS7 brought out the grid to top cap and Brimar advertised it for electrometer use where high resistance grid resistors used. Any grid current would produce an offset voltage.
The data sheet has a graph at the end showing the expected percentage of valves against grid current. If you can work out from your circuit what the grid resistance is you could calculate the grid current required to give the 60/600mV offset.
https://frank.pocnet.net/sheets/184/6/6BS7.pdf
Can you live with it - does it affect normal readings?

[Cathode Ray]

Hi George,
If I understand the graph correctly, Marconi probably selected matched valves where the grid leakage was similar in each ?

If that's the case I'm a bit snookered, I've no chance to select from a cupboard of valves !

The meter will read voltages provide the source impedance is considerably less than 100 megohms, otherwise I expect the offset will creap into the reading and give incorrect results.

I could just chance buying a 6BS7 on the basis that at least one valve has gone well out of spec to cause the offset.
What you think ?