Mystery amplifier

[Kokotoni Will]

Originally Posted by trobbins

It would seem that current is reforming the first 8uF e-cap

Do you mean PSU cap C2? That has 381V across it, and is paper-in-oil. There aren't any 8uf caps in the amp itself. HT on the amp is 178V after the 10k load, and just over 200V at the PSU.

Originally Posted by trobbins

This test setup can also measure leakage current through the coupling caps C10, C11 and C2 in marked up schematic in post #141, by measuring the Vdc across the resistor connecting each coupling cap to 0V.

Forgive me if I'm missing something obvious, but I can't see any one resistor that connects any of these caps to 0V. C10 and C11 are 450V filter caps.

Originally Posted by PJL

When you first applied HT without the resistor the capacitors were forcibly reformed otherwise they would have overheated and leaked. You could look for evidence of swelling/leaking on the caps as they might have been damaged. Assuming all is well, what is left is likely residual leakage current and is unlikely to do any damage, 2.5mA @200V is just 0.5W.

There's no visible damage to the filter caps, although I expect there wouldn't be even if they were damaged, as they're cans with electrolytics inside.

[PJL]

I see you have disconnected the earth and 1K resistor from the secondary which I missed from my instructions.

For push-pull transformers the correct measurement for turns ratio is primary 'anode to anode' to secondary. Attached is the A2134 datasheet which shows how the optimal load impedance varies from 3-7.5K dependant on operating conditions (particularly the HT) none of which match our example with UL taps and low anode current.

With 250V HT I would aim for about 6K. The required turns ratio for various speaker impedances is then:
4 ohms = square root(6000/4) = 39
8 ohms = square root(6000/8) = 27
16 ohms = square root(6000/16) = 19

Remove valves and disconnect HT. Confirm anodes are primary 3 and 4.
I am a little concerned about unwanted capacitive coupling in the transformer so put a 1K resistor load on the first secondary (AB). Then put your 6.3V on 3 and 4 of the primary and measure the AC voltage on the primary 34 then measure the first secondary (AB). Move the 1K resistor load to the other secondary (CD) then measure the AC voltage on the primary 34 (again) and measure the AC voltage on this.

[trobbins]

I was referencing the updated schematic by PJL in post #141. Should we be using a different circuit schematic to follow along?

[Kokotoni Will]

Originally Posted by PJL

I see you have disconnected the earth and 1K resistor from the secondary which I missed from my instructions.

For push-pull transformers the correct measurement for turns ratio is primary 'anode to anode' to secondary. Attached is the A2134 datasheet which shows how the optimal load impedance varies from 3-7.5K dependant on operating conditions (particularly the HT) none of which match our example with UL taps and low anode current.

With 250V HT I would aim for about 6K. The required turns ratio for various speaker impedances is then:
4 ohms = square root(6000/4) = 39
8 ohms = square root(6000/8) = 27
16 ohms = square root(6000/16) = 19

Remove valves and disconnect HT. Confirm anodes are primary 3 and 4.
I am a little concerned about unwanted capacitive coupling in the transformer so put a 1K resistor load on the first secondary (AB). Then put your 6.3V on 3 and 4 of the primary and measure the AC voltage on the primary 34 then measure the first secondary (AB). Move the 1K resistor load to the other secondary (CD) then measure the AC voltage on the primary 34 (again) and measure the AC voltage on this.

Primary terminals 3 and 4 do go to the power valve anodes.

I managed to break my other meter, i.e. the one that gave correct LT readings, so had to make do with the one with the 500k resistor across AC. With LT strapped across 3 and 4, my multimeter measures 5.8V. Slightly lower than 6.3V, but as PJL said this is likely because of that 500k resistor.

I measured AC voltage across secondary terminals AB, but got 0V, with and without a 1K resistor across AB. Same with CD.

[Kokotoni Will]

Originally Posted by trobbins

I was referencing the updated schematic by PJL in post #141. Should we be using a different circuit schematic to follow along?

If you could refer to this one, that would probably help avoid confusion:

https://harnessontheleftnape.blogspot.com/2025/11/mystery-amp-schematic.html

[PJL]

The secondary connections depend on the choice of left/right transformer but you can confirm by checking continuity of A to B and C to D. It's also possible your meter may not be able to measure low AC voltages.

[Kokotoni Will]

There is continuity between A and B, and there is continuity between C and D.

With LT across primary terminals 1-2, the meter measures 0.4V AC across secondary terminals A-B. Again, less than the 0.7V AC measured with the other meter. I'll have a look at getting a better meter.

[Kokotoni Will]

My new valves have arrived. The guy couldn't find the valves I'd ordered, so I ordered a different pair, described as giving "similar emissions". I've just tested the amp with them plugged in. Their cathode voltages are 11.78V and 12.5V DC. Not overly well matched.

This time both speakers made a hum that got louder as the valves warmed up - I didn't notice any sound last time. HT fell from 232V with no load to 211V with the amp plugged in.

I've ordered a Fluke 25 which should arrive in a few days.

[Diabolical Artificer]

I've been trying to follow this far, a few observations. 200 to 250v HT is tad low to get the best out this amp. For a PP amp you want 300v at least. I'd get the amp mostly going & not worry too much about getting everything perfect. One you have an OP you can then tweak. Still, we all go about things differently.

Their cathode voltages are 11.78V and 12.5V DC. Not overly well matched.

it really doesn't matter. In terms of THD a 5 or 10% difference is insignificant.

I've ordered a Fluke 25 which should arrive in a few days.

Well done, you won't regret it, they are really hard to kill. My two have survived significant stupidity on my part. Check the fuses when you get yours, pop in a new PP9 & it should last you years.

It's a shame you don't have a scope. There are free PC ones which are good & with a cheap soundcard & a few resistors it would will enable to test amps better going forward.

Well chuffed to see you get this far, Andy.

[PJL]

3.8%, that match is very close.

You put the links and connections to the speaker terminals back exactly as they were originally?

You are running without the lamp limiter and all valves in?

Once you have the new meter you need to complete the output transformer test so we can work out how to wire it for your 4 ohm speakers.

Be careful measuring current with your Fluke as the fuses are very expensive.

[Kokotoni Will]

Thanks Andy!

Originally Posted by PJL

You put the links and connections to the speaker terminals back exactly as they were originally?

Not exactly, I've used alligator leads to save having to re-solder (see pic)

[Diabolical Artificer]

Posts crossed with PJL -

Be careful measuring current with your Fluke as the fuses are very expensive.

There are 3 fuses in the Fluke 25 - two big HRC types but the one mostly used is the 630mA one when set to mA/uA. I've used a fast blo 1A in mine for years without issue. You still have to careful though when the meters set to A/mA as it's dead short.

Andy.

[Kokotoni Will]

Originally Posted by Diabolical Artificer

It's a shame you don't have a scope. There are free PC ones which are good & with a cheap soundcard & a few resistors it would will enable to test amps better going forward.

I've got a FNIRSI 5012H.

[Kokotoni Will]

I've re-fitted the valves and the hum is much quieter now.

[PJL]

I also use those patch leads and they are not great so solder it all back together.

Switch on and adjust the HT slowly up to 250V. If the hum gets unbearable switch off as we will need dummy loads (10W power resistors) of say 4-16 ohms.

Measure all 4 cathode voltages again.
Measure anode voltages CV4006 pin 6, ECC83 pin 6 and pin 1.

Only use the FNIRSI 5012H with the 10X probe as >80V will damage it but it should work well for use with the amp.

[Kokotoni Will]

Cathode voltages V3 - V6:

15.2, 14.2, 14.4, 14.6V DC

V1 anode: 38V
V2 anode 1: 150V
V2 anode 2: 151V
V7 anode 1: 84V
V7 anode 2: 88V
V8 anode: 21V

Hum got reasonably loud when valves warmed up, but quickly quietened down.

[PJL]

And the anode voltages?

[Kokotoni Will]

Originally Posted by PJL

And the anode voltages?

I've just edited my post.

[Kokotoni Will]

Looks like voltages on the right channel are about 50% those on the left. I've taken voltages across the filter caps, and it's the same story:

C10: 92V
C11: 192V
C12/13: 250V
C23: 106V
C24: 50V
C25/26: 250V

I've noticed that, as well as humming quietly, the right channel picks up interference from mobile phones, and pops when the unit is powered off, whereas the left channel just hisses quietly.

[PJL]

V8 anode voltage is too low and that sets the bias conditions for V7 and that affects all the other voltages.

Swap the CV4006's over and see if the anode voltages change.