Valve amplifier service

[simon]

They've already been changed, change them again?

[Nick]

If replacing them the first time did nothing, then no.

[simon]

I'll not bother then

[Mike H]

Question: what is 'ANPIO'?

Poking R7 makes horrible noises is saying to me RF oscillation. But I'm a bit surprised as I wouldn't have thought an ECC83 could do that (ECC81, or 82 even, yes).

I would want to put grid stoppers on both pin 2 and pin 7. Pin 7 being grounded via the '0.22u 400V' (can't read the 'C?' number) might be exacerbating the problem. You have a grounded grid amplifier with wide bandwidth (assuming the cap has good HF performance).

Or it could just be 'dirty' socket contacts.

I assume you've done the swap valves around to see whether the noise moves to the other channel thing?

[Cressy Snr]

Question: what is 'ANPIO'?

Probably Audio Note Paper In Oil.

[Mike H]

Looking at the photo again, you could move the 1 meg R7 to the tag board, so that pin 2 has 820R and a wire in series, and pin 7 has an 820R and a wire in series. If that makes sense.

[Mike H]

Question: what is 'ANPIO'?

Probably Audio Note Paper In Oil.

Right, which I was kind of afraid of, maybe it's got inductance or too much ESR or something. So Simon could try bridging it with something like a high Voltage ceramic disc cap, if he's got one(!), see if that does anything.

Incidentally studying the diagram more, discrepancy of anode Voltages V2A vs V2B, both L & R V2B anodes are higher than V2A, and not equal, this is suggesting C4 (it looks like C4 if I zoom in) is leaky, and the left channel C4 is more leaky than the right.

EDIT: admittedly because R7 is 1 meg, but even so.....

[Nick]

Yep, that PIO would never be my choice for C4, but the voltages show that both r7's are dropping about 10v, so that's 10uA about equal both side. But I agree a polyprop would be the normal choice there. But in general its just guess work without a scope showing us if its LF or RF that's happening.

[simon]

Yes Audio Note PIO. Great big mahoosive things they are that sat right on top of the tag strip. I took them all out when I replaced all the caps early on, and have left them out for access. C4 is now a LCR PP, and C5 and C6 are orange drops.

Yep, I've switched the valves, Dave did a full revalve before I started looking at it too. I've cleaned the sockets with contact cleaner also.

I'll try 820R on pin 2 too.

What I can't quite get my head around is what's changed. It didn't hum before so why would it need grid stoppers? I've replaced the caps and that didn't make a difference. Resistors perhaps?

Okay, I'll investigate with the scope (I was waiting for someone to say the magic S word ). I'm not sure how reliable it will be given my "experiences" with the C3g phono (which I still need to fettle) but I'll see what happens.

Thanks for the help, really appreciate not being alone here.

[Nick]

I find it hard to imaging a ecc83 needing grid stoppers TBH.

[Morgan Jones]

Post #17 said: "What I did notice was the small amount of hum stopped when I put the meter probe on the left channel grid of V2B."

That shrieks RF oscillation to me, and RF oscillation can cause hum. I doubt if the phase splitter is oscillating, but the cathode follower in that SRPP is a definite possibility. Agreed, ECC83 is usually too lazy to oscillate, but usually isn't the same as never. A 10k grid stopper on the upper grid of that input SRPP might cure the problem. Cathode followers sometimes need cathode stoppers too. A 1k in series with the cathode's lead to the phase splitter is another possibility. A third possibility is that the power supply isn't adequately decoupled at RF to that stage. A 100nF to chassis might work.

I'm horrified to see the comment that modern PIO capacitors only last a few years. Are they really that bad? I wouldn't dream of using them anyway, but that's appalling if true. I would expect a nice polypropylene (such as an LCR) to last for decades unless abused.

[Nick]

I'm horrified to see the comment that modern PIO capacitors only last a few years. Are they really that bad?

In general I don't think they are, but Audio Note definitely had a problem at some time in the past with getting conductive. Don't know if its still the case, it may be the result of a manufacturing problem that's now fixed, or may still happen given enough time. I can remember at least three times where I had a repair where a AN PIO was driving a large valve resistor (say 1M area) and developing a voltage across that resistor.

[Nick]

That shrieks RF oscillation to me, and RF oscillation can cause hum. I doubt if the phase splitter is oscillating, but the cathode follower in that SRPP is a definite possibility

Maybe, but I would then expect the V2A grid to fix the problem, not

stopped when I put the meter probe on the left channel grid of V2B

Which has a 1M resistor and a cap to ground between the CF and V2B grid.

But for me, the first step is get a dual channel scope on one of the output valves grid when its "humming", and see if there is RF on that point (and depending on which one it is), and then see how the level of RF changes as you put the other probe elsewhere. Decoupling the rail is a good idea though, may be interaction from the output stage pentodes which could sing to the earlier stage.

[simon]

Thanks both, I'll try to investigate further tomorrow

[simon]

I've had a look with the scope but I'm not sure I should be allowed to really. I struggled to get two probes to be useful - the traces were so erratic it wasn't any use.

BTW the second channel was in V DC and I couldn't change it to V AC. Is that just the way they are? It's a Tektronix TAS 485 200MHz.

So what I did find:

I couldn't detect anything on the speaker terminal, other then the low level mush I had with the C3g phono.

Both 5881 anodes had a nice trace, perhaps everso slightly asymmetric. Perhaps this is what PP does?
.

20230831_171200.jpg (211.57 KiB) Viewed 11397 times

.
Putting the probe on the second ECC83 grid didn't make any difference.

The 5881 grid was a bit more saw tooth, with a small spike at the bottom of the wave. This is the best photo I could get.
.

20230831_171118.jpg (160.74 KiB) Viewed 11397 times

.
Putting the meter probe on the second ECC83 grid (probably) appeared to remove the small spike as far as I could tell. Again it was difficult to get a good photo.
.

20230831_173157.jpg (173.75 KiB) Viewed 11397 times

.
Does this mean anything to anyone? Not sure where that little spike is coming from, but the scope doesn't seem to be picking up any obvious nasty RF???