Mike express build

[Miket]

From your sound clip, I would guess that your amp is suffering from ultrasonic oscillation, which largely boils down to lead dress, lead dress and lead dress.

My lead dress can be seen in page 1 of this thread. I have gone through the lead dress with a fine tooth comb. I may be wrong, but I think it's ok.

[Tillydog]

I think 300mV is still way too 'hot' a signal to diagnose this - The PI voltages and clipping you are seeing are about those predicted by LTSpice with 300mV in & gain at 50%. Clipping starts at ~3mV in with gain at 50%, give or take. It starts at the output stage / PI. I would knock the signal down again (maybe 5-10mV??) and reduce the gain to get everything clean, then see what happens as you turn the gain up.

I don't know why your grid resistors should be running hot, especially if one is very hot. I consider that to be significant - does the hot resistor move if you swap V4 V5?

From your photos, I would have said that the proximity of the DC blocking cap on the input jack to the bright caps was a dead-cert problem, but I note that you have taken these "out of the circuit" - have you physically removed them, or just disconnected one end? (In which case they could still be the cause of the problem.) - How does it look now that you've removed the DC blocking caps, etc. It is very finicky.

A small grid stopper on V2 made a big difference on my build, so it is definitely worth a try.

[Miket]

One thing you may want to try is reducing the value of R7. I changed mine to 67.9k and it cleaned up the abnormal phasing sound but still left plenty of gain. YMMV.

II tried 100k and 56k in r7. Neither made any difference.

Also added 100k grid stopper to v2a and phase inverter. No difference there. (value too low? I know you see 470k on some amps)

I removed the nfb and audio probed around for the origin on the swirl. Seems its being made by the power tubes, as the only place it appears is at the ot secondaries.

[Miket]

I think 300mV is still way too 'hot' a signal to diagnose this - The PI voltages and clipping you are seeing are about those predicted by LTSpice with 300mV in & gain at 50%. Clipping starts at ~3mV in with gain at 50%, give or take. It starts at the output stage / PI. I would knock the signal down again (maybe 5-10mV??) and reduce the gain to get everything clean, then see what happens as you turn the gain up.

I don't know why your grid resistors should be running hot, especially if one is very hot. I consider that to be significant - does the hot resistor move if you swap V4 V5?

From your photos, I would have said that the proximity of the DC blocking cap on the input jack to the bright caps was a dead-cert problem, but I note that you have taken these "out of the circuit" - have you physically removed them, or just disconnected one end? (In which case they could still be the cause of the problem.) - How does it look now that you've removed the DC blocking caps, etc. It is very finicky.

A small grid stopper on V2 made a big difference on my build, so it is definitely worth a try.

Tilly,
Thanks for the scope tips, this is the first time I've used one on an amp, so I'm still working out how to do it properly.

I was wrong when I said it was the v4 grid resistors getting hot, I ment screen resistors. I measured up to 44v dropping across it under load- which means it would be dissipating around 2w at that point. Within range I guess?

The bright switch caps are removed completely from the circuit. As are all the DC blocking caps related to the vvr, and the vvr itself.

I'll do some more scoping tomorrow, and update pics of the guts.

[Tillydog]

I noticed you have the connections to the power stage grid stoppers looped through the connection from V3 pin 6 - try shortening the pin 6 wire and run it clear of the grid connections like the Francesca photo attached. Also, push the blue and yellow wires close together and tight against the chassis where they run from the perf board to to V 4 & 5. Try to make sure that you have as much clearance as possible between the OT primary wires and V4/V5 grid connections. No promises, but something to try.

Just to eliminate something else, could you confirm that with T, M & B all on zero that it kills the signal? (Gain still on 50%l, say.)

[hitchcaster]

maybe you got the schematics mixed up and you built a synth by accident? or a mutron octaver?

honestly I'd follow kens lead dress and design EXACTLY.... then from there you have a baseline and you can try and outsmart him and add things or change lead dress etc.... you'll be able to hear changes as you go, and then probably end up putting it back to kens way 99% of The time... he did alot of listening I think

[Miket]

Just to eliminate something else, could you confirm that with T, M & B all on zero that it kills the signal? (Gain still on 50%l, say.)

If t,m and b are all on 0, gain 50% I get a clean yet much quieter signal, if I max the gain pot, I get almost no signal, with gain pot at 0, no signal

[Miket]

I bypassed the first half of the preamp by attaching the input directly to the grid of v2a. And physically removed v1.
The gain was less, but the phaser swirl effect was still there. I then boosted the signal with an opamp boost pedal, and he amp was back to its square wave synth self. Even with a low gain ecc81 in the pi.

I've attached current photos. Please excuse the temporary installation of the grid stoppers.

[Miket]

I just did some more scoping, with the lowest voltage signal I could. The clipping order is as designed, power tubes>pi>v2

The power tubes or output is a square wave before the pi starts to clip. The strange thing are these spikes that appear. I'll attach a photo, note the photos was taken with gain at 3 o'clock.
After taking the photo, I connected my guitar and hit just the bottom e string, as the note rang at 440hz the readout on the scope was more or less the same (8ohm tap connected to dummy load)

Also, under these conditions, using the amp with a dummy load as apposed to a speaker. Is it normal for the ot to humm loudly at the same frequency its outputting?

[Miket]

Double post

[Tillydog]

If t,m and b are all on 0, gain 50% I get a clean yet much quieter signal, if I max the gain pot, I get almost no signal, with gain pot at 0, no signal

OK, that proves that the tonestack is correctly grounded, but that there is some coupling beween the 1st stage and the rest of the amp. I'm not sure if this is significant at the moment, so putting it on the back burner for now, because it seems you still have the same problem with this stage bypassed...

I haven't had a chance to look at your pictures properly, but I will. (I noted that you sorted out the connection from the PI )

Oh, and I noticed before that it looks like your NFB connection is onto the last tap on the impedance selector switch which is usually the 4 or 16 ohm tap, rather than the 8 ohm - Did I see right? Any reason why?

I don't think it's unusual for the OT to make a little noise, but I wouldn't have described it as 'loud' - it might be that the clamp bolts need tightening a little.

What OT are you using? (Is it the Marstran 3801?)

What are you using for a dummy load?

Do you get the phaser sound on a clean signal, or only distorted?

Sorry - no answers, only more questions :-/

[Miket]

In the photos the nfb is not connected. When it is connected, it goes to the 8 ohm tap.

Yes I'm using that marstran transformer, and the dummy load is the air brake, then into an 8ohm 50w resistor bank.

The phaser sound is still there on clean signal, but much more noticeable as the gain increases. Seems to coincide with the sharp peaks and dips on the last scope pic.

[Tillydog]

In the photos the nfb is not connected. When it is connected, it goes to the 8 ohm tap.

I meant in the earlier photos - Do you not have the taps arranged 4-8-16?

I've seen peaks like you have in the scope photo, and they sound bad. It suggests to me that the load you're driving may be inductive, or mismatched or maybe it is something to do with driving the amp so hard - I think it is being driven much harder than it would ordinarily see with a guitar signal.

With the nfb disconnected, it will take less signal to fully drive the amp. A square wave will sound like a synth.

If nothing shows up, you could probably kill the peaks with a carefully chosen Zobel network (a resistor & capacitor in series) across the OT secondary. I've found that trial and error yields quicker results for these than calculation, and the values are quite critical. I guess you's be looking ar R of about 47 ohm (try values between 20 - 80 ohm, say) and C of about 2.2uF (try values between 470n - 4u7, say). I can't help thinking that there is something else wrong, though.

[Miket]

I haven't tried the zobel network yet, as I feel that I would rather find the cause than deal with its effects.

Today I swapped in another output transformer, replaced the bias caps, tried 6v6's, replaced power power supply filter caps, cooled the pi bias, re flowed every soldier joint in the output section, changed the power tube grid resistors to 3k9(closer to Marshall value), increased C12 to 500pf.

And NOTHING helped!

This amp is driving me crazy!

[Colossal]

In the photo above, your second speaker jack is not grounded...unless that photo is old and you already took care of that.