DIY 1959 with horrible ... distortion? oscillation? not sure WHAT it is!

[Gtr0]

Hi.

I'm back with a terrible noise that I just can not seem to find.
I just finished an amp this morning. Powered on w/out load, took some measurements, all fine. Added tubes, took measurements... good. Biased, no problem.

But this is what I got on first play.

1959_noise.mp3

Yikes!! Reversed output tranny primaries, disconnected NFB, swapped pre tubes, finally swapped power tubes... trying the amp everytime without change.

I traced the wiring several times but I can't find anything - doesn't mean I didn't make a mistake, obviously. The sound is so crazy though perhaps someone will know right away from listening and can advise? Thanks.

[Chuck H]

Could be an oscillation due to lead dress. It's not "swirling" so I don't think it's crossover distortion. Can you provide a high res pic of the innards? Maybe we can pic something out visibly. If you have a scope you could put a high input signal through the front end (maybe 200mV or so) to induce clipping and trace the signal to see if and where it shows up in the chain.

And, FWIW, I think the basic clipping character of the amp is pretty damned good Just get rid of that background buzz and I think you have a winner.

[Justin Thomas]

Agree with oscillation. But, isn't a 1959 a non-Master amp? And if you're getting THAT sound out of it, either you play LOUD, you modded it, or you're slamming the front end/guitar is superhuman.

Justin & Jusrin

[Chuck H]

This^^^^^^^^^^^^^^^^^^

I assumed it was cranked. It wouldn't be fair to plug in a dirt box for an amp diagnosis. Possible channel stack? Master circuit? Which would definitely up the ante on taking care with the lead dress. I REALLY HATE IT when a design is presented for diagnosis (1959 in this case) when the actual amp is something very different. And the 1959 design doesn't have a channel stack or a master volume. So...

[Gtr0]

Agree with oscillation. But, isn't a 1959 a non-Master amp? And if you're getting THAT sound out of it, either you play LOUD, you modded it, or you're slamming the front end/guitar is superhuman.

Justin & Jusrin

Cranked

This^^^^^^^^^^^^^^^^^^

I assumed it was cranked. It wouldn't be fair to plug in a dirt box for an amp diagnosis. Possible channel stack? Master circuit? Which would definitely up the ante on taking care with the lead dress. I REALLY HATE IT when a design is presented for diagnosis (1959 in this case) when the actual amp is something very different. And the 1959 design doesn't have a channel stack or a master volume. So...

Could be an oscillation due to lead dress. It's not "swirling" so I don't think it's crossover distortion. Can you provide a high res pic of the innards? Maybe we can pic something out visibly. If you have a scope you could put a high input signal through the front end (maybe 200mV or so) to induce clipping and trace the signal to see if and where it shows up in the chain.

And, FWIW, I think the basic clipping character of the amp is pretty damned good Just get rid of that background buzz and I think you have a winner.

I will post a pick, but it is a 1959 circuit with a switch to add a 2203 cascade - in the clip it's a standard 1959 bright channel, all knobs on 10 except bass. It seemed to aggravate the oscillation more. I have another identical one, mostly same lead dress. I will post a pic.

[Gtr0]

okay - here are some pics, from my iphone, as fast and high res as I could get without busting out a DSLR.





the power section is a bit messy with the way the filter cans are setup, typical 1959 layout.

So yes, modded in that it can cascade like a 2203, but that is on a switch. so the only difference between this and a standard 1959 is the .68uF cath bypass v2a (which I think is a stock item for one of those metal face early 70's models). and it does the same oscillation in 1959 or 2203 mode.

BTW, some of those 10k 2W power filtering dropping resistors don't like like 10kΩ, they look more like 10Ω, but they are 10k.

[Chuck H]

Excellent and fast action. Now I feel like a kill joy because I have to go do a day job thing. But I'll look the pics over when I get home. You may have a bunch of other (and possible better) observations by then. Someone always finds something.

[Gtr0]

i have 225VDC on pin 1 of v1.. I don't think thats right so I am triple checking the filtering all through the chain.

[pdf64]

Assuming it’s an oscillation issue (scoped it?), I’m not keen with the way the regular Marshall layout brings the OT primary wires through the chassis into such close proximity with the tone controls; seems to be inviting unpredictable coupling. I’d consider (carefully!) improvising some screening between them to see if that helped. If so, then turning the OT around so that the secondaries came through that hole.

Also try 100pF caps across each of the 3 preamp 100k plate loads.

Try tacking in a type 3 master vol, to see if it’s a preamp only, or whole amp positive feedback loop.

[Gtr0]

Found it.... pffffffff... after tearing out all of the 10k dropping resistors to make sure they weren't 100Ω. And that was like after 5 hours of scratching my head.

BUT, it was the negative side of the rectifier and the PI filter can ground lug were connected, but never made it to a chassis ground. I connected a wire from PI filter can ground to the chassis and it was like magic.

But I agree with you PDF, in fact I am scratching my head wondering why the PI filter cap is way over in the corner... running literally feet of wire to get there and back. I know it was probably done with trial and error, the way Marshall laid it out, but I think it's crazy inefficient - The rectifiers should be way over there with the Power supply, and the PI filter cap where the rectifiers are... probably would quickly minimize some wire runs.

Anyway, I do thank you all for your attention and help - I am glad it was a simple error on my part - you guys would have never seen that in those photos!

THANK YOU!!!!

[Gtr0]

... and I kept meaning to say, my Tektronix 465 has recently stopped giving me a trace :-( - time for finding something else I guess.

[Leo_Gnardo]

... and I kept meaning to say, my Tektronix 465 has recently stopped giving me a trace :-( - time for finding something else I guess.

You can probably fix that too. Of course it helps to have a schematic. Often the problem is in the power supply, worth checking out first. Sometimes it's the CRT. I've revived a couple of scopes that had bad high voltage caps - say 1 uF at 1500V. Newark Electronics had them at reasonable price - they're also known as Element 14 - I think they're Farnham in Europe. No more difficult than swapping out caps in a guitar amp. But you had better make sure the high voltage in a scope is discharged, 1500V is gonna hurt a lot more than 400-500V if you get zapped.

[Chuck H]

I thought it sounded pretty good other than the bee hive. How about another sound clip now that it's polished up.?.

[Gtr0]

I thought it sounded pretty good other than the bee hive. How about another sound clip now that it's polished up.?.

haha - i used to play at a place called the bee hive when i was a teenager.

1959_noise_fixed.mp3

Here's a short clip... let's see... this is in 1959 mode, everything on 10 except the volume on 7'ish. the cab is a tiny 1x12 with an Emi Governor in it - I typically play heritage greenbacks, but at home on the weekend I tend to play this little guy because he's stuffed in the back in a small iso booth i made. Oh yes, I am also attenuating it by 8db :-( with a Rivera Rock Crusher else I'll smoke the speaker which is 75 watts.

[Gtr0]

oh yea, 2203 mode...
I mostly keep it set up as a 1959, but once in a while I need something with a bit more headroom for palmy mutes etc...

2203_mode.mp3

this is gain and mast vol on 7, presence on 10 (as always!!) EQ all at 5.