Marshall 1987X crackling distortion

[cassiddidy]

Hey, all. I've been working on this Marshall that somebody brought in, and after replacing both the power and output transformers, it began to have this weird distortion. First half of the clip is with presence maxed out, second half is with it turned below 4 on the dial

mrsh_1987x.mp3

Which I really think is unrelated to the transformers, it just didn't show before. I've tried swapping with known good tubes, the sound is still there. I thought maybe it was the NFB, so I tried disconnecting it. The sound is still there, although different. Sound was consistent with the fact that the presence knob is essentially a tone control for the NFB, reintroducing high frequencies back into the main signal. I've even gone through and resoldered the entire board in case there were any cracked solder joints.

Then I got an ESR tester and checked the capacitors. One of the filter caps (half of a multicap) showed a high reading (over 2 ohms), while the others seemed to be just a little higher than they should. No idea about the age of the amp, but I highly doubt it's younger than 10 years old.

The question I have, while I wait for these new caps to show up, is do you all think that this could be caused by bad caps? Or am I going to replace these things and find the sound still there? I've been looking around for a solution to this, and some posts have led me to believe this could be a cold bias issue. I've biased them to 70% (well, one of them... you know what, that's complicated...)

Working with a pair of matched EL34s, and it turns out one only comes up to 50% of tube's capacity when the other is at 70. I oddly enough have a spare pair of tubes that comes out to the same difference, so I tried putting 2 of the "hotter" tubes (and then the 2 "colder" tubes) together and ended up with the same distortion. I then tried a pair of properly matched 6CA7s, biased them to 70%, and got the same exact distortion.

I'm hoping it's the caps. But I'd like to be prepared if it isn't. Any ideas?

Thanks in advance

Cass

[Reeltarded]

I cant listen to the mp3.. some error.. I am confused by a couple things you have said but do this:

Using a very small amount of a nitrogen spray with the amp operating... sssst the tiniest amount on all plate resistors from input to output.

Don't flood anything and dont contact the parts with the tube from the can. Those cans of compressed air are fine. You have to invert the can and only tap the trigger. You will hear it if you have found it.

Tap a little everywhere on the board with a non-conductive probe.

Why were the txfrmrs changed?

[martin manning]

Sounds like it's breaking into oscillation... a lead dress issue perhaps. Noisy plate resistors make a continuous bacon frying sound.

[bepone]

this sound reminds me of:
1. unsoldered transformer wire on some output tube socket pin ?
2. sparks and lightnings in the output tube (too high voltage)?

[Littlewyan]

Sounds like a dodgdy solder joint or perhaps a grounding issue. Start with basics, check voltages everywhere and especially check the bias voltage on both tube sockets to it's the same on both. Check voltages all over the PI too.

[cassiddidy]

The transformers both failed a shorting test. I've built a tester that utilizes a 30V neon bulb lit by the voltage kickback from "charging" a transformer with a <5V DC supply. The primary on the OT failed, as did the heater winding on the PT

I did leave some things out of my initial post. In addition to reflowing the solder joints on the board, I did all the joints on the sockets and pots as well. I replaced one of the power tube sockets when I did the transformers (the pins seemed loose so I decided to be preemptive). The other socket had already been replaced before the client purchased this amp. Seems like there was some corrosion around that socket, the old one might've had a short, might've burned out. I cleaned the contacts on the socket and resoldered it.

I've basically resoldered this amp a few times. I've also checked tubes in different sockets. Still had the same 70-50% ratio no matter which sockets I put them in. I've rerouted the leads running from board to power tubes so that they run over the board instead of alongside the transformer leads. And I've shortened the NFB lead, rerouting that to avoid the transformer leads. I've also tapped EVERYwhere with a chopstick, nothing. Dang, I did a lot now that I'm writing it all down...

I haven't checked for arcing inside the tubes (definitely no arcing inside the amp). I also don't have canned air, so I might wait to finish the cap job before I go out to buy some. I also haven't checked voltages aside from the plate voltage (490V). I wasn't sure what to expect from the other pins, so I left that alone. What should I be looking for for voltages?

Also, while I don't think I can rule out a lead dress issue, I'm not sure what else I could do with that that I haven't already tried.

490 is a little high for a plate voltage, isn't it? I know it's still within the limits for an EL34, but I was expecting something more like 460

[cassiddidy]

Actually, that's 490 unloaded. More like 475 with the tubes in. Still, this issue seemed to be happening regardless of where I set the bias or what tubes I used

[Littlewyan]

That plate voltage sounds right for a 1987x. Check your bias voltage on pin 5 of both EL34s, then check it where the 220K grid leak resistors meet. It should be the same in all three points.

[pdf64]

That plate voltage sounds right for a 1987x. Check your bias voltage on pin 5 of both EL34s, then check it where the 220K grid leak resistors meet. It should be the same in all three points.

Just to note that it can’t be exactly the same though, unless a vtvm is used.
The 10M loading of a typical DMM will result in the pin5 voltage being a little less than 0.98 (10M/(10M+220k)) of the bias supply voltage, eg with -50V bias it’ll measure a volt or so less.

Has the amp been checked for oscillation? Because that’s what it sounds like to me. And it may also explain the differing idles, ie they’d not at idle.

[bepone]

... The other socket had already been replaced before the client purchased this amp...

i'm missing something if i'm reading "client" here who will be charged for service? and we are providing free repair here?
if so.. i'm out sorry. for diy no problem i have unlimited time and help but for other cases.. 0 time

[Littlewyan]

That plate voltage sounds right for a 1987x. Check your bias voltage on pin 5 of both EL34s, then check it where the 220K grid leak resistors meet. It should be the same in all three points.

Just to note that it can’t be exactly the same though, unless a vtvm is used.
The 10M loading of a typical DMM will result in the pin5 voltage being a little less than 0.98 (10M/(10M+220k)) of the bias supply voltage, eg with -50V bias it’ll measure a volt or so less.

Has the amp been checked for oscillation? Because that’s what it sounds like to me. And it may also explain the differing idles, ie they’d not at idle.

Sorry, should have been clearer. I'm more interested in it being the same for both valves. Trying to rule out a leaky PI capacitor. It'd also be good to know how the bias is being measured.

[pompeiisneaks]

... The other socket had already been replaced before the client purchased this amp...

i'm missing something if i'm reading "client" here who will be charged for service? and we are providing free repair here?
if so.. i'm out sorry. for diy no problem i have unlimited time and help but for other cases.. 0 time

I don't really understand the tone of this reply, but it seems a bit rude/hostile. If you don't want to help, that's absolutely fine, but no need to berate someone for asking for help just because they are doing it for a paying customer instead of a hobbyist. There are plenty of people here that don't mind helping anyone no matter who they might be. There's no need to make negative comments because you feel someone's asking you for something you're not willing to give, just don't give. Nobody's the wiser, and nobody has to feel slighted. I've seen many people post on the forum about repairs or custom builds for customers and they deserve equal assistance. All people, professionals included, need insight and help from time to time.

My primary goal on this forum is to keep it civil.

~Phil

[cassiddidy]

Thank you. I wish that didn't need to be said. I mean, I did mention at the very beginning of this post that this amp was something that somebody had brought in. I thought for sure that would have explained the situation.

I don't run some big shop or anything, this just helps me cover a little grocery money. It's something I started doing for myself, and after years of looking through forums and troubleshooting dead amps and synths felt good enough to offer services to people in the building.

Reading forums like these helped me learn a lot. And even if I were to figure this out on my own, posting about it and seeing myself and others discussing the problem could help someone else out. Feeling the need to dig at somebody for that is pretty silly

[cassiddidy]

... moving on...

The reading at the grid leak resistors is -45.6 and at both pin 5s it's about -44.6. That's with one of the tubes current reading around 36.8mA and the other reading 24ish. If I adjust the bias trim down, I can't get it to -50V. It only gets to around -48. But this brings the bias current to around 24-26mA/11-16mA on the tubes.

But the sound is there in both cases (feels more aggressive when the tubes are hotter).

The tester I'm using is tube depot's "bias scout kit". It gives test points for plate voltage and cathode current. This is a link to the pdf manual. https://s3.amazonaws.com/tubedepot-com- ... 1477059879

The caps showed up this morning, so I've made the switch. And... well, yes. The problem is still there.

One other thing I've noticed, the bias current doesn't seem to stop rising. It's very slow, but I haven't found the top point and I'm not sure that it's related to this distortion. Only thing I can think of at the moment.

[bepone]

[

I don't really understand the tone of this reply, but it seems a bit rude/hostile.

My primary goal on this forum is to keep it civil.

~Phil

not rude just right question. if the guy has a repair shop he is charging other people knowledge to me this is apsolutečly wrong. i saw also before here this situation. from my point of view apsolutelly wrong.if for you is ok then ok proceed. in our side of sea this is rude what the guy is asking. maybe diferent mentality of nation.