Final voltage test, episode 2 [resolved]

[Suger]

Hello again.

So I finished the first amp. Of course, soldered the black wire to terminals 6 and 7 instead of 7 and 8 at first, so had a pretty weak sound, but I found the issue easily (thanks to an older post on the forum with a slightly different issue, but this was on the list of things to be checked).

Anyhow, the amp is hooked now, and the sound is amazing, as expected. Well worth the effort and (sometimes) the frustration, even when you're as untalented for these things (but stubborn) as I am.

Second build was a breeze almost all the way through to the last meter test. And here, deception. The measurement on the A side is as follow: A spike at 400V, then down to 300, where it stabilises, before going down again, steadily, to 89V. Any suggestions? (Yes, I know, the solders, but apart from that?

 

[Paul Birkeland]

Can you post some photos of this build so far?

 

[Suger]

Here you go

 

[Paul Birkeland]

Can you post IA, OA, IB, OB, KregA, and KregB voltages on that board?

 

[Suger]

Of course, here they are:

IA 389V
OA 88V
IB 86 V
OA 85 V
Kreg B 0V (which, I guess, isn't good)

Quick edit: did a continuity test, turns out Kreg B is currently grounded, which obviously shouldn't be. So I have a path to trace back. Will let you know.

Longer edit: if my understanding of the driver and shunt regulator schematics are correct, current from Kreg goes four places: A2, 431, R4(B) and Rc(B). So here were my next steps:

A2 was grounded (obviously). So I unsoldered the white wire from A2, and checked. It wasn't grounded anymore, but Kreg still was. So the issue wasn't on that side. Resistance to ground at R4(B) is 148 k ohms, as could be expected. Resistance to ground at Rc(B) is, logically, infinite as the capacitor stands in the way.

So that leaves 431 as the likely culprit. Resistance to ground at the lead furthest from Kreg B is 2.5 kohms, as expected (connection to R3(B)). Resistance to ground at the middle pin is near null (less than .3 ohms on the multimeter), as, if I read the circuits correctly, is expected. This leaves the pin closest to Kreg B, where the resistance is less than 1 ohm too (0.8 ohms on the multimeter). So I'm guessing that this is the culprit. There was a big (at least, once looking at it with the magnifyer) blob of lead at the foot of that pin, so I removed it with a solder pump. As far as I can see under magnification, there is no solder bridge either under or above the board, so I'm a little puzzled. Shall I try to remove it, clean it and resolder it?

Some time later: so, I've removed the LM431A shunt regulator, and I've apparently managed to damage it during installation: I cannot see any lead bridge for the life of me, yet two of the pins are connected when I test them with the multimeter. So I guess the next step is to find a new LM431A and replace.

 

[Paul Birkeland]

Yes, that could definitely be an issue.  With that 431 pulled out, can you run the amp and check everything else?  OA and IB will be a lot higher than they should be, but you should still be able to perform the biasing adjustment to get OB to proper spec.

 

[Suger]

So, the day's update.

Without the LM431, I get pretty high voltage readings, c. 400V on both sides, but anyhow, I've now added a new 431 and it turns out that, for the reading on the A side, the 431 was the culprit. I now get readings of 297V on the A side, right within the 300v ±1% range.

The issue, of course, is that I also get 297V on the B side, and that tweaking the potentiometer has absolutely no effect. I'm guessing I have an issue somewhere in that circuit (which, if I read the scheme correctly, is MJE5731A to 9 pin socket and output, 9 pin socket to Kreg A, Kreg A to breg A to R4 (the potentiometer) to R3 to ground, and Kreg to 431, 431 to R3 and 431 to ground, so I'm going to check these.

In the meantime, I had a small question about the schemes and the manual:
In the scheme (p.6 of the pdf), pin 8 of the 9 pin socket is connected to Kreg B, and pin 2 to Kreg A. But, in the manual, the connections are white wire from A8 to Kreg A (p.62) and white wire from Kreg B to A2 (p.61). Is this a typo by any chance, or have I confusded everything?

Quick edit: 0V in Kreg A, so there is an issue around there...

 

[Paul Birkeland]

With your current list of symptoms, I would reheat the solder on the middle leg of the B side MJE5731A since everything is indicating zero current flow from the B side C4S.

The schematic shows the 6, 7, 8 triode as the regulator and the 2, 3, 4 triode as the driver, but they are swapped in the actual build so definitely follow the written instructions.

 

[Suger]

So there is something very wrong in my build, as for the third time now, after some tests and all that, I end up with the Kreg B side of the 431 grounded. It would seem that I have a reaction somewhere that happens and damages it, so I guess I'll just reorder the whole PCB set and see what happens.

 

[Paul Birkeland]

Is Kreg showing continuity to ground on your meter?  0V doesn't necessarily mean grounded.

 

[Suger]

Alas, yes. Each time, I end up having continuity between two of the prongs of the 431 (the ground one and the Kreg one), even if I remove them from the board.

 

[Paul Birkeland]

So every time you put in a new 431 it shorts out?

 

[Suger]

Well, I've now shorted three, so I guess yes, everytime.

 

[Paul Birkeland]

Can you post some photos of your build of the board with the 431s on it?

 

[Suger]

OK. So, two weeks later, new C4S with new components. Still no go:

IA 397V
OA 394V
IB 394V
OB 0V
Kreg A 0V
Kreg B 15V

Any suggestions welcome, of course.

 

[Paul Birkeland]

Is the 5670 visibly glowing?

 

[Suger]

It is. About as much as the 5670 on the other block, as far as I can judge.

 

[Suger]

Problem solved. I had somehow skipped the breg to bB wire. It's now working fine.

Many, many thanks for the support.

 

[Paul Birkeland]

I've definitely forgotten to install that one a few times.

 

[Suger]

That's reassuring (or is it?)

Anyhow, the beasts are now both working. Trying them up with a little seasonal Bach. And God do they sound well. Well worth the effort, and then some.