Hi,

The PSU negative rails on a synth I'm repairing are outputting double their specified values. The -15V rails (A and B) are outputting -27-28V, and the -5V rail is outputting -10V.

The positive rails are fine. Well, the +15V is outputting a little low @ 14.2V but still nothing like the negative rails. Also, the PSU is disconnected from the rest of the synth.

The synth is old and I had measured AC on the DC rails so decided to recap and replace a bunch of old components, including the rectifiers and all of the transistors. Many of the parts are obsolete so I replaced them with suitable substitutes. The substituted parts are as follows (originals on left, replacements on right):

Diodes

D1-D4: 1S2473 - 1N4148 x 4
_________________

Transistors

Q1-Q4: 2SD313E - BD243C x 4
Q5-Q7: 2SC536G - KSC1815YTA x 3
Q8-Q10: 2SB507E - KSA940TU x 3 (BD244C)
Q11-Q12: 2SA564R - BC557BTA x 2
Q13: 2SB560E - BC640 x 1

_________________

IC

IC1: UPC141C - UPC305C x 1

I replaced the caps with good quality like for likes.

It's a bit of a noodle scratcher. I have no idea why this would be happening, particularly on the negative rails only.

I initially replaced the B507E transistors with the KSA940TU, which should be a suitable replacement. Given that those transistors are all on the negative rail part of the circuit, I thought it might be something to do with them, so replaced them with BD244C, which are complementary transistors to the BD243C on the positive rail circuit. But the problem persists.

As I said, a noodle scratcher. Hope someone can help.

The schematics of the PSU are attached.

 

I would be replacing the op amps, Lm741 is an old chip, I1, I3,I4 these chips set the regulation voltage,

personally i would bin the op amps and replace them with Lm79xx Lm78xx regulators..

 

Hey, thanks for the reply. Actually, the ICs were 4558s and I swapped them out for brand new ones (4558) in sockets.

Could it be a dodgy chip that's causing this?

 

I would say so, look at the negative rail op amp regulator, and swap them for the positive supply and see if the fault swaps or clears, start by checking the voltages at the pins 2,3,6...

 

Great, thanks. I'll have a look tomorrow morning and report back.

 

I may have found the issue. I double checked the datasheets of the transistors against their substitutes and found that the 2SA564 and BC557 pin outs are unmatched (ECB CBE respectively). Those particular ones are on the inverting part of the circuit so it looks like they could be the culprits. Now to find some replacements...

 

OK. So I readjusted the pins on the 557 trans so that they are sitting correctly: CBE (incorrect) - ECB (correct).

Now there's even more weirdness going on. When there's no load, stuff seems OK if a little bit low (and some irregularities on the +-15B rails).

However, with a load connected, stuff gets very odd. Here are the readings:

+15A: 9.56V
-15A: -9.96V
+10: 3.63V

+15B: -1.97V (!!)
-15B: -16.99V

+5: 1.86V
-5: -1.86V

Also, Q1 (BD243C) is extremely hot.

 

Some new readings. I replaced IC3 but it doesn't seem to have made any difference (though Q1 is no longer outputting a high temperature).

I disconnected all external load and these are the readings:

+15A: 14.2V
-15A: -14.1V
+10: 9.5V

+15B: -10.1V (!!)
-15B: -22.5V

+5: 4.7V
-5: -4.6V

+-5V are OK with no load. +-15V A are OK if a little bit low. +10V is OK.

So really, the outlier is still those +-15V B rails. Why might there be negative voltages on the positive rail...

 

EDIT: nevermind, by measuring the pins on the transistors I found that one transistor pair was acting wildly different to the others. Process of elimination revealed a bad Q6 transistor. Replaced it and all is working now!