Peavey XR 696F - I made a booboo

[nickb]

Looking at the schematic it looks to me like Q112 and Q109 are implementing some negative feedback from the last output stage back to the driver stage. The voltages and currents seen by these transistors should be relatively small. The BE voltages are across the 0.22 ohm resistors on the output stage, so that should be quite low voltage, the collector voltages come off the emitters of Q106 and Q117 which seems like they have to be limited by the +/-25V rails. Looking at the available SOT-23 package BJT parts it seems like most are 40-60V parts. I don't think these require special parts, but I'm sure someone will tell me if I am wrong. Obviously you have to watch the pinout as you can't bend the legs around on SMT parts.

These are for over current protection. The collectors are a maximum of +/-15V. You'd want them to sink all the current that the drivers Q206/ Q207 can provide under over current conditions, so say 500mA.

[jondoe]

I can pick up BC807-25 locally, they are good for 500mA, and I only have to by 10!

[jondoe]

A nice chap from Peavey has informed me I need a ROHM MMST8598, which was the one I thought it might based on it's markings (R2K).

I have some BC807's, which are close to the MMST8598 specs, but not the same, question is, do I risk it? Or hunt down some MMST8598's, which seem to have long lead times everywhere.

[glebert]

I'm confused, is this for Q112? Is yours actually a resistor not a transistor? Was the original actually bad then?

[jondoe]

This is for Q112 "2N6533/34 PNP SOT-23 TRAN" - Mine tested dead on my transistor tester, showing only low resistance between two legs and not identifying it as a transistor at all.

Peavey have suggested a MMST8598 but I have some BC807-25, listed on the same data sheet linked. The part I removed is marked R2K (printed on it, not on the board), which matches the datasheet for the MMST8598.

[glebert]

R2K is a terrible way to mark a transistor.

[jondoe]

R2K is a terrible way to mark a transistor.

Agreed!

[jondoe]

OK so I took a chance and fitted the BC807-25, this did not resolve the problem of the bright bulb and is now causing Q116 to become quite warm to the touch, much warmer than the others as indicated by a little isopropyl alcohol on each. So I guess either this part isn't suitable or I have still have problems elsewhere

[glebert]

So I guess either this part isn't suitable or I have still have problems elsewhere

I can't see any reason for the BC807 part to cause an issue (especially at idle), assuming the pinout of the two parts is the same.

[jondoe]

I'm pretty sure that the pin out is correct. Here's a better datasheet for the BC807 while I'm at it; https://docs.rs-online.com/6e47/0900766b813ec026.pdf

I think I'll compare some readings from the other side of the board around Q212 there see if there are any obvious differences. Q116 tests OK removed from the board, as do Q115/Q121.

[glebert]

I think you are probably at a point where you need to start taking some voltage readings and using those to guide the search. You can do it for both sides of the amp and use the ones that seem right to judge the other side. I would start at the power transistors and work back upstream.

[jondoe]

I think you are probably at a point where you need to start taking some voltage readings and using those to guide the search. You can do it for both sides of the amp and use the ones that seem right to judge the other side. I would start at the power transistors and work back upstream.

I've taken voltage readings on the large transistors with just the high voltage supply side of the board plugged in (J107/J202) and both channels of the amp match each other very closely in that regard.

After turning it off and discharging the caps I took some resistance/diode readings around the big transistors and found some differences from one channel of the amp to the other, around Q114. I noticed the resistance wasn't reading the same across R136 as it was on R236, so I attempted to lift a leg to test it in isolation, at which point the resistor broke in half very easily I have some on order and if I'm very lucky the postman might bring them by Saturday.

Although I don't know how to interpret my findings, the resistance between base and emitter of Q114 was and still is around 2k, where as at Q214 the resistance is around 100ohms.

[glebert]

I've taken voltage readings on the large transistors with just the high voltage supply side of the board plugged in (J107/J202) and both channels of the amp match each other very closely in that regard.

I think the more interesting data would be the voltages driving the power transistors.

I don't know that resistance readings of transistors while in the circuit are going to be helpful (other than to note a difference as you have) since there are so many other components included (active and passive). BTW, are you keeping the polarity of the DMM leads consistant comparing Q114 and Q214? The resistance reading can forward bias diode junctions so if you reverse them you will get different readings for sure.

[jondoe]

I think the more interesting data would be the voltages driving the power transistors.

I don't know that resistance readings of transistors while in the circuit are going to be helpful (other than to note a difference as you have) since there are so many other components included (active and passive). BTW, are you keeping the polarity of the DMM leads consistant comparing Q114 and Q214? The resistance reading can forward bias diode junctions so if you reverse them you will get different readings for sure.

After some archaeological style digging through my boxes of junk, i.e. all those small orders I made for repairs and then never put the parts some place sensible, I found a 5W 0.22 resistor, it's no quite the same size/shape, but it'll do! I've put that in place and I'm now getting the same resistance reading across the Q114 and Q214. I get the same resistance reading regardless of polarity on those two legs on both transistors now.

I've taken some readings from the big transistors with everything plugged in on the bright bulb side;

(B/C/E)
Q116 -0.8v/-1.8v/-0.1v
Q119 -12v/1.8v/-12v
Q120 -12v/1.8v/-12v
Q114 -0.8v/-1.8v/-0.1v

(E/C/B)
Q101 20.3v/4.9v/20.3v
Q110 0.3v/4.9v/0.3v
Q107 0.3v/4.9v/0.3v
Q102 20.3v/4.9v/20.3v

[glebert]

I've taken voltage readings on the large transistors with just the high voltage supply side of the board plugged in (J107/J202) and both channels of the amp match each other very closely in that regard.

So when you do this is it with a dim bulb? Are the 40V and 90V rails symmetrical?

On your "bright bulb" voltages it looks like something is really pulling down the negative voltages. Can you measure the voltages across all of the 5w resistors (R105, R106, R126, R133, R136, R137, R178, R180)? Something is pulling current and those should show you where it is flowing.