Peavey KB100 - overheating power transistor

[dissonance]

Hi folks,

I'm trying to troubleshoot a Peavey KB100 that was given to me by a friend. He told me that it would intermittently cut out during use. It had a few input pots that were sheared off, so I went ahead and replaced those. I'm finding now that one of the two power transistors mounted to the aluminum heatsink is overheating, and the amp is shutting down as a result. The transistor in question is Q105 on the schematics. I'm also noticing that there appears to be an unacceptable level of ripple current present- I'm measuring 2-3VAC across C134 and C142 (the two largest caps on the board). Could this be what is overheating Q105, or am I dealing with two different issues? Q105 is labeled 70474140, if that helps at all. Any thoughts or advice would be much appreciated.

Thanks,

George

[Enzo]

The transistors are running hot from excess current flowing through them. The excess ripple on the caps is not causing this, it is the result of this. The more current you draw from a power supply, the more ripple will appear on the caps as they try to keep up.

I might suspect Q105 is OK and the circuit around it is the problem, but who knows. Verify those 0.33 ohm resistors on both sides. Make sure R157 is not open. CHeck Q100, Q106 and their 100 ohm resistors.

Are the power rails even? Or are they skewed - one higher than the other?

The op amp has its own little zener power supplies. Are both the +/-15v pins at proper voltage on the chip?

[dissonance]

Thanks for the reply, Enzo. I've confirmed that all of the resistors you mentioned are within spec. I'm not sure how to go about checking Q100 & Q106. Where do you suggest that I check the power rails for skewing? Also, is U7 the op amp you are referring to?

[Enzo]

The main high voltage rails are something - I'd have to guess maybe 40 volts each? Ground your volt meter and measure what is on each main cap. Should be the 40 volts (or 45 or 39 or whatever) on BOTH. If one is 48v and the other is -32v then they are skewed 8 volts. If they are even with respect to ground, then the system is at least in balance.

I'd at least check the transistors with the diode test function of your DMM. Perhaps even remove them to do so.

Yes, U7, the one that drives the output stage.

[dissonance]

The high voltage rails fluctuate due to the excessive ripple, but they both do so within the same range (~31v-39v), so they don't appear to be excessively skewed. This was measured at C134 and C142.

At the op amp (U7) I measured a fluctuating voltage of ~8.4v-13.0v at V+, and ~10.2v-14.5v at V-.

Both Q100 and Q106 seem to test OK in the circuit, using the diode method. Should I pull them from the circuit and re-test just to be certain?

[Enzo]

FLuctuate from ripple? Ripple is 120Hz, far faster than your meter can change. Excess ripple will lower the DC reading - DC readings are averages - but shouldn;t hop around unless the whole thing is hopping around.


Your power supply is screwed, until you get that right, the rest can wait.

STart over. Pull the red wires off the board - I assume they are red? 3-pin Molex in the corner? The schematic says blue and vlack for ALL the wires, that isn;t right. There SHOULD be red wires there, and then orange wires on the other one next to it. Measure the AC coming out the redwires - I'd expect 50-60v center tapped. Is it there and stable?

Leave the orange wires connected. There are two fuses for them - F1,2 - they make the +/-25VDC supplies. Are those fuses OK - pull them from the clips and measure ohms - and are those 25v supplies Up to voltage and stable and clean? Those also drop through a couple resistors to 15v zeners. These are for the preamp ICs, but are they OK anyway?

I am concerned your 25v rails might be missing - and one of their jobs is to hold up the 15v rails for U7. Since they are bouncing around I am suspicious.

[dissonance]

At the 3-pin Molex connector with the red wires, I measure a stable 67.8v across the outer terminals (34v from center to either of the outer pins).
Both fuses test fine, however at the orange wires, I measure a stable 36.8v across the outer terminals (18v from center to either of the outer pins).
At CR136 and CR138 I see a stable +/-15VDC, but at C149 and C153 I measure only +/-20.9VDC.

[Enzo]

AH, 21v is probably close enough, they both are the same and your 15s Are OK.

I don't get this fluctuating thing. Working with no speaker load, connect the red wires back and what happens across the two main filters C134,142? Measure AC voltage across each. SHould be more or less none.

[dissonance]

I measure a varying current of 1-4VAC across C134,142. Incidentally, this fluctuation seems to correspond to a hum that appears to be generated by the transformer itself.

[Enzo]

Wait, I'm stupid.

Right by where the speaker wires come off the main board is a 5 watt 5.6 ohm resistor, and next to it in series a small .1uf cap. R180 and C147. Is that resistor open?

Scope the speaker leads, is there a large amount of RF signal there?

[dissonance]

R180 tested fine. I didn't really get any signal at the speaker connector, unless I'm not checking for RF correctly. You meant for me to connect the scope at the connector on the main board, with the speaker lead disconnected, correct?

[Enzo]

Speaker or no speaker, RF would show up if the amp were unstable like that. Oh well, it was worth a shot.

If the amp goes unstable, it will crank out full powr for all it is worth, but at a freq high above audio - 100kHz maybe. On a scope set for audio, the trace will be come a wide band of light as tall as the power rails. Speed the sweep way up it resolves into a waveform, but set for viewing audio it is just a swath of light. It will be 50-60VAC p-p. That won't sound through the speaker, but the amp will get real hot.

Beats me, start scoping everything. That fluctuating popwer rail thing. What do those rails look like on a scope? U7 pins 4, 8, what do they look like on a scope?

[dissonance]

Finally got a few minutes to get back to this one . . .

I'm still poking around with the scope, trying to find something interesting. I've noticed that when I probe one of the large caps (C134,142) with the scope set to AC, I see a triangle wave pattern when it swings positive, and pretty much a flat line at the negative peak. This corresponds with the hum I mentioned previously. All I'm getting when I probe U7, pins 4 & 8, is a flat line on the positive and negative peaks.

[dissonance]

Just started taking a fresh look at this problem, and was surprised to discover something that I hadn't noticed originally. If I disconnect the 2-wire harness (blue and yellow) that heads over to the output board, Q105 does not overheat. So, I assume that the load of the speaker is causing things to go unstable. As soon as I reconnect that harness, the P/S transformer starts buzzing off and on, and the waveform of the main rails goes crazy when scoped at C134 or C142.

Given this information, any thoughts on where to look next?

[Enzo]

You are disconnecting the speaker, yes, but what is on the output terminals now? DC?