Fender stage 100 oscillates with new output transistors

[marku52]

It had a blown fuse, 2 shorted output transistors. TIP 142 and 147 darlingtons. I replaced all 6 transistors, and now it oscillates at about 750 kHz. Fun. I guess the new parts are higher gain or higher Ft.

Anybody solved this before I go sprinkling capacitors all around the neighbourhood?

[Kevina]

It had a blown fuse, 2 shorted output transistors. TIP 142 and 147 darlingtons. I replaced all 6 transistors, and now it oscillates at about 750 kHz. Fun. I guess the new parts are higher gain or higher Ft.

Anybody solved this before I go sprinkling capacitors all around the neighbourhood?

R104, R105, R106 or C48 open circuit?

[TomCarlos]

Fender Stage 100 Schematic...

Update: Let me try a new file without the hyphen.

[Helmholtz]

Fender Stage 100 Schematic...

Sorry, doesn't open here.

[TomCarlos]

Let me try a zip file...

[marku52]

R104, R105, R106 or C48 open circuit?

Thanks
Interesting. Have you seen failures there? Looks like a zero in the FB loop.

[Kevina]

Thanks
Interesting. Have you seen failures there? Looks like a zero in the FB loop.

No, never seen one, but I spent my life remote diagnosing faults in the music industry.

[marku52]

No, never seen one, but I spent my life remote diagnosing faults in the music industry.

Here's some more data. This amp has a weird output structure, the usual output node is grounded, and the speaker output actually is tied to the center tap of the power supply. I notice with the oscillation going, the power supply voltages are unbalanced (NPN side lower), but there is no unusual power consumption out of the AC mains.

My theory of the nonce is that the NPN darlingtons are self-oscillating. Something I distantly remember from E school was that an emitter follower with inductance on the emitter gets converted into a negative R on the base. Leading to an oscillation in the NPNs alone (PNPs are usually slower). This output stage is complementary emitter follower. With inductive emitter resistors.

I'm going to try adding some R to the bases of the NPNs, maybe 100 ohms or so.

I'll report what happens.

[Enzo]

That is a fairly common arrangement. Look up "flying rail" amplifiers.

[marku52]

Is there any point to it other than "We could do it so we did!"? I've seen it used where you could expand the output range of an op amp by flying output transistors on either side of the op amp rails.

But this......Why bother?

[Enzo]

I don't see it as a why bother. Imagine a battery circuit. Say a wireless transmitter or even a transistor radio. it works fine, but has no ground connection. it has a "ground" inside, but that is really a circuit common, not a ground.

Now imagine a solid state amp. It has power supplies with a common. You have output stages connected to that power supply, and it gathers into a common, a common we call the output bus. The speaker load connects between those two commons, power supply and output. If we built that on a wooden plank, it would work even without a real ground.

Well we DO have a real ground, so we can connect one end of the speaker to it, or the other. We are used to connecting the power supply to it, but we can just as easily do it the other way, the way this amp is wired. otherwise the circuits are darn similar.

[g1]

I believe one of the advantages is being able to run the front end of the power amp off the low voltage supplies. I think there are others but of course there are trade offs as well.

edit: with a bit of different implementation, this design allows all collectors to be directly bolted to heatsink ("grounded").