Ampeg BA115

[lowell]

Ahhh so that's what that diode is doing in the MOSFET diagram.

[J M Fahey]

Q1/4 look to me like they are in common base configuration. Is this right?

Yes.

And this provides more voltage gain than a common collector arrangement right?

No. there's umity gain (220 r on collector = 220 r on emitter; lower case r replacing the ohm sign)

Also, current gain is less than unity from what I remember, but this doesn't really matter in this amp because there is no gate current in power MOSFETS right? So we need not drive them with a low Z source?

a) For all practical means, consider current gain=1.
b) MosFets have very high input capacitance, so they benefit from being driven by a low impedance.
That said, you are driving them from 220r , pretty low

Common base arrangement is non-inverting so it looks to me like the global feedback in this amp is POSITIVE feedback. Is this correct? And if so how is this not an oscillator?

Common base is not inverting, but output MosFets are connected common source, which is inverting.
It also means they provide (a lot) of voltage gain.
That's one of the reasons these amps are not exactly stable.
And since they are inverting, you must apply NFB to "the other" Op amp input; in this case theone labelled + .
The power amp proper is U3b>Q4/5>Q2/5 ans NFB is R19/R22(plus C10.
Power amp gain is 33/1=33.

U3a is the compressor/limiter.

jumper the opto out of the circuit

Does not make much sense; what does that mean?
What is actually done ?

These amps can't be "modded for stability" , unless you call "modding" gutting it, throwing the power PCB in the junk basket and substituting a normal topology one.
In fact they *do*work, only that any fringe load triggers the self destruct gen, while the same weird load might not hurt a more normal (or classic) one.

[lowell]

Ah that's what I was missing, that the mosfets aren't oriented as voltage followers, like most amps are. Thanks JM that makes sense. How did u calculate the PA overall gain?

[lowell]

What about a fast acting fuse in series w the load? Or is this not *fast* enough?

[J M Fahey]

How did u calculate the PA overall gain?

From U3b to the output MosFets you have "big" Op Amp, connected as an inverter.
If you used only U3b, the + output (Pin 5) should be grounded , you'd need a feedback resistor from its output (Pin 7) to its - input (Pin 6) and an input resistor in series with the - input , *but* in the "big" one, Mosfets invert signal, so now the inputs must be exchanged; Pin 5 becomes inverting and Pin 6 non inverting .
So now, we ground Pin 6 and connect feedback and input resistors to pin 5.
The feedback resistor (which does not come from Pin 7 but from the output) is R19, anbd the input resistor is R22.
Gain is R19/R22 or 33K/1K=33.

Of course, a power amplifier with 1K input impedance *can* be used, but is not too versatile, so they add U3a which is easier to drive (10K input impedance (R25) and has gain electronically variable thanks to the LDR in parallel with it.
Before it they add *another* inverting Op Amp (U2B) which now has 100K input impedance, and also serves to mix the CD in.
The fuse will not hurt, but not protect it much either.
Anyway this amp's problem is not shorts (it does have protection) but I suspect under certain circunstances it oscillates violently and self destructs.
Just a hunch, eh !!