Mackie SR1232 amp issues

[Mars Amp Repair]

Hello folks,
I apologize ahead of time for the diagtibe!

Been working on a Mackie SR1232 power amp section of this powered monitor. I think it has 2x12" woofers & a horn.
The first issue is I received it from another tech who could not repair it and actually went backwards shorting out some FET output transistors. FETS Q11 & 17 have been replaced an surrounding circuitry been checked. the symptom once the amp was restored to it's original condition w/ he received it is the same. Past a certain volume level the sound gets all flabby & distored.

My take on the power supply arrangement to the rails of the output stage is that up to a certain power demand, +-25V is supplied through D27 & D3. Past that level the circuit in the power supply from Bridge rect D1 composing of U1 and Q11 & Q17 switches in the higher voltage via FB thru U1.
I'm again surmising (guessing really) that this circuitry endeavors to keep the + and - rails equal as it appears to only 'regulate' the negative supply thru L3 & L2 via the feedback to U1 Pins 1 & 2.
However, when you pass a certain level L3 (or L2 can't recall) in the power supply squeals as if there is an inordinate load on the negative higher voltage supply to the output stage that appears to kick in at that point.
Again in hopefully a slightly educated guess to isolate the issue, I believe that Q21 and Q23 to the right of the output stage act as crowbar protection if there is any inordinate DC offset from the output stage, so I disconnected the drains of those 2 FETS and can see that the circuit that supplies the higher +- voltage to the output stage does indeed try supply the higher voltage to the + half (I only went up to 100V with volume level in effort to not blow shit up) but the negative half has some kind of issue where it most likely is drawing too much current that causes L3 or L2 to squeal and not supply the higher voltage.

Issue I have is that I cannot find any obvious bad components in the Neg side of the output stage that would be causing this in ordinate current draw so I'm still not certain if the issue is in the output stage or this weird circuit with L2 & L3 in it.
I've also tried to figure a way to disable some of the protection circuitry to the right of the output stage & cannot figure a safe way to isolate that circuitry.

I have replaced the opto-IC-1 and Q17 & Q11 as mentioned previously.

Weirdly IC-1 outputs a square wave to the FETS Q11 and Q17. I really don't see any PWM generator anywhere there. I suppose it may be a pwm type output that is ultimately filtered thru L3 & L2.
Anyway I know this is kinda like the car mechanic hears a long explanation from the consumer when the mechanic a the end of it says: 'well maam, if you hold the car up to the phone, maybe I can tell you what's wrong with it'!
Anyone had any experience with this beast?

Thanx, Glen

[J M Fahey]

This is a wild beast of an amp, completely different to anything else.

This is a seldom used Class G amplifier, pulled to its limits.

Available rails would allow it to get 300W into 8 ohm ... but they get 1200W RMS instead.
How?

At low power they are fed from the +/-25V rails.

When peak signal reaches that, a comparator detects it and switches one of the rails HARD to ground sing Q21 or Q23 as needed, which are used as switches, thatīs why they get a squarewave.

So when signal swings UP, Q21 GROUNDS the Negative rail (indicated as #2) and power amp has the FULL rail to rail voltage available for that half sinewave, a beast, when going Negative same happens with Q23 grounding positive rail.
Twice the voltage means 4X the power, just like that.

So do NOT disconnect Q21/23 , they are vitally needed.

That amp or a very similar one (same working principle) has been repaired earlier here (search under the Mackie Thump! keyword), the one who did may chime in or you read that old thread.

[Mars Amp Repair]

Ah HAH! That explains a lot! Makes much more sense. I'll have to attack the issue with that in mind. I will say that disconnecting Q21/23 did allow me to somewhat isolate the issue as it just wouldn't switch to that mode with those 2 connected.
Now that I think of it on the output one half of the driving sinewave I used for a signal was greatly surpressed. I'm supposing the output stage is ok given that it works fine up to when the amp tries to switch to the 'G' class mode. I can probably focus on the circuit to the right of the output stage where all the 'G' class stuff goes on. I'll keep you and the rest posted. I'll see if I can find the Thump post.
THANX again! glen