[morzh]

In the post 43 (p4) I asked to short (put in the shorted position) those VR3 and 4.
I guess they did not short after all.
These dictate the quiescent current of the output tranistors.
But even at maximum they will not make it dangerous.
If however they are open or way out of tolerance (on the high side) it is different the output pair gets fully open, especially the upper transistor (Q5 and Q6).

As for FM, I am no expert, maybe others will chime in.


PS. If both those trimpots went bad and the two others, VR1 and VR2 are the same type, maybe you want to change them too.
I would measure them first in circuit, should be the same. They dictate the gain of the amp.
One thing though, both VR1-2 and VR3-4 will need calibration. Before you make the whole thing work I would keep the 3-4 in shorted state.

You will need an audio generator, an ammeter and, if possible, a scope.
VR3-4 should establish such a quiescent current (measured by the ammeter in series with B Power as the difference between the shorted position and a non-shorted one) where the crossover distortion on the scope disappears (or you stop hearing it) and the quiescent current is on the order of 50-200 mA or so.
VR1-2 will establish a desired gain, say full power at 250mV input signal, which can be measured by a scope or a true RMS voltmeter.