May BJT's measure ok but still be broken?

I have a Dynamic DP-1000 on my bench.
The amp have a floating supply with 5 OP pairs pulling current to Gnd, and 5 pairs pulling current to Spkr out.

There are F15030 and F15031 as drivers and I tested them with DCA-75 and they produced nice curves.
But in cirquit they pulls the outputs full on when bases are connected (current limited to 2A ~1v6), when far from 2 Vbe that would be needed, but there are also 10R between BE in OP that pulls 5x6mA per side so 2 Vbe's are propably on the high side.

Someone have made an earlier attempt to repair the amp.
When I got it it was like this:
Ch1: OP BJT SA and SC swapped, VAS SA and SC swapped.
Ch2: Drivers 030/031 swapped.
Silicon pads on all OP where collectors are supposed to connect to heatsink Gnd/speaker return.
So, some of the transistors may have had large reverse voltages, maybe reverse currents to.

In IPS I have also swapped OP275's, BC550/560 and B737/D868 that measured strange or bad.
Tried AD797 as replacements for OP176 as gain/CMRR regulation, but they oscillate wild. Way to much gain.

BR

Figge
 
Yes they are, but ground side connects to HS with transistor backside.

Read about avalanche breakdown. I think the VAS, drivers and op's may be damadged by they have been inserted with wrong polaritet.

Forgot to mention that speaker side predrives by 2SA1837/2SC4793 biased by a TL431, and uses J351 and K2221 IIRC, as OP drivers. That side are insulated from HS.

Anyway, I ordered MJL21193/21194 for OP's and 15032/15033 for drivers. The F-type were out of stock, but their collectors are supposed to connect to Gnd anyway.

Figge