Leech PA120 Service

[pdf64]

A friend wants to put his old 70s Leech PA120 4 channel PA amp made in the late 70s back into service.
He tried and popped its 5A mains fuse, then tried a 2A fast blow in it, which popped straight away.
I put a new T5A fuse in and powered up through a 100W light bulb limiter and it seemed fine, and off the limiter it runs fine, puts out pretty good power into a dummy load.
But on removing the back panel assembly -
1/ the heatsink has broken in half, probably from the amp being dropped.
2/ the 2 main ecaps for the power supply rails have vented, so probably best to replace them; it may have been the surge current from their initial power up, and subsequent outgassing from that, which popped the fuse the first time.
The 40mm diameter makes it a little bit tricky to source ones that will fit the clips but they are available (63V 4700uF).

The power transistors look like they will still have a fair chunk of heatsink to work with, so I was planning to use some steel bracing to bolt the heatsink / back panel back together. But to mitigate the risk of the power transistors overheating and shorting, I was wondering about replacing them with beefier modern 20A 250W types?
They are 2N6254 RCA, date code 7902, quite a respectable 15A 150W.
The only speaker protection looks to be the F5A power supply rail fuses on the power amp pcb

Sorry, I can't find a schematic.

So just fix the heatsink back together, replace the 2 big ecaps and clean up the positive and negative rail fuses mounted on the pcb, and leave well alone regarding the power transistors?
Or replace the power transistors too, eg with MJI5003?
https://imgur.com/gallery/O7fOhao

[Leo_Gnardo]

So just fix the heatsink back together, replace the 2 big ecaps and clean up the positive and negative rail fuses mounted on the pcb, and leave well alone regarding the power transistors?

That's what I would do, also replace any other old electrolytic caps. Scan for any roasted or drifted resistors. You've already proven the amp works. I'd leave the silicon parts alone unless any signs of trouble crop up.

[Helmholtz]

As the limiting parameter is max junction temperature I would base my decision on measurement of the case temperature in built-in condition at the point of max transistor dissipation (40% of max output power? Would need to look this up.).
Means thermocouple epoxied to the case and calculating the junction temperature using the Rthj-case from the transistor datasheet and the max transistor dissipation (IIRC about 20% of max output power (Vcc)²/2*Rload).

https://www.seas.upenn.edu/~ese319/L...plifier_08.pdf

[Leo_Gnardo]

If you can position a fan to cool that half-a-heatsink, so much the better.

Those who wish can sing Monty Python's "Eric The Half-A-Bee" now, in honor of their 50th birthday.