Ampeg SVT-4 Pro heatsink question

[glebert]

Working on a project SVT-4 Pro because I am a masochist and have run out of less frustrating projects to work on. I found shorted output FETs on Channel B, after removing those Channel A voltages looked good so I tried it and that channel seems to work fine (actually a really nice sounding amp). Looking at the voltages on channel B now that the FETs have been removed it looks like I have power resistors to replace in addition to transistors. There are two massive heatsinks on this, but instead of one for channel A and one for B they have all the N channels for both channels on one and all the P channels on the other. The resistors are underneath the heatsinks, so I need to remove them (at least one) but I had a heck of a time getting the channel B FETs off from the heatsink. They used adhesive thermal tape instead of insulators and grease. I had to use a chisel to get between the FETs and the heatsink and pry them up like they were almost welded down. Is this normal, or just what happens if they get massively overheated? I took the screws out of some of the channel A FETs and they seem stuck down just as hard. I don't want to break good devices by prying them off the heatsink. What is the best way to do this? I may be able to unsolder the devices and remove them while still attached to the heatsink. Hoping I can get away with only taking one heatsink off.

Thanks in advance,
Greg

[Jon Snell]

The transistors are fixed to the heatsink with thermal adhesive. There is no need for mica insulators as the heat sink is live and connected to the power rails.
It is very unusual for the MOSFETS to fail unless the bias set pre set has bad connections.

[glebert]

The transistors are fixed to the heatsink with thermal adhesive. There is no need for mica insulators as the heat sink is live and connected to the power rails.
It is very unusual for the MOSFETS to fail unless the bias set pre set has bad connections.

My heatsinks are not live, they are grounded through the fan bracket to the chassis, the thermal adhesive tape is not electrically conductive. The real question though is how to remove a heatsink without breaking the FETs that are still working. The bias will be checked when I start replacing the relevant components.

[Jon Snell]

My heatsinks are not live, they are grounded through the fan bracket to the chassis, the thermal adhesive tape is not electrically conductive. The real question though is how to remove a heatsink without breaking the FETs that are still working. The bias will be checked when I start replacing the relevant components.

My mistake, wrong model. Some have live heatsinks and others not.
I would de-solder the MOSFETS from the board first and would use with Metcal de-soldering station.
If you do not have de-soldering equipment, you will need lots of de-solder braid and plenty of time with a hot iron.
Good luck, you will fix it.
Did you find out why they blew in the first place?

[glebert]

I got the heatsink off. I used my $15 desolderer/solder sucker + iron to get off most of the solder and then followed up with braid.

[nevetslab]

Long Term Thermoset does that with the Bergquist K6 or similar greased-mica insulator substitutes on the TO-247, TO-126 & TO-3 devices. For sure, getting the P-channel heat sink removed from the PCB is a tedious operation. Even if you find the non-mics insulators appearing to have no damage (no peeling away of the top surface, sticking onto the devices' heat spreader), I'll not reuse them, having had breakdown & failure of the devices. I go back to Mica and thermal grease.

Thermoset is usually a long-term time phenomena. Short-term high temperature cycling, as can happen if the fan(s) fail can also cause that, I think, though I'd have to read up on that. Pretty sure Bergquist has data on that and probably available on their website (or other manufacturers of similar materials).