Marshall amp driver overheating

[Rotus623]

So I have troubleshooted many issues from this forsaken amp. The issue that I face now is when the amp is on, it works fine, when it is cranked past 5 it starts to decrease in volume and then increase. I checked the TDA7293 power IC, and it is very hot. The cooling fan is running just fine. Any suggestions as to why this power chip is overheating? BTW, the amp is an MG100hdfx from 2003.

[Rotus623]

So check this out fellers. The way the amp was when I got it, there was a plastic washer spacing the IC from the heatsink. I took the plastic washer out (cause the heatsink is not grounded) and screwed the IC directly to the heatsink, without the washer. I had the amp cranked to 8 with mad gain, and the IC stayed cooler than room temperature for over ten minutes. No more sound issues. Here is the next question. If the heatsink cools the part so much better with a direct contact, how come the manufacturer put a plastic washer in between them?

[leadbelly]

You've created a ground loop by removing the isolator. In a TDA7293, the tab is internally connected to the inverting input. Are you sure you want to just keep playing around and risk destroying it?

[kevinkr]

Originally Posted by leadbelly

You've created a ground loop by removing the isolator. In a TDA7293, the tab is internally connected to the inverting input. Are you sure you want to just keep playing around and risk destroying it?

Actually it isn't, it is connected to -Vs which is the negative power supply connection for the input circuitry - the power output stage has separate supply pins to allow for a class G output topology with external discrete circuitry. (I was peripherally involved in the development of this amplifier chip family with my former employer, and used the 7294 the next one up in the line.)

The downside of connecting the tab directly to the heatsink is that the heatsink is now at the negative supply rail which if the sink is exposed could be hazardous. There are better options for making sure the thermal resistance between the device and heatsink is reasonable. (Not necessarily a silpad - consider mica and thermal grease.)

[Enzo]

By plastic washer, do you mean the little shoulder washer around the screw, or the mica sheet between chip and heatsink? As kevin mentions, the mica sheet has to be smeared with silicone heatsink grease on both sides before putting it in there. The grease makes good thermal conductivity.

And make sure the chip is dead flat against the heatsink and tightened up snug

[Enzo]

How many threads you got going on this same amp now, four?

[Rotus623]

Thanks Enzo, and Kevin. Enzo, I just figured that they are all separate topics, even if it is the same amp, and the other ones really aren't going now, they are kind of dead and settled.

[Rotus623]

Originally Posted by kevinkr

Actually it isn't, it is connected to -Vs which is the negative power supply connection for the input circuitry - the power output stage has separate supply pins to allow for a class G output topology with external discrete circuitry. (I was peripherally involved in the development of this amplifier chip family with my former employer, and used the 7294 the next one up in the line.)

The downside of connecting the tab directly to the heatsink is that the heatsink is now at the negative supply rail which if the sink is exposed could be hazardous. There are better options for making sure the thermal resistance between the device and heatsink is reasonable. (Not necessarily a silpad - consider mica and thermal grease.)

Kevin, when I check for continuity of the TDA and the heatsink, there is none. I think that since the heatsink is painted, and aluminum, so maybe it is okay. I will look into some thermal grease and just grease the washer instead though. Thanks for your input.