Would like to re-use an SS power section

[ChopSauce]

Despite my interest for this challenge, the priority of fixing this is lowering day after day. I no more intend to throw the board without delay, but I have many more important things to do right now.

Since there aren't small screws with nylon shoulder washers thru the TO-220 xstr heat sink tabs, you'll have to look to make sure there isn't any continuity between either driver's collector (metal tab & middle terminal) AND the heat sink.

I wish it had been so straightforward, but maybe I have been missing something even more evident.

I'm obviously back to having to trace the circuit, but - with your help - I have now much more information available to help me understand what I'm looking at, & possibly find a fix for it.

Thanks again!

[ChopSauce]

Today, I had some time to further investigate this issue. Well, my controler is not that much reliable but I could find no short at all. Neither any component directly connected to the heat sink, nor any seemingly faulty semi-conductor.

Just out of curiosity: what about definitely isolating the heat sink (removing the small 100R resistor & the 100nF cap in parallel):

- those latter two shouldn't be doing anything obvious in the circuit, should them?

[nevetslab]

That's worth a try. I would then connect your voltmeter in DC Volts mode between Ground and the heat sink to see if a voltage develops as you power it up. If so, what potential is it at? Same as the collector of either the NPN or PNP power transistors? If so, it suggests a bad insulator (mica or thermal pad meant to isolate the collector surface and the heat sink.

[ChopSauce]

I'll do that, thanks. That and possibly running the amp on idle with chips physically lifted from the heat sink - just to see if it changes something.

I also envisionned a counterfeit power chip - as that also exist - and tracked a thermal dependent failure, by (seriously) warming the heat sink with a heat gun, but nothing changed (much).

I still feel like I have been missing something obvious but ... what?

EDIT: I know! I have been missing the fact that the failure can only come from the power chips, or the driver ones. These are the sole components physically connected to the heat sink. I'm gonna check the assembly.

[nevetslab]

Let us know what you find. While removing the power transistors physically from the heat sink will completely isolate them from it, they will get warm or even hot if left running for any length of time, as under nominal conditions, there is some current flowing...though mostly the TO-220 drivers. They would be biased such that the power transistors would be near turn-on, but not in full conduction...unless there still is a circuit fault. If you don't find any problem with the transistors lifted up, you could just replace the insulators to be safe. I will re-use Mica insulators, after inspecting them for fractures. The rubberized-silicon-embedded insulators, those are NOT for re-use, as the part has undergone 'thermoset' from long-term use, and prying them up breaks down the insulation. Pretty sure you have greased mica insulators. Also need to be certain the three axial lead diodes' leads are not in contaact with the heat sink...only their body, which isuusually gooped with thermal grease so they track the temperature change

[ChopSauce]

Got it
(does it seem)

On that cheap amp, the mica insulators are actually plastic ones. One was leaking. Note sure this was the insulator or the termal paste - maybe hygrospic or ... whatever, though.

I wiped the leaky insulator to check for cracks or ... and interchanged the insulators of both power chips: no more voltage on the heat sink

Afterwards I guess this power amp unit was know to leak and the 100R resistor was there to drain the heat sink. What else?

I don't know where to buy mica insulators so I'm gonna re-assemble the chips with the old plastic ones, but new thermal compound paste, and proceed to some running tests.

An amp failing for a 1 cent component ...

... it seems time is coming for me to upgrade my controller, also - it was reading ~45 volts between the heat sink & the ground before the fix, for what it's worth.

Thank you very much, anyway

[g1]

I think when properly assembled the heatsink will be physically grounded by it's screws. The resistor is there in case the physical ground is compromised. Because you had a fault putting voltage on the heatsink, and you were missing the physical ground, the resistor burnt.
I bet when it is properly working and assembled, you will measure 0 ohms across that 100 ohm resistor.

[nevetslab]

You can find TO-220 Mica insulators and, I belive MT-200 Mics insulators on ebay. The MT-200 is for your Sanken power transistors. You'll also need to buy silicon grease, which is applied to the back of the power transistors/drivers and then, after putting the mica insulators on, and mushing it around to get a good thin coating of the grease on all of the hata spreader surface or each transistor, apply same amount to the back side of the mica. Then re-mount to the heat sink. Make sure you have th etiny nylon shouder washers for the drivers...unless those are just clamped down by that metal clamp we see on your heat sik. I now forget if I saw mounting screws on the drivers.

If you can't find any of the MT-200 Mica insulators, I have some here in California that I could send you.

Once you've got the transistors properly mounted, that should eliminate the DC leakage, and that 100 ohm resistor & 100nF cap can be restored. And, there's a good chance your amp will work. If not, we'll find the next problem as we go forth.

Good work on the findings!!

[ChopSauce]

I already have some good thermal compound paste left, but I noted the reference (MT-200) for the insulators of the power chips, thanks a lot - one more time

Indeed I will have to improve this ... ... assembly.

EDIT : I was a bit rude. After further inspection, it seems the mica insulations actually are made of mica, though one indeed leaks ...

I think when properly assembled the heatsink will be physically grounded by it's screws. The resistor is there in case the physical ground is compromised. Because you had a fault putting voltage on the heatsink, and you were missing the physical ground, the resistor burnt.
I bet when it is properly working and assembled, you will measure 0 ohms across that 100 ohm resistor.

I don't think so, otherwise why mouting the chips with an insulator?

Please note that he heat sink -*-is-*- (and has always be) properly connected - to its own ... how do you say: trace(?) on the PCB, which trace is connected to ground only (could I check...) through the (little) 100R resistor. So the heat sink cannot be more physically grounded but if it were through the chips, whose cases are around 45V above ground.

Not sure I am clear enough on this one ...

[nevetslab]

MT-200 & TO-220 Mica Insulators

I"ve taken some photos of one of my Sanken Power Transistors, a 2SC3264 NPN Pwr Xstr, MT-200 package, along with a typical TO-220 driver transistor. I've also shown the associated mica insulators for each. The TO-220 in this case has the metal collector tab. The TO-221 case type is all plastic, and NO exposed collector surface on the back, so those types can be mounted with just thermal grease or grease-less pads to transfer the heat to the heat sink. On the larger Power Transistor, the metal slab on the back is at collector potential. The two thru-holes in the plastic body is INDEED electrically isolated from the collector. On both transistor types, the middle pin is collector, and is common with the metal surface seen on the back.

I've also shown the two Mica insulators for each type. On the TO-220, if the only mounting meands is a small size screw, there is a nylon shoulder washer that insulates the screw from the metal tab. In your case, I believe there is a metal clamp that is common to both driver xstrs as well as the three bias diodes.



As to the failure mode you experienced, you stated having found a cracked or damaged insulator installed between teh transistor and the heat sink. That is enough to cause the failure, as enough current came thru to burn the 100 ohm resistor between heat sink and ground.