What kind of thermal paste to use for IGBTs

Thread Starter

Dimmz69

Joined Feb 11, 2017
11
I have a p7000s amplifier that had a problem whereby it would switch off after a couple of hours. After having a look at the PS i found that the IGBTs q407 & q406 had been switched out with G4PH50UD, so i ordered some IRG4PC50U from UTsource. The parts finally came and i installed them. The amp ran perfectly for two hours then the minute i stepped out of the room it went off. When i tried to turn it on again there was a spark at q406 and it went off again. Now before I reinstall q406 I'm trying to find out the possible causes from what I had changed because there had never been a short before. Could the IGBT have been bad? Also, I used some arctic silver thermal paste that I had on hand which i discovered is slightly capacitive. So could that also have been part of the problem?

I've attached the service manual. thanks
 

Attachments

Thread Starter

Dimmz69

Joined Feb 11, 2017
11
It's not the paiste . You have other circuitry related issues.
Any chance it's just a bad igbt? Nothing used to burn out before. It would just go off after a while with the older ones and it would run really hot! I have another similar amp for comparison in the temps.
 

shortbus

Joined Sep 30, 2009
10,045
Are the IGBT mounted on a heat sink? If so did you use the correct mounting pad insulators and bushings? Just the heat sink grease alone isn't enough. It would work for a while as insulation, until it broke down, then short out to the heat sink.
 

tcmtech

Joined Nov 4, 2013
2,867
Any chance it's just a bad igbt? Nothing used to burn out before. It would just go off after a while with the older ones and it would run really hot! I have another similar amp for comparison in the temps.
If you have two units and one runs hot and the other doesn't while in the same working environment that's pretty good indicator it has something wrong with it beyond heats sink grease or a random weak IGBT component.
 

Thread Starter

Dimmz69

Joined Feb 11, 2017
11
If you have two units and one runs hot and the other doesn't while in the same working environment that's pretty good indicator it has something wrong with it beyond heats sink grease or a random weak IGBT component.
The igbts that were running really hot were the G4PH50UDs (which were not as per the service manual). After swapping them out with IRG4PC50Us temperatures were normal, but then the short happened.

That's why I was asking if the short could just have been a weak igbt?
 

Thread Starter

Dimmz69

Joined Feb 11, 2017
11
Are the IGBT mounted on a heat sink? If so did you use the correct mounting pad insulators and bushings? Just the heat sink grease alone isn't enough. It would work for a while as insulation, until it broke down, then short out to the heat sink.
Hi, yeah they're mounted on a heat sink. I used the mica insulator washers and I greased the heat sink side and the collector of the igbt. I was suspecting I may have used a bit too much grease and both sides happened to meet. And seeing how that grease is capacitive it may have caused some problems.

Haven't managed to sit with it yet since it happened yesterday. But I should be able to have a proper look tomorrow.
 

RichardO

Joined May 4, 2013
2,270
Silver sounds conductive, both thermally and electrically.
It's not electrically conductive, but they do give a warning about it's slight capacitance. I hadn't read the small print till after the fact.
The specifications for Arctic Silver 5 do not say if there is a breakdown voltage. In other words, the paste may very well be non-conductive for CPU's running at a few volts but conduct at the higher voltages in an IGBT circuit.

Thermal pastes I have looked at state things like this for a Wakefield 126 series product:
Volume Resistivity 5 X 1014 ohm-cm
Dielectric Strength 225 volts/mil

If I were doing it, I would use alcohol to remove all vestiges of the Arctic Silver paste and use a more conventional paste just to be safe.
 

Attachments

tcmtech

Joined Nov 4, 2013
2,867
Given the non isolated case design of those IGBTs there is the possibility that it shorted through the mica insulator.

As for the greases capacitive effects at audio frequencies it would be a non issue. The simple capacitive effects of the metal back of the IGBT with the mic insulator between it and the aluminum heatsink would be far higher than anything the grease would add to anything.

Still without knowing what shorted to what it it's impossible to say anything definitive. My first guess if a circuitry component failing didnt blow up the IGBT a short through its back to the heatsink would be my first guess as to where to look.

Some poking around with a basic continuity tester before disassembly should show whats shorted to what. If you have continuity between the center pin and the heatsink the mica insulator is the likely fail point.
 

ian field

Joined Oct 27, 2012
6,536
Are the IGBT mounted on a heat sink? If so did you use the correct mounting pad insulators and bushings? Just the heat sink grease alone isn't enough. It would work for a while as insulation, until it broke down, then short out to the heat sink.
If a previous owner swapped them out - they may have forgotten to refit the insulators.

You can actually get glass bead thermal paste that works without an insulator kit - but I think the thermal resistance is probably higher.

The old standard was mica washers and thermal paste, but many manufacturers switched to silicone rubber pads. Those will deteriorate if you put paste on them.

Some manufacturers do the mica washers with plain old silicone grease - the white paste with magnesium oxide is much better.

CPU cooler paste is expensive and you don't get much - but most brands list all the thermal resistance data you're likely to want.
 

Thread Starter

Dimmz69

Joined Feb 11, 2017
11
Hi guys, thanks for all the responses. So as it turns out, it was exactly the paste that caused the problem. It shorted the two igbts to the heat sink(it has no nylon bushings to isolate the mounting screws, ill get a drill and add some), fortunately nothing else suffered from it. Paste must have broken down and started conducting like @RichardO had suggested.

Anyway, I alcoholed away the paste and remounted with the normal heat sink compound (For some reason the only place i could get some was some distance, a 32km round trip :p). Seems to work okay now. Just run the unit for two hours with no problems. Tomorrow I'll leave it on for much longer to see if the original problem of switching itself off for no apparent reason has vanished. Fingers crossed.;)

You guys are great!!:cool::cool:
 

tcmtech

Joined Nov 4, 2013
2,867
So as it turns out, it was exactly the paste that caused the problem. It shorted the two igbts to the heat sink(it has no nylon bushings to isolate the mounting screws, ill get a drill and add some), fortunately nothing else suffered from it. Paste must have broken down and started conducting like @RichardO had suggested.
Are you sure it wasn't the fact that by not having the bushings that are suposed to prevent the IGBT package from being misaligned on installation plus limit physical creeping from thermal cycling until they move so far as to short out against their attachment screws was not the real problem?

To me improper device installation seems far more plausible than thermal grease dielectric breakdown. Especially at the voltages an audio amplifier such as yours runs at and the fact that the previous owner put in the wrong device to begin with showing that they likely had no clue what they were doing in the change out from start to finish?
 

Thread Starter

Dimmz69

Joined Feb 11, 2017
11
Are you sure it wasn't the fact that by not having the bushings that are suposed to prevent the IGBT package from being misaligned on installation plus limit physical creeping from thermal cycling until they move so far as to short out against their attachment screws was not the real problem?

To me improper device installation seems far more plausible than thermal grease dielectric breakdown. Especially at the voltages an audio amplifier such as yours runs at and the fact that the previous owner put in the wrong device to begin with showing that they likely had no clue what they were doing in the change out from start to finish?
The heat sink has threaded holes for the mounting screws, so it's a bit difficult to misalign the IGBTs.
 

tcmtech

Joined Nov 4, 2013
2,867
The heat sink has threaded holes for the mounting screws, so it's a bit difficult to misalign the IGBTs.
But do the IGBT's have fully insulated holes that the screws pass through? Many don't and thusly allow for bare back plate contact with the screw if there is no alignment bushing or insulation sleeve on the screw.

Also for the grease to short out the IGBT to the heat sink it had to either be put on so excessively heavy it overflowed the mica insulator which if so has me wondering what sort of grease was used to be that conductive being I have yet to ever find any that were let alone to the point they could short out a power switching device to the point of destruction that ran on low voltage.

The thing is that if you are in fact using IRG4PC50U IGBT's which are 600 volt 55 amp continous (220 amp peak) 200 watt rated devices and destroying them via dead shorts the numbers aren't adding up for such a short circuit to go though common heatsink grease. The electrical conductivity and absolute worse case cross sectional exposure area a mounted device could have don't add up to allow it.

That and according to the Arctic Silver website their grease is non conductive to begin with which makes it impossible for it have conducted enough current to destroy a 600 volt 55 amp capable IGBT though short circuit to anything.

Not Electrically Conductive:
Arctic Silver 5 was formulated to conduct heat, not electricity.
(While much safer than electrically conductive silver and copper greases, Arctic Silver 5 should be kept away from electrical traces, pins, and leads. While it is not electrically conductive, the compound is very slightly capacitive and could potentially cause problems if it bridges two close-proximity electrical paths.)
http://www.arcticsilver.com/as5.htm

Plus I have worked with higher powered electronics for many years and I do know that many of the devices in the TO-247Ac case do not have any insulation inside the fastener pass through hole either and thusly require fasteners with alignment bushings and or insulating sleeves as well when they need to operate electrically isolated from their heatsink in a application.

That's why I think it was simple physical issues that caused the short and the grease is getting blamed for it to hide a simple amateurish screw up on how the device was mounted improperly. After all the person who worked on it put the wrong devices in to begin with and you did point out he did leave out the spacer bushing that is there to keep the device from coming in contact with the mounting screw.
 

RichardO

Joined May 4, 2013
2,270
I agree with @tcmtech. There still seems a high probability that the short was something other than the thermal compound. Checking the mounting screw's isolation is definitely required.
 
Top