Yes. This is what happens when I post at stupid hours of the morning.
As for liquid PC cooling, trust me, that did cross my mind, and so did liquid cooling WITH peltiers. Very expensive solutions there.
Well folks, thanks to the kindness of Hikari1 I have good old fashioned mica installed. Results: virtually no difference over the silicone pads I was using. They may even be running slightly warmer.
Since I'm now dealing with a fairly well known and tested insulation method, I'm starting to wonder what I'm measuring or if the measurements are accurate. I'm looking at fins of 51C, the base directly behind the MOSFET of 55C, and as I measure with the IR thermometer, the hottest spot on the top of the MOSFET is measuring 79C, which is slightly warmer than the silicone. Otherwise the cases are measuring upper 60's to low 70's. I can hold my finger on the sides of the MOSFETs for about 10 seconds before it gets truly painful. I'm assuming either my measurements aren't accurate or I'm managing to measure the actual temperature of the die through the case. Frankly, I'm leaning towards the latter seeing as how I'm getting relatively realistic measurements when I hit the source resisters. They're measuring mid 60's and feel about as warm as the sides of the MOSFETs.
So.... Opinions? Analysis? Have I done my due diligence and in the safe?
Since I'm now dealing with a fairly well known and tested insulation method, I'm starting to wonder what I'm measuring or if the measurements are accurate. I'm looking at fins of 51C, the base directly behind the MOSFET of 55C, and as I measure with the IR thermometer, the hottest spot on the top of the MOSFET is measuring 79C, which is slightly warmer than the silicone. Otherwise the cases are measuring upper 60's to low 70's. I can hold my finger on the sides of the MOSFETs for about 10 seconds before it gets truly painful. I'm assuming either my measurements aren't accurate or I'm managing to measure the actual temperature of the die through the case. Frankly, I'm leaning towards the latter seeing as how I'm getting relatively realistic measurements when I hit the source resisters. They're measuring mid 60's and feel about as warm as the sides of the MOSFETs.
So.... Opinions? Analysis? Have I done my due diligence and in the safe?
Just to update
What is the Bias?
What voltage on power supply?
What heatsinks are you using?
IR thermometers are affected by the emissivity of the surface you are measuring.
A polished metallic surface eg heatsink is very different to a black surface. Not sure whether this is what you have.
What is the Bias?
What voltage on power supply?
What heatsinks are you using?
IR thermometers are affected by the emissivity of the surface you are measuring.
A polished metallic surface eg heatsink is very different to a black surface. Not sure whether this is what you have.
Grease???? You're kidding, right?
Yeah, I greased them up right. I just biased it up to 1.3 amp as 2picoDumbs suggested in the comment that disappeared. I'm looking at heatsink base temps of 57.6C, fins around 54C, cases in mid 70's, and that freaking hot spot up around 85-86C. Not as hot as I was expecting... Is this looking good?
Yeah, I greased them up right. I just biased it up to 1.3 amp as 2picoDumbs suggested in the comment that disappeared. I'm looking at heatsink base temps of 57.6C, fins around 54C, cases in mid 70's, and that freaking hot spot up around 85-86C. Not as hot as I was expecting... Is this looking good?
I consider that too hot for long term reliability.
Aiming for a Ts <50degrees C is a safe target.
But you are reporting 57.8degrees C for Ts and 85-86degreesC for a hot spot. Is that the Tc of the hottest output device?
That gives a deltaTc-s of 27.4Cdegrees.
That is very high.
If your Rth c-s is 0.2C/W then when dissipating 40W you would expect a deltaTc-s of ~8Cdegrees, not 27C
You need to do some checking.
Aiming for a Ts <50degrees C is a safe target.
But you are reporting 57.8degrees C for Ts and 85-86degreesC for a hot spot. Is that the Tc of the hottest output device?
That gives a deltaTc-s of 27.4Cdegrees.
That is very high.
If your Rth c-s is 0.2C/W then when dissipating 40W you would expect a deltaTc-s of ~8Cdegrees, not 27C
You need to do some checking.
No one can do that (at least not without gloves). Check your thermometer.
No way - if it really were 79C you would get blisters instantly.
It's obvious that you have a massive measuring error.
You're not the first to tell me I might be reading way higher than things actually are. I'm thinking the heatsink readings are the most honest.
Test your thermometer. Stick it in one of your body openings and see what you get.
You can't measure the junction temperature. It is buried inside the thermal insulation of the plastic package.
The best you can do is measure the package surface temperature or the heatsink temperature adjacent to the device.
I'm assuming if we know the heatsink temperature, and we've got a good idea as to how the interface material is moving heat, we can come up with a pretty good guess as to the junction. So if the sink immediately next to the device and behind it is reading 58, we can assume it's running warm, but not in danger of letting the magic smoke out?
