Combination Heat Sink & Adhesive? Good ones?

[jackerbes]

I've got an ESC (Atlas Black ABC20) that was originally cased in black heat shrink tubing.

Beneath the heat shrink it has a small square aluminum plate heat sink that rests on top of the array of IC's that get hot.

That ESC runs so hot that I exposed the aluminum heat sink plate to try to improve the cooling. Then the plate fell off. It looks like it was placed on the IC's with a layer of heat conductive tape type, the tape came off with the plate.

Is there a good combination heat sink compound and adhesive that can be used to both bond the plate to the IC's and keep it in place?

Thanks for any help,

Jack

[Julez]

Hi Jack!

It was most likely a slicon pad unter the heatsink. It has no adhesive power, the shrink tube was there for a reason.
If your ESC runs very hot, you can do the following:

Get a better aluminium heatsink from the electronics store you least distrust. There are really tiny ones, which look like this:
http://www.cartft.com/image_db/Passive_heatsink.jpg
http://www.moreleds.com/heatsink2.jpg
Put the silicon pad, or thermal paste under it, and fix it with zip ties.

If you want to be really cool, check the side of the fets where all the little legs are one contact (marked red in my picture). (At the other side, 1 leg is separate).
The legs which are all one contact can be used to dissipate heat more effectively. Solder 0.5mm copper on this side.

Cheers,

Julez

[jackerbes]

Thanks for the suggestion Julez.

I have a small passive heat sink like that I can put on there. Would or should I use heat conducting paste too? Like the kind of paste that is (was?) used on early Pentium CPUs? I have some of that also.

I don't think I'm ready to try soldering the copper on there though.

Jack

[Julez]

Hi Jack,

Yes and yes. Heat conducting paste is of paramount importance. Use the one you have.
Put a little drop on the center of each FET, and lower the heatsink vertically on the ESC without shaking. This way, you make sure that there will be no air bubbles in the paste layer between the FET/ heatsink surfaces.

Cheers,

Julez

[Tomapowa]

A few other options beside thermal grease and zip-ties:

Re-tape using stuff like this:
http://www.crazypc.com/products/PQ5-5025.html

or re-apply with heat conductive adhesive:
http://www.directron.com/silverad.html

Digikey also sells heatsinks, some with adhesive already applied (i.e. peel&stick):
http://onlinecatalog.digikey.com/WebProject.asp?BookCode=dus08flx&SectionIndex=0&PageIndex=1154#

If overheating is a big issue, you might look into why it is over-heating... possibly you are over-working the ESC due to over-prop or under-voltage conditions.

[jackerbes]

Thanks for the rest of the suggestions guys, now I'm sure I'll find something to get this working better.

"..If overheating is a big issue, you might look into why it is over-heating..."

That is sort of what I am doing. I started out with little or no impression as how hot they actually got in use, and moved on to getting a feeling for the actual heat using a non-scientific finger and a scale of "that's pretty warm, "that's hot", "that's too hot to keep my finger on", and even got as far as "Wow! that's really hot!".

This all started when I watched a 3 +/- year old Phoenix 25 go up in a flash of flame and a puff of smoke while bench testing a motor and battery combo. All three motor leads unsoldered themselves at the same moment in time and the problem stopped. Along with the ESC of course.

That was at mid throttle pulling 12-13 Amps at about 12V from a 4S A123 1100 mAh pack. My Watt's Up reported a peak of 62.5A and I think that peak resulted from the P25 failing, not from a load I put on the P25.

So all this is what kindled my interest in how hot ESCs can and do get. I got a new Phoenix 25 from Castle on an out-of-warranty repair basis (cost me $25.00) and am being more careful with it.

I put the new one on the same motor and battery and it quickly got too hot to keep my finger on. So I called CC tech and got a little lecture on linear BECs and some advice. I had a switching UBEC laying around and put that into the same setup to take the BEC load off of the P25. And found I could easily keep my finger on the P25 at 12-13 Amps. And even at 18 Amps where that motor was maxed out.

So I'm already a little older and smarter about it all. The "Atlas Black" ABC20 20A ESC we are talking about here came to me with a Atlas 2909/20 motor.

As I was getting ready to use it I did the finger tests on it and found it was getting too hot to maintain a touch at 10-12 Amps.

I opened the heat shrink over the aluminum plate, and it still stayed too hot. As I was playing around with it, the plate finally fell off (I had assumed it was glued in place, not just held in contact) and that brings us to this point.

I tried using the UBEC on the ABC20 ESC but it did not seem to make any difference. It seemed to get just as hot with the power lead removed from the BEC connector and the UBEC powering the receiver. I'm under the impression that the linear BEC in the ESC is essentially "off" and generating no heat if there is no load on it but now I'm wondering if that is the case or not.

At any rate I'm going to play around with the ESC some more when I get the heat sink on and see if I can get it to run any cooler. The plan now is to get some silver based conductive put that on the IC's, then get a heat sink in contact with them, and tie the whole thing together with cable ties to see if it looks any better.

Worst case I'll just use the ESC within its limits and see if it survives. It is rated for "do not exceed 3x Li-Poly or 4-5 servos". So I was right at that voltage with a 4S A123 (about 12V) and only had to micro servos in the plane. I was only pulling about half of the rated 20 Amps.

If I use it on another plane and the finger tests make it seem appropriate, I'll mount it with the heat sink externally exposed to get better cooling.

Jack

[Tomapowa]

HI Jack,
I'm confused my CC lectured you on Linear BECs. The heat you are feeling on the ESC FETs (if that is in fact where it is radiating from) has nothing to do with the onboard BEC. Using micro servos like you are doing should draw little current... no where near the BEC limit, therefore it should not be getting hot (sounds like over-use of a UBEC). You also have to understand that running the ESC static (i.e. with no air flow) will certainly allow it to heat up beyond the ESC specs, in fact most are now rating ESC with the understandsing/wording that there is airflow (some specify even how much) across the heat-sink. Properly mounting that heat-sing (i.e. thermal paste, etc...) is a start... see where you go from there... keep us updated!

[jackerbes]

I've not gotten to getting the heat sink on the Atlas Black yet.

But talking about the Phoenix 25, I'm not sure exactly where the heat the BEC generates is located on the ESC. The heat I am feeling is located and centered on the label side where that "field" of small IC's is located. That is the place that gets hottest when the BEC is in use.

The Phoenix has a linear BEC and it has to take the 12V I had there and convert it to 5V and does that by turning the excess voltage into heat. So the CC guy said, so have others said here, and so say the CC FAQ too:

http://www.castlecreations.com/support/faq/faq-phoenix.html#phx1

But I'm the pretty much new to all this guy here, I might have it wrong.

As for the static testing, that Phoenix 25 is not going to run any cooler in the more restricted air flow inside the fuselage than it does on my bench exposed to the prop wash and hanging out in the open.

I know there is also some effect from the prop unloading a little bit in flight. As I understand that, it will only make a minor downward change in the Amps drawn from what I am seeing on the bench. I think the changes are summarized as being that the motor will turn a little faster and use a little less capacity in the air than it does at a given throttle setting static.

I just ran the Phoenix 25 at 9-10 Amps for 6:30 and it was barely warm to the touch. I gets much hotter if I do the same thing using the BEC instead of the UBEC.

I have to change a setting on the P25 and will have to put the power lead back to do that. I'm also going to run a test again just to make sure I am not confused about the heating issue when the BEC is in use.

And I will report back on the Atlas Black ESC too, I appreciate the help and what I am learning from this.

Jack

[jackerbes]

I finally got a heat sink mounted on the Atlas Black ABC20 ESC.

It was an adhesive backed finned aluminum stick on type much like was used on Intel 486 CPUs. I narrowed it a little, used some Arctic Silver thermal compound from Rat Shack, got a dab on each of the IC's, then put the original small aluminum plate in contact with those. Then I coated the aluminum plate and got the heat sink in full contact with that.

I had to use the original aluminum plate too to get things up high enough for the new heat sink to clear the big capacitor.

Cased it in heat shrink, cut that away to clear the cooling fins, and it looked pretty good. There is hole that runs length wise through the heat sink and those are open on the ends so the air flow it pretty good through it.

I fired it up with the BEC in use and two micro servos on it ran it at 5 Amps until the temperature stabilized and it was just slightly warm to the touch. Maybe 90-95 at the most. I ran it at 9 Amps for 30 seconds or so more and it did not get noticeably warmer.

So I think that ESC is now good to go.

Now for the strange part. I used the same motor and receiver with the Phoenix 25, again using the P25 BEC and at the same 6 Amp load, and it gets a lot hotter to the touch. Like just to the point where I want to get my finger off of it. And the temp went up when I cranked it up to 9 Amps.

Then I tried the Phoenix 25 with the UBEC. It does run a little cooler with the UBEC than it does when I am using the Phoenix BEC but not a lot cooler. When I ran that at 6 Amps, it was pretty warm to the touch but not hot. At 9 Amps the Phoenix 25 is to the heat where I want to let go of it.

So I guess it is just the nature of the Phoenix for it to run that hot. And using the UBEC with it is probably not essential but does help a little. I'm definately going to work harder at getting it some good airflow and keeping it as cool as I can in use.

If I get bored enough this winter I may try to get a thermistor on that Phoenix to see just how hot it is getting.

But case closed for now and thanks for the help.

Jack