The website claims that you can run 1.5A per phase without active cooling, but I find that when I run about 1.1A continuous that the chip gets too hot for my liking. I have some small aluminum finned heat sinks that I can attach, but only to the top of the plastic casing of the chip. I realize that they would be better on the bottom side of this board, but I'm using a board with sockets for plugging it in that won't allow for that.
The question is whether or not thermal paste will actually bond to that plastic casing to attach the aluminum heat sink. Can this be expected to wick away a fair amount of heat? For what it's worth, 3d printing folks often use double sided tape to attach these little heat sinks but that seems like a very poor solution to me.
I would suspect that if you rough up the plastic surface with some emery paper to eliminate the shiny surface then the thermal glue will stick.
Use as thin a layer of paste/glue as possible.
Ive seen heatsinks glued to DIP package in commercial units, for example an amplifier. Too bad i wanted to see the chip part number and could not see it because of that.
What i'd like to see though is some real world numbers. Temperature of die without the heat sink and then with the heat sink. I would want to see a significant difference too.
Ive seen heatsinks glued to DIP package in commercial units, for example an amplifier. Too bad i wanted to see the chip part number and could not see it because of that.
You can get thermally conductive double sided tape too. I've used it in the past to attach Luxeon star leds to heat sinks which can get quite hot. It may look like normal double sided tape but the adhesives are different.
Gluing a heatsink on the top of a plastic cased IC is a Mickey Mouse way of trying to cool it. Obviously it is dissipating too much heat (too much current and/or too much voltage across it).
That sounds interesting, but do you know the thermal properties of the super glue?
That is, not the thermal conductance, but how the stuff reacts to heat...will it let go when heated?
If we could find some specs on the glue that would be good.
I think most super glues are not very resistance to heat, unless they have a special additive.
+1 That's the best choice. Make sure the surface is smooth so there are no air gaps. Heatsinking on a plastic face is weak at best because the thermal resistance is so high from the die coming out, but it's better than nothing and I have done it before.
The glue's thermal resistance is negligible compared to the plastic of the package that the heat has to pass through. The glue won't let go if it gets warm. Since it's attached to a heatsink it probably will never get hotter than about 100C.
Don't know the real heat range before it breaks down. When still working as a die/mold maker we used it to hold parts together to do multiple parts in one setup. You had to heat them with a propane torch to get them apart when done. It didn't "color" the steel parts so it was under ~400 degrees F. So normal heat from a component/IC should be OK.
The glue's thermal resistance is negligible compared to the plastic of the package that the heat has to pass through. The glue won't let go if it gets warm. Since it's attached to a heatsink it probably will never get hotter than about 100C.
Yes i know about the thermal conductance, in fact i've used high temperature epoxy for high power LEDs in the past and got results similar to using a brand name thermal adhesive. Almost anything 'thinly' applied will work.
I was more concerned with the temperature stability. Should be easy to test though with some boiling water.