Electronics > Projects, Designs, and Technical Stuff
Improving Thermal Performance?
(1/1)
jpanhalt:
Boron nitride (BN) is an exciting thermally conductive, electrically insulative material. It is fibrous and highly anisotropic w.r.t. thermal conductivity. Thermal conductivity along the axis of the fibers is 400 W/m*K (gold = 315W/m*K). Perpendicular to the fibers, it is at least 2 orders of magnitude less. There are lots of research papers from China and South Korea related to using it as a filler for epoxy to improve its thermal conductivity, particularly for IC encapsulation. As an epoxy filler, it works well but may require pretreatment (e.g., silyation) to avoid interface losses (e.g, from air). When mixed with aluminium oxide (Al2O3), it seems to work even better. The spherical particles of Al2O3 help align the BN particles to facilitate thermal conduction.
Arctic Silver sells a BN-filled/Al2O3 epoxy thermal epoxy:
https://www.amazon.com/gp/product/B0009IQ1BU/ref=ox_sc_act_title_1?smid=AKFWYWVZ484P5&psc=1
The TDS claims a thermal conductivity of 4 W/m*K. That is quite high for an epoxy. I am interested in improving the thermal performance of a MSOP-10 package, specifically the Microchip MCP73833/4.
Does anyone have experience using a thermally conductive adhesive with that chip to increase its charging capability to 500 mA without having thermal regulation cut-in.
T3sl4co1l:
Sure:
1. Reduce supply voltage (use a tracking regulator?)
2. Reduce chip or ambient temperature (or max amb spec if this is an extreme case condition)
3. Add thermal vias, use heavier copper, multilayer PCB
4. Heatsink on top maybe helps but the package thermal resistance isn't specified
5. "Doctor, it hurts when I do this!" "Then don't do that." Get a bigger/better chip!
Wait, what the hell chip are you using, there is no MCP7383... Did you mean 73831?
There are also thermal pads with 5, even 10 W/(m.K) with a squishy durometer that a spring-loaded heatsink can conform around the parts. That'll do as well as the epoxy you describe, without the mess. :-+
You didn't mention any restrictions on your present case, so there are a huge number of possibilities, that hold much greater promise, than simply trying to sink the heat out. :)
Tim
jpanhalt:
My typo on the chip -- missing a "3". Actual chip will be the MCP73834 (not MCP73833).
MCP73833/4
Had considered the MCP73831 but the MCP73833/4 has better heat performance.
jbb:
Yeah, it’s better to avoid the waste heat. I had a lot of trouble a while ago with a product warming up it’s own battery to the point it wouldn’t charge.
If you must leave the chip the same, do consider ways to get the heat from the PCB to the outside world.
Navigation
[0] Message Index
Go to full version