Actually, what I was hoping to do was use 4 oz. copper on one side and 2 oz. copper on the other.  The 4 oz. side would have the heavy power stuff and the 2 oz side would have the surface mount I.C.'s and control circuits.  Talk about short signal paths from logic to power MOSFET's, it would be the thickness of the board. We have used bus bars in the past for TO-220 designs and plan to at least to connect some of the sources of the MOSFET's.

I was thinking of making kind of square "U" shaped piece of copper 5 mm thick that had a cutout of 10 mm square to fit around the D2PAK.  The PCB would have copper that conducted heat from the MOSFET to the heatsink.  This leaves a strip of less than 1 mm around the D2PAK.  When I do the math of 1 mm *30mm(for three sides)*.07mm(for 2 oz. copper) I get 1.2 degrees C per watt.  If I go to 4 oz. copper I get .6 degrees C per watt.  This copper is thick enough so that it is higher than the top of the D2PAK so a big Aluminum heatsink could be through bolted on to cover lots of these pieces of copper.  If I use a Silpad (Berquiest) of the cheap type (gray color) the copper to Aluminum thermal resistance would be .58 degrees C per watt.  If I use the Silpad K10 (tan color) it goes down to .26.  So, if I add the two for D2PAK to Aluminum thermal resistance I get 1.46 degrees C per watt for 2 oz. copper and .86 for 4 oz. copper.  For reference a TO-247 with a gray silpad is 1.5 degrees C per watt and with a Silpad K10 (tan) is .66.

At any rate, with my proposed system I could imagine 50 watts continuous especially for 4 oz. copper.  BUT, I realize these are just a bunch of numbers and formulas.  50 watts on the 4 oz. copper would mean a temperature gradient of 30 degrees C and 2 oz. would give 60 degrees C over a distance of just 1 mm.  Can an FR4 board handle that?

Does anyone know the thickness of the copper on a plated through hole?  The cross sectional copper area will of course depend on the diameter of the hole.  Also, its thermal properties would be greatly effected if it were full of solder.

What am I missing here?  Any comments appreciated.  This has to be done right for high volume production, 10,000 per quarter.

Heart Transverters
Costa Rica

Gary Allbee wrote:

Heart: In addition to Rob Malos reply, I have also "heard" of 4oz copper available upon request Rob commented:"If you do plan to use 2oz or 3oz copper do discuss it with your SMD
assembly house as it will affect their temperature profiles
significantly. They will hate you lots if you mix very fine pitch stuff
with great thick slabs of copper." This is important, but also in the fabrication of the pcb you must be careful on trace width, and that is you will have a tremendous amount of etch back on a 3oz copper board and you should work closely with your fab shop for their minimum trace widths etc on a 3oz copper board.  I beleive you will have more problems with fine pitch components on the fab side then you will on the assembly side. Heart:"Our current design will use 28 of these MOSFET's. " I would recommend using the 3oz board only for power devices and put the controlling circuitry on a seperate pcb.  Other ideas for power handling could be to pattern plate more copper onto the specific power handling traces, but again you have to work closely with your fab shop. The best all around advise is to work closely with your vendors - fab, assembly, and component suppliers and make sure you stay within their capabilities - or find a better fab or asembly house that is capable of a little more.  Some fab shops wont even touch 3oz copper never mind 4oz so you may have to shop around a little. Also, loading the board with vias as a thermal relief is a contraversial heatsink technique where some swear it does nothing, and others say it does lots, and others say well it cant hurt.  My philosophy is - it can't hurt, then I mount a heatsink on the bottom side of the board, and for your power dissipation I would recommend a LOT of air flow.  Air flow is amazing, and with enough airflow you may not even need the heatsink.  I have not had the luxory of using thermal modeling software, but I would be very interested in seeing some hard data on air flow. Anyone out there have more info on air flow? Regards,Gary AllbeeAlta Industrial Automation
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * To post a message: mailto:[EMAIL PROTECTED] * * To join or leave this list visit: * * - or email - * mailto:[EMAIL PROTECTED]?body=leave%20proteledaforum * * Contact the list manager: * mailto:[EMAIL PROTECTED] * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Reply via email to