[Amps] attaching a heat spreader to a heat sink

[kg7hf at comcast.net]

I agree with everything here, I just think that it's likely overkill to need to do the final sanding step.   If you have a good planer, and are careful in the setup, the runout should be very minimal, maybe on the order of 0.001 to 0.0001.  More than likely, the problem you will see is not a "bow" or warpage issue, but uneven an uneven plane where one side is slightly thicker than the other.  I wouldn't think that would be a problem either given the differences, there might be some slightly greater heating on the thin side/end.

I have also noticed the same as you, having the heat spreader sheered, not cut seems to induce bowing/warpage, but that is easy to take care of too.  Now, I don't think I would try running my copper heat spreader through the planer, maybe if I was desperate and wanted to waste a set of knives, but that is pretty hard stuff for something spinning at 10,000 rpm to deal with.  But, to get rid of burrs an such on the end, just use a flat file and file it, who cares if the very edges don't contact the heat sink?  

I've only done one solid state amp so far, based on the MRF141G (based on the an313 app note), and it seems to work out fine, but I do have quite a big heat spreader, I think that likely with fans, I could get away with not having an al heat sink.

I'm in the process of making a new amp, based on the ARF1500, these are the same guys that are in the new Tokyo HP 2.5 amp.  I've actually been considering takeing a different approach on getting rid of the al heat on this one.  Here is what I'm toying with.  Not use an al heat sink at all, and go with an integrated copper heat sink.  This could be made for example by cutting 1/2 or 3/8 in copper pipe, and stacking them on top of each other, then soldering that stack together, perhaps by dipping them in a solder pot.  Then, attach that stack of short pipes to the heat spreader in the same way.  In this way, the heat sink would become one unit.  A fan could easly be attached to one end of hte pipes to force cool air through them.  


73s Paul kg7hf
-------------- Original message -------------- 
From: "Tom W8JI" <w8ji at w8ji.com> 

> > I would think that might be overkill, because when you 
> > bolt the spreader to the heat sink, the two pieces are 
> > going to conform to each other, in fact, they are going to 
> > expand and contract as the heat is applied and dissipated 
> > anyway. 
> 
> 
> As I recall making temperature measurements for a Heathkit 
> FET amp it could not have air gap, even at the micro level, 
> between the plates using the suggested thickness of copper. 
> The finish had to be pretty good. 
> 
> One solution is to use a really thick copper spreader. I 
> think I doubled it. If it bows up a little or if it has 
> small surface irregularities (like milling lines) there will 
> still be plenty of area along edges for transfer to the 
> aluminum. If you thicken up the copper then the only real 
> critical area is the surface below each FET. 
> 
> The big problem areas are near where the copper is sheered. 
> Any bend or burr there can hold the copper off the aluminum 
> or even warp the whole piece. Better be sure they saw it and 
> not sheer it, or be prepared to do some machine work. 
> 
> By the way, the FET's will have a tendency to blow up on 
> higher bands using Motorola's original layout.... especially 
> if you let the drain voltage exceed 55 volts. Be sure to 
> look at the drains right at the FET's with a scope and watch 
> the peak voltage!!! The boards I played with had a nasty 
> third harmonic resonance that created some pretty high 
> voltages. 
> 
> 73 Tom 
> 
> 
> 
> 
>