Using solid mass as a heatsink for MOSFETS

[n3lrx]

I have these heavy cast aluminum boxes the walls are about 1/2 thick, they will provide a perfect home for my current project. Originally I was going to mount the PCB's inside and mount a heatsink on top for the two FQAF11N40's in the final and do the same in another box for the FAQF9N50 in the modulator. But, I'm wondering if I could just bolt them directly to the box and use the box itself for a heatsink would this be safe? Right now I've got one of the boxes sitting on a radiator and although the radiator is way too hot to touch the box is relatively  cool, warm to the touch, but not too hot to hold. Would that be a good sign for them being suitable for a heatsink?

I've got 2 Thermaltake Volcano 7 CPU coolers (solid copper, finned heatsink) that I was going to bolt the temperature sensitive devices to but if the boxes they're mounted in will suffice I can use the Volcano's elsewhere. Or would it be best to use a finned sink with forced air cooling?

I thought about getting some of those really cool 'Heat Pipe' design heatsinks you know,  those ones that look like an engine with headers! But maybe that's a little overkill! Maybe they'd look really cool if I decide to add a lighted fan, neon lighting and flashing LED's  in the transmitter case it would look neat but I don't think I'll go that far! LOL

[WA1GFZ]

It all boils down to surface area: 130 degrees C temperature rise per square inch per watt. So if you have 10 square inches to dissipate 1 watt the temperature rise will be 13 degrees, If you have 100 square inches to dissipate 100 watts then the temperature rise will be 130 degrees. If you have 1000 square inches to dissipate 100 watts then you get 13 degrees of rise. I run about 2000 square inches of surface area in my 75 meter final if I remember. This is for avreage dissipation not peak. A 1000 watt final will have an average dissipation of 100 watts if built correctly.
Peak dissipation will be around 500 watts at modulation peaks

[WU2D]

Forced air over a dense finned heatsink is the current low cost approach with CPU's as with your Volcanos. Forced Convection allows you to use thinner denser fins too. Some of these exotic folded designs only work with air moving through them and are not effective at all with natural convection. I saw an interesting approach where the user attached the switching devices to the inside of the box as you want to do. They carefully attached a conventional extruded heatsink on the outside of the box directly in back of the devices using a low thermal resistance interface. Then they fashioned a shroud which held two fans which had some vents which drew air up and out of the heatsink fins. This lowered the thermal resistance of the heatsink drastically, yet did not compromise the environmental integrity of the enclosure since the entire assembly was external.

As you can see from the curves for some unknown heatsink, forced air works effectively to lower thermal resistance and allows a given heatsink to dissipate much more power. I assume that Frank hates fans for all of the right reasons - they fail, make noise and get dirty and must be cleaned!

Mike WU2D

[n3lrx]

Thanks,
I'm gonna try it as a heatsink. The box I have is pretty hefty, close to 1-1/2lbs of cast aluminum. Inside it is about 5 1/2"L x 3"W x 3"D with a screw down plate aluminum lid. I had the box setting on top of a 1500w Oil Filled Radiator, which was way too hot to touch. Yet even after leaving the box on the top of the radiator for over an hour it got hot, but by no means uncomfortable to hold. Within a few minutes off of the heat the box was cool to the touch.

I think this will work for the Modulator board. If I find that it gets a little too warm I can bring the box into work and machine the side where the devices are and mount an external heatsink. I have a few of these boxes but the TX board is a little longer than expected so it won't fit in one of the same enclosures, I do however have another enclosure that it will fit in only without as much mass. I'll use a heatsink on that one. The RF Board, Modulator, T/R switching, and small Audio Processor will all be inside the same case within their own sub enclosures. The entire case will be cooled with two 80mm box fans in push/pull for cross flow air. I'd also like to include a proper AM monitor as well with -/+ peak LED's and meters. I have planned on adding metering to monitor the power supply and Mod/Final current. 

I just got the PCB's in today, I havent had a chance to start on them yet since they arrived just as I was leaving for work. I did however manage to wind a few transformers and inductors during downtime. I'll have to work on the slobering when I get home.

Hopefully everything turns out right.. As I progress in the project I may change things around a bit. But as I have it planned the transmitter will be in a 12"H x 12"W x 7-1/2"D enclosure with the switches, power indicator, TX indicator, meters and LED's mounted on the face. Audio and power input on the bottom, and RF IN/OUT on the top. It's got flanges on the back so the case can be mounted to the wall behind the operating desk.

[n3lrx]

Man, Do me a favor.. Next time I say I'm going to build another one of these make sure I haven't forgotten to take my medication!! This is unbelievable soldering with a pair of tweezers, a lighted magnifying glass and a 15w soldering iron! It will be a bloody miracle if this thing works!

Anyone got a flow solder machine? Even with a magnifier these old eye's aren't as good as they used to be.. And to be honest, I've been 4 eyed as long as I can remember and they even with my google specs they never were that good!