开云体育

More power to you.

Remember in the "thermal cascade" the JUNCTION of the device is what counts.? ?Needs to stay well below 175 C at all times.? Mosfets have a problem with a localized heating effect as well that I don't well understand, has to do with local variations in sensitivity to bias.??

To get heat away from the junction there are several thermal resistors in series, and at every one, temperature is lost, keeping the junction HOT and reducing the efficiciency of moving heat OUT

Junction to tab -- this is pretty high in the TO-220 case, esp for MOSFETS that really were not built for this RF stuff.? ?Maybe 2 deg C/watt?? ? This means you CANNOT know the actual temp of the JUNCTION.

CAse to heatsink -- this is made worse by the mica (or other insulator).? ?So the heatsink is even further COOLER than the junction

heatsink to ambient room -- this is made worse by lack of air movement

Ambient room -- if the room is HOT things are even worse.

Measuring closer and closer to the junction is problematic, too!? The Drain is RF-HOT!? ?Lots of RF voltage/power there.? ?Potential for FEEDBACK if you couple something that carries RF away (wires, etc).? ?So could be a problem where the measurement....causes damage to the radio!? ? Not certain I would be so quick to put measurment devices there that go back toward the raspberry pi....

Putting a little thermistor to an external consumer temp measurement (think, the "external" probe of a consumer thermometer) next to the tab, separated by some insulating but thermally somewhat conductive glue or caulk.... then maybe making a common-mode choke ofthe sensing wire system by putting 5 or 6 turns of a coil of it over a pencil shape.... or putting some snap-on ferrites?? ?Might make this work and not CAUSE TROUBLE.? ?Dunno.? ?

Easier to measure the temp of the heat sink NEAR the mosfets, but that is considerably removed from the JUNCTION, too.? ??

It isn't a slam dunk simple project, I think.

Let usknow what works!

Gordon KX4Z