:: SMD thermal test platform 2

TEST 2. same TO252 12R. however a backing plate of brass is soldered onto the PCB first. the brass is 0.64mm thick (equivalent of 0.16mm copper i think). the brass platform is approx 25mm wide, 50mm high with TO252 sitting on 1 of the long end)

top tabs showing some thermal stress as it cools (too much solder :p)

it is quite hard/impossible to get the solder to stick uniformly throughout the entire under surface of the brass and PCB copper, so i opted to "overheat" the iron and make the solder watery on the sides and let it all seep in as much as it could. i have purposely allowed a top left corner to be improperly solder to see the effects when the thermal pics are captured.

slightly cleaned up and ready to pump some voltage !

with this soldering of a brass heat tab, i think i am beginning to see why a reflow oven is an absolute perfect tool together with soldering paste to get SMD soldered in the right way.

12R pumped with 7.1v (4.2w). even with more power, having a small sinking brass plate lowers the temperature by over 15degrees C.

i think it is time i make a better power supply to cope with these thermal tests. the LM317 i chopped up hastily and plonked into a plastic food "tupperware" is having problems with voltage stability :(

but this is not the end. the actual aim is to shoot a thermal gradient of the PCB and see how much heat it spreads over the intended area of the brass tab. 

this is SMD version 3. 75x15mm (0.8mm thick brass), over it another 50x15mm brass (same thickness). stacked like a 2 layer pyramid. the rear thermograph speaks volumes.

in this version. 12R is loaded at 8volts, pushing 5.3w. top side the package reads 75degrees C.

without even going further, based on this sad scenario. either choose to use full on heatsinkable TO220 variants or, mod these TO252 to mount on heatsinks direct via soldering to a brass plate proxy. for this, a different sized brass bar is required, to slightly match the pad print of the SMD and carry approx 2mm thickness.

when finished, the DPAK will look more like a brother of a RF transistor.

so for now, with additional brass heat sink. it is possible to add an additional few more watts to the operating thermal envelop without using expensive copper plating. and with the spreading, there is now an even larger conducting surface on the reverse side of the PCB which could be fixated to an actual finned heatsink block, although having to pass through the PCB itself is detrimental, the ability to take off any heat is a plus.