:: DIY thermal tests (SMD /PCB etc)

**update. starting of stage 2 thermal tests. ordered 1 compound X thermal interface silicone rated @ 1.5w/mK. tests will be on ability of compound(s) to transfer heat to heatsink.
the new heatsink is prepared simply. some bits of sanding down, and black acrylic paint coat, and some more finer rub down. at dimension of 120mmx150mmx28mm, fin depth 24mm, fin thickness 1.4mm (corrugated) x 14 fins.

using ESP's xls calculator, we estimate the thermal rating to be 0.82°C/w. with an active fan going with approximated ratio of 0.4, the heatsink could be looking at 0.33°C/w

taking general specs from a typical 92mm dc fan, with a CFM capacity of 20-40, lets say we get an approx LFM rate of 200-500. we insert this into this next calculator, which is easier to use.

10LFM = 3.37°C/w
100LFM = 1.07°C/w
250LFM = 0.67°C/w
350LFM = 0.57°C/w (approx NIDEC beta SL 12v)
500LFM = 0.48°C/w (now the faster the fan, the lesser the total effect)
1000LFM = 0.33°C/w (now the faster the fan, the lesser the total effect)
however it must be noted this is for bonded fin. nonetheless can be used as a nice guide as there is no other online calculator this simple that has LFM direct to °C/w number crunch (big thank you!).

ambient air temperature is now 28C. this heatsink block is really sucking the heat form the environment, 4 below. being not a thermodynamic expert, i dont really know what the -4 number correlate to.

for this test and comparison. heat source will still be TO-xxx based resistor packages. they will be mounted on the sink using different thermal interface compounds. however this time, i may have a good power source ready to pump some precise amount of heat into these suckers.

the thermal interface medium to be tested are
1) normal ALTECO super glue
2) Silcoset 153
3) toothpaste (lion systema gum care)
4) Loctite 272
5) compound X
6) another loctite type super glue (401?)
7) adhesive KAPTON tape only
8) normal ALTECO super glue + full layer of KAPTON
9) Silcoset 153 + full layer of KAPTON
10) compound X + full layer of KAPTON

item 1 to 5 will feature a special test format where the ends of the package are separated from the sink by KAPTON tape layer. assuming the tape is manufactured well and has a uniform thickness, this thickness will be used as a "standard" so that all the interface medium achieve a similar STANDARD interfacing thickness throughout all test sets.

unfortunately i do not have a insulation tester, with this, i may be able to test the breakdown voltage of the above setup. but wait, i maybe able to do it as i may have an experimental low current 12kV source.

to be continued...

(previous entry)
And this is the final leg of the experiment. 

with some generic black matte spray paint, i prepped 2 sets of final test subjects. on the left side, is the PCB with the 2x 0.8mm dual stacked and soldered onto the PCB (the same unit as before, except that this time, it is painted) due to my PSU problems, i am only running this at 8.3v (5.7w).

with the painted surface, we can now see the heat emission gradient very easily.
the test sample on the right is a new variation. it is the idea of using a brass backing plate to stick the SMD onto an actual heatsink (120mm x 60mm x 22mm in dimension, 16fins of 1.5mm thickness with a base of 4mm). the brass plate is badly secured to the heatsink via just the use of generic thermal silicone (not even thermal grease). as you can see, there is a rather big step in temperature down from the brass which is soldered direct to the 12R device. the silicone layer is thick, and is quite thermally resistive.

the brass contact piece is 12.7mm x 30mm x 1.6mm (about double of the 0.8mm but 3x smaller area approx). the brass primary heat spreader carries a total surface area of approx 381mm2. the PCB version has a total surface area of approx 1125mm2. but with the alu sink block behind it, it seems to be dissipating heat 3x more efficiently (and guesstimating from the heat gradient, the heat load is too small for the ALU sink system)
i will definately need to make a more serious PSU and try to pump 15-20watts into these.

1 thing is for sure, the brass plating mod to an actual heatsink is working like a charm.

now i will need to construct a better PSU in order to try and pump these 2 very nice thermal samples with more heat.

Popular Posts