Different temperature on both side of same heatsink?

[fivefish]

So I have (2) power transistors mounted on a heatsink (fin type).
Transistors are bolted on the flat side (let's call this component side), with the fins naturally on the other side.

I measured the temperature using those laser thermometer (eBay/china). On the component side of the heatsink, between the 2 transistors, I measure 58*C.   

On the fin side, measuring the same point at the opposite side, I'm measuring 36C. 

That got me wondering why the difference, when it's the same piece of aluminum, same spot (just opposite sides), will there be a big temperature difference like this.

[mikeselectricstuff]

heatsinks can be fairly shiny, even if black anodised, so you are probably seeing reflected IR.

[electr_peter]

I measured the temperature using those laser thermometer (eBay/china).

IR thermometers show accurate results only in ideal conditions, do not trust actual temperature reading too much. As mikeselectricstuff mentioned, temperature reading depends of reflectivity/emissivity of material as well as surrounding environment temperature. Also, distance to object matters, as it "averages" the temperature of wide spot.

Use more robust measurement equipment to be sure - try doing it with your fingers or thermocouple. I would think there is at most few degrees of difference on the heatsink.

[Andy Watson]

Wouldn't you expect the side where the heat is dissipated to be cooler then the heat source?

[fivefish]

Thanks mike and peter... you're right. I'm using the IR thermomenter without knowing it's limitations.

I grabbed my Fluke thermocouple and did the same measurement, now fluke meter reports approx. 52C on both sides (+/- 1C variation) of heatsink.

How do IR thermometer works anyway? It's a bit voodoo for me.

[ez24]

heatsinks can be fairly shiny, even if black anodised, so you are probably seeing reflected IR.

ie it is hard using these IR meters accurately.  If important I think you can paint on some flat black muffler paint and take measurements off of that, maybe fingernail polish.  Maybe roughen up the heatsink with sandpaper to reduce reflections.  But anyway physics trumps these IR meters.

[fivefish]

Wouldn't you expect the side where the heat is dissipated to be cooler then the heat source?

Yeah, that was my initial reasoning too... but as mike & peter said, it could be the IR thermometer giving me the wrong readings.  And theyre right.

Using a proper thermocouple showed equal temp readings on both sides.

[SeanB]

Best is to first make a black mark using a dry wipe marker on the measuring spot. That way the emissivity of the spot is the same all round. I noticed that simply by moving the sense spot on a heater from the edge, where it is well anodised aluminium plate, to the top where it is a coating of a PTFE lining, firmly bonded to the plate. Around a 5C difference in 1cm of movement. Even a difference between putting the pocket probe in the hole and using the IR on the black oxide surface of the brass heater near it.

[ez24]

Best is to first make a black mark using a dry wipe marker on the measuring spot.

   thanks

[mikeselectricstuff]

I measured the temperature using those laser thermometer (eBay/china). On the component side of the heatsink, between the 2 transistors, I measure 58*C.   
On the fin side, measuring the same point at the opposite side, I'm measuring 36C. 

Classic case of not trusting readings that are clearly suspect and need a reality check - a quick prod with a finger would have revealed the misleading figures.