Heat sink temperature

[xuxxx679]

Hi,

Here is a brief graph of a test that I've been doing. The material is copper and system is under natural convection. At first, I thought the temperature measured at thermocouple 2 will be higher since the thermal resistance from the bottom of the base to tc2 should be much smaller than to tc1 (larger cross-section area, less distance). But the test actually shows that the tc1 measurement is higher. Could anyone provide some thoughts on why?

Thanks,
Danny

 

[3DDave]

Probably the heat is not confined to the path you expect.

 

[drawoh]

xuxxx679,

Have you reviewed all your assumptions? Are there any sources of flowing air you did not notice? Are you sure your heat is applied as shown? Have you tested your thermo-couples?

Have you tried measuring the outside face at thermo-couple[ ]1? Heat is being dumped at most of the outside faces. It is logical that the internal material should be hotter.

--
JHG

 

[btrueblood]

What is the difference in temperature, what type are the thermocouples, how are you measuring them (i.e could it be explained by measurement error).
Oh, and did you do a calibration check of the two thermocouples (ice/water bath or even just ambient un-powered comparison).

 

[LittleInch]

Not sure which way around your dimensions are (W x ? x ?).

Also how are your TCs fixed? Surface mounted is not easy to get good contact.


Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[IRstuff]

They're thermocouples, BTW, not thermal couples

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[itsmoked]

Of course TC1 will be higher.

TC2 is only half touching the heated surface the other half of it is another 'cooling fin'. Place some tiny amount of insulation over it and see the correction.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[xuxxx679]

Thanks everyone for your response, the accuracy of the thermocouples should be fine. @itsmoked, I put a small amount of insulation over it and the temperature did go up but it is still lower than TC1. Is there a way to attach the TC so I can make sure the TC is measuring the surface temperature, not a combination of surface temperature and air temperature?

 

[3DDave]

It's almost like a photo would help diagnose this.

 

[chicopee]

With the insulation over Tc2, what was the difference in temperature and what was the difference in temperature without the insulation? and how is the heat applied?

 

[btrueblood]

"the accuracy of the thermocouples should be fine"

Yup. Should be. And so should the accuracy of the measuring equipment. Assume what you want, but did you ever do a power-off reading of the temperatures and confirm both tc's were reading the same? Did you look at what happened at the moment you turned the heater power on?

I once got good and spoofed by a ground loop, where the low side wire of one thermocouple brought stray power from the unit in test back to my data recorder, and injected just enough stray voltage to give me about a 10-20 F drop in that TC reading. Calibration and power off checks did not catch the problem, it only manifested after the TC was mounted and power turned on. I've seen weirder things happen too, like where sneak shorts in the tc leadwires or connector plugs created false couples.

But whatever, you're the man in charge and you say they are fine. Try mounting Tc1 into a drilled hole, or epoxy-ing it to the surface with a good filled epoxy. The next thing to do is think about what 3DDave said, maybe dismount your heat sink and see if your thermal contact paste is smooth and uniform across the interface, i.e. you aren't biasing the heat path away from Tc 1.

 

[btrueblood]

Oh, and another thing, are you using independent meters to read the tc's? Have you tried switching devices?