TEC testing with variable DC PSU

[Ls7corvete]

I have been interested in using TECs at different voltages for some time now.http://www.overclockers.com/tips1211/ I feel that kryotherm and some simple calculations can do wonders for TEC cooling but I do not like to put much faith in this. Peltiers do not have the well documented nature that the rest of cooling has. Simple questions like how much better is a 5002 than a maze4? or whats better a WCed TEC at 5v or an air cooled one at 7v? Or how much does the large heat area of a TEC improve performance?

Anyways, many unanswered questions out there. Alot of bad info as well, some blatantly wrong and the mistakes are even ignored.http://www.xtremesystems.org/forums/...ad.php?t=67927

For now though I want to test various heat loads on the popular TECs running at a range of voltages to determine dT and amperage/wattage drawn. It is these questions that are crucial to planning a TEC setup and very often is where many people fail at setting up their systems. I am looking for qualitative data here not quantitative, ie, I want to know whats going to be best to cool XXX system, not how much better its going to cool down to +/-.001c of error. People need to look at real test results from a typical system and that is lacking right now.

Anyways enough rambling on what I need to know is what is a good price for a 0-60V / 30A Sorensen DCR 60-30B minutes from my parents house, tested. I will most likely just do some testing and pass this on so I need to make sure I don't end up too far in the hole with this. Also any help finding devices to get some good data on this would be nice. I plan to use the PSUs to read the voltage and amperage on the TEC and the heating elements, possibly thin film resistors for the heating. Maybe a fluke for temp reading, is it possible to use a switch to read multiple RTDs from one meter?

I will start something at WBTA if that is best, im waiting on registration. So who is interested in seeing the results? What do I need to make this happen? And who can help me out with this subject?

More plans for this later... right now I need to know about that PSU.

[laxman]

I totally agree with you on the bad information and misconceptions about TECs. In fact right now I a creating a site that right now has a handful of articles on TEC. I will release the site when I edit my articles and have written up a procedure for my testbench.

Onto what you said about testing, I do not think that undervolting really helps with peltiers because the the Dtmax is also a function of current. So a simple way to model the Dt from water inlet is. Dt= C/W(Qcm)+Dtm(Q/Qcm-1). Both Dtm and Qcm obey Dtm(I)=Dtmo*(I/Imax) and Qcm(I)=Qcmo*(I/Imax). So when you put those into the DT equation you get that DT(I) increases linearly. Meaning you should run peltiers at the max voltage no matter what. Although what throws this model off is the fact that TECs become more effcient as I/Im decreases. I would geuss that your testing would predict that around I/Im=.8->.9 . If we assume that their is some optimal voltage/current to run the peltier at, that voltage and current will change with inlet temp, C/W, Qcm (at max voltage), and Dtm (at max voltage). To say the least that makes things interesting. So I think that the best thing to if would be the optimal voltage, and Qcm for any waterblock with given C/W.

I will help you in any way I can because I am in the same boat as you.

[bigben2k]

Originally Posted by Ls7corvete

...
Anyways enough rambling on what I need to know is what is a good price for a 0-60V / 30A Sorensen DCR 60-30B minutes from my parents house, tested.
...
is it possible to use a switch to read multiple RTDs from one meter?
...
I will start something at WBTA if that is best, im waiting on registration. So who is interested in seeing the results? What do I need to make this happen? And who can help me out with this subject?

More plans for this later... right now I need to know about that PSU.

Google shows a going price of ~$750 (I know, it doesn't help).

Yes, it's possible. Fluke makes both a manual and an automatic switch (scanner).
See the specs: http://wbta.us/bigben2k/fluke2180a_2...190a_Y200X.pdf

Your account is active (not sure why it wasn't). Send me a PM so I can open up a forum just for you. Let me know if you want it as read-only, or open to all.

[Ls7corvete]

Originally Posted by laxman

I totally agree with you on the bad information and misconceptions about TECs. In fact right now I a creating a site that right now has a handful of articles on TEC. I will release the site when I edit my articles and have written up a procedure for my testbench.

Onto what you said about testing, I do not think that undervolting really helps with peltiers because the the Dtmax is also a function of current. So a simple way to model the Dt from water inlet is. Dt= C/W(Qcm)+Dtm(Q/Qcm-1). Both Dtm and Qcm obey Dtm(I)=Dtmo*(I/Imax) and Qcm(I)=Qcmo*(I/Imax). So when you put those into the DT equation you get that DT(I) increases linearly. Meaning you should run peltiers at the max voltage no matter what. Although what throws this model off is the fact that TECs become more effcient as I/Im decreases. I would geuss that your testing would predict that around I/Im=.8->.9 . If we assume that their is some optimal voltage/current to run the peltier at, that voltage and current will change with inlet temp, C/W, Qcm (at max voltage), and Dtm (at max voltage). To say the least that makes things interesting. So I think that the best thing to if would be the optimal voltage, and Qcm for any waterblock with given C/W.

I will help you in any way I can because I am in the same boat as you.

Heh, indeed you are saying just what I am thinking as well as some other things like the change inperformance of the TECs at diff temps, ie, the example of the 320w peltier that is only capable of 320w at 50c, at 25c its a 230w IIRC. There is enough variables that I no longer feel plugging in numbers to the calculations is acceptable.

Optimal is hard to say though as it is very much determined by preference. Lower wattage not only brings less heat to deal with but less PSUs and less radiator space.

Thanks for the input, I will continue discusion at WBTA, BB is setting a forum up for me there.

[Ls7corvete]

http://wbta.us/forums/index.php?board=36.0

[Ls7corvete]

I got the PSU, would be nice if someone would be kind enough to lend some stuff for testing. I have some but there is much more to test out there.

[bigben2k]

What do you need?

[Ls7corvete]

Originally Posted by bigben2k

What do you need?

TECs and blocks to test, popular ones to do a comparison.

Still need:
swiftech block- 5002, others? Modified 6002 would be neat.
I would like to test a Winstch Labs unit and their 437w TEC if anyone has one to lend
172w TEC...others? some 40mm to test would be awsome.
a maze 2-2 would be nice if anyone still has any, running two TECs at low voltage interests me obviously.

I would like to test as many TECs as possible, the block data is pretty well document the only issue being the larger die size of the TEC compared to the cpu/die sims used in other tests.

Some test equipment would be nice, temp reading and something to control/measure the heater die. I am not against buying some stuff off of knowledgeable members.

For now though I will test with a 5.25" temp reader until I have everything setup and enough equipment to do a good test.

[bigben2k]

As a WBTA member, you have access to a Swiftech 5002 and 6002 (among others), but I don't have the TEC setup (cold plates) for either. Of course I'd have to ask that you don't modify them in a way that would prevent other testers from using them.

[Ls7corvete]

Originally Posted by bigben2k

As a WBTA member, you have access to a Swiftech 5002 and 6002 (among others), but I don't have the TEC setup (cold plates) for either. Of course I'd have to ask that you don't modify them in a way that would prevent other testers from using them.

Thanks, I will keep that in mind.