cold plate

You can probably find a full solution off the shelf from Omega, for small dollars. There are many other suppliers. I've done this to control a kiln. Look for PID Controller.

Joe Gwinn

Put the DUT in a (probably stainless steel) vacuum flask with the top off and pump in cold (or hot) air fairly slowly through a silicone rubber tube going all the way to the bottom. Bung up the open top with some sort of insulating wadding. This gives a nice even temperature.

I've done this with heating and it works well - an advantage is the low power requirement, so pretty safe to leave on for a long time. I used a 'water' cooled power resistor with a tube through the middle for the heating and a fish tank bubble pump for the air.

For the cold/hot air supply, a miniature vortex tube may be useful. I haven't tried this, but it sounds much more fun than Peltier devices, which is the most important thing.

Now, who will invent a MEMS vortex tube with pump for on-chip cooling?

Cheers

--
Syd

The open-circuit Seebeck effect voltage *opposes* the applied voltage, like back-EMF on a motor.

They can run away thermally, but for different reasons.

--sp

--
Best regards,  
Spehro Pefhany 
Amazon link for AoE 3rd Edition:            http://tinyurl.com/ntrpwu8 
Microchip link for 2015 Masters in Phoenix: http://tinyurl.com/l7g2k48

Quite right, thanks. The other point, that voltage control is very much inferior and can destroy the device in bad cases, is true.

Cheers

Phil Hobbs

Ha, I hacked something like that from our diode laser controller.

(not a very good picture.)

I don't think it would do your -20 C and I limit the max T to ~65C. Something around 0 C is the lowest it will go, depending on load and the room air.

There look to be lots of TEC's online, you just need the right search term... my google-fu is not working today.

George H.

Well (before I knew any better) I used voltage control for our TEC cooler. It works just fine. The whole thing is in a feed back loop and (I guess) that takes care of any weirdness. I certainly wouldn't call it "very much inferior". Then again, I haven't look at the loop dynamics in any detail.

George H.

Hey, how about the smaller unit from the web site you found.

$2.5k

George H.

I don't want to make an engineering/machining/programming project out of this. If I can't buy something all built, it's not worth doing.

I understand. there was a smaller $2.5k unit on the website you found. That seems like a reasonable number to me.

George H.

That's my point. In a feedback loop, the quantity being regulated will be controlled. So unless there are effects happening faster than the feedback loop...

Still, a current source is very easy to build.

--

Rick

I was looking for a larger TEC once and didn't find much on eBay other than the standard size 12706. I found one a bit larger that I wanted to test and the vendor couldn't actually get one in my hands after claiming to have shipped twice. So I gave up on it.

--

Rick

Feedback covers a multitude of sins, especially at low speed. If you're using the TEC near zero delta-T (as I usually do), it's no biggie. However, if you're trying to get a big delta-T, voltage control will put a _huge_ current in at turn-on, which can melt the TEC. (Stacked TECs are especially vulnerable to this.)

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

That might be the answer. It has manual controls and a serial interface, and it's all done. We could stick little boards to the plate with some of that iccky tacky Bergquist or Laird gap-pad stuff.

I'm hoping it will settle faster than the big blue chamber.

This is the sort of thing we need to do:

That's amazing for a 22M thick-film resistor.

I guess you (I) might worry about how the gain changes (inside the loop) when the TEC is at different temperatures. Does some delta V always give about the same change in heat input. Empirically the controller I made works fine from a few C to 65 C. Though as I said I didn't look to close at the dynamics. It did get a little squirrelier* (sp) at high temperature, but I've always assumed that's because it's working against a large delta T and room temperature air currents and such have a bigger impact.

George H.

*Well I don't monitor the TEC, but a suragate for the temperature which is the wavelength of a diode laser.

It's the turn-on transient that's lethal in a voltage-driven TEC set to run at high delta-T. The open circuit voltage changes by volts.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Hmm, Well maybe I'm just lucky. I've certainly run the TEC's starting from both the hot and cold temperature to full power in the other direction. Maybe my maximum voltage limits the current, or the power opamp has a maximum current. (OK I had to go look at the schematic, I drive the TEC from +/-15 V with a opa544.. which has internal current limits at 2A. perhaps that's my savior.)

George H.

The old-school science way to work this, is with a tall Dewar flask. Put a little liquid nitrogen in the bottom, and you establish a gradient in temperature- room temp at the brim, much colder down below. Then, just lower the device under test, slowly, until it hits the target temperature. There's a little more to it, you want some baffles to keep the cold gas from convecting.

If you want hot, though: maybe just a thermos and let the initial charge of liquid N2 boil off, then watch temperature rise. When it doesn't rise fast enough, turn on a heater.

Condensation of moisture from the air will be an issue, but a little liquid nitrogen will supply dry gas to purge the system at start.

I want something like this, maybe smaller,

Well, they do cost $3. That's around $270k per pound, if I haven't made a mistake somewhere.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net