The Cooling Solution

[RoboTech]

IMO incorporating a water chiller into the test bench to maintain constant temperature is a must. Even using a huge reservoir (several 100 gallons) wouldn't hold temps for long term testing and I doubt creating a workable fan/radiator setup would be practical (size, response time and ultimately cost).

Refrigerated Circulating Water Baths are readily available but expensive. This could be the single most expensive piece of equipment if purchased new. Expect to pay $2,000 ~ $5,000 US.

http://www.vwrsp.com (VWR Scientific Products)

http://coleparmer.com (Cole-Parmer)

VWR and Cole-Parmer both sell PolyScience chillers under their own name. Other popular brands are Haake, Lauda, and Julabo. I personally use VWR and Haake chillers in the lab where I work. Haake chillers have an excellent reputation but are also more expensive.

Most lab chillers incorporate a refrigeration system and PID control, while a few use TECs. I would highly recommend the refer units.

Most chillers are available with either analog or digital controllers. The analog units are less expensive but typically can only maintain a +/- 0.2 temp. Some of the better digital controls claim +/-0.01 temps.

Even the small lab chillers can handle around 200 watts~300 watts @ 20 deg. C.

Check the sources above and then keep and eye on E-bay...

[pHaestus]

You CAN get water temps to within a 0.3C tolerance over an entire test with a radiator and fans. I have done it. It is a bit of a pain because it requires a lot of fiddling with the fan speed, waiting a minute, fiddling again, and so on. Still it is possible and could in theory be automated.

What is the real problem with rads and fans is the sensitivity of the system to room temp changes. ACs wreak havok. Changes from summer to winter are an issue. Over an 8 hour cycle or so the room temp is hard to keep stable. A chiller deals with all of these. In principle, choosing water temp set point suitably warm makes it at least POSSIBLE to work all year round. You may actually need to switch to a poorer radiator if it gets really cold in your basement in the winter.

Anyway there are some things that preclude success (using a bucket to measure flow rates, using a compunurse to measure temperatures). Using a radiator as a cooling source is NOT the kiss of death but it will quickly make testing into a chore best left to trained monkeys.

Just wanted to throw that in there.

[bigben2k]

pHaestus: given that you've actually done it, do you foresee an automated solution as being capable of maintaining +/- 0.1 deg C?

It's something I'd like to pursue myself.

If I get into some cash, I'll dump the idea for a chiller, and just be done with it. Maybe someone else can pick it up.


For those who aren't in the loop: I'm building an 8' by 8' by 8' office/lab in my garage. Given the hot weather here in Texas, I'm going to have to slap an AC unit on it. That of course, may cause major ambient temp fluctuations. Of course I could simply restrict my testing to two or three seasons out of the year...:shrug:

[ThyKingdomCome]

I have a chiller bb2k. Perhaps we could work something out. Email me before attempting to purchase one.