Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 beta 3/9/83; site aat.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!mb2c!uofm-cv!aat!sleat
From: sleat@aat.UUCP
Newsgroups: net.analog,net.physics
Subject: Re: thermal diodes
Message-ID: <323@aat.UUCP>
Date: Sun, 28-Oct-84 15:48:20 EST
Article-I.D.: aat.323
Posted: Sun Oct 28 15:48:20 1984
Date-Received: Mon, 29-Oct-84 03:11:50 EST
References: <3259@rabbit.UUCP>
Organization: Ann Arbor Terminals
Lines: 57


[Apologies if this is a duplicate.  Local evidence suggests the first
posting never made it out of this system.]

>I'm not sure that "thermal diodes" is the correct term, but I recall
>hearing a few years ago about the development of semiconductor devices
>that acted as heat pumps or refrigerators.  Pass a current through
>them and one end got hot, while the other got cold.  I think some
>company actually tried marketing a portable refrigerator using these.
>Does anybody know where I could get some specs on these devices, or
>buy some of them?
>-- 
>Jan Wolitzky, AT&T Bell Laboratories, Murray Hill, NJ; (201) 582-2998



You might try:

	Cambridge Thermionic Corporation
	445 Concord Ave.
	Cambridge, MA  02238
	617-491-5400

I have an old data sheet (ca. 1979) for their model 801-1029-01-00-00
"Ceramic Module".  Basic spec's are:

	Max Temp Difference	>60degC
	Max Current		9 Amps
	Nominal Voltage		0.7Vdc
	Hot Side Temperature	50degC (max?)
	Max Heat Pumping Cap	>3.3 Watts
	Max operating Temp	125degC

The unit is a rectangular block approx .2 x .3 x 1 inches, with two
18awg pigtail leads.  Heat is transferred across the .2 inch dimension
(i.e., from one .3" x 1" surface to the other).

The energy transfer rate, at a fixed current, is of course inversely
proportional to the temperature difference.  The given transfer capacity
of 3.3 Watts is at 9 Amps with a delta-T of 0.  From the rather cryptic
performance curves given, it looks like you get about 0.2W across with
9 Amps at a delta-T of 50degC.  Below 5 Amps, transfer rate is approxi-
mately linear with current.  Above that it begins to fall off (as
resistance effects begin to dominate, I suspect), the slope reaching
zero at about 9 Amps.  Above 9 Amps, the slope goes negative.

Nowhere on the data sheet does it mention that this is a semiconductor
device, though the leads are color coded for polarity and the 0.7V
suggests a semiconductor junction breakdown voltage.  I believe that
this exploits a phenomenon other than the Peltier effect, though my
recollection on such matters is hazy at best.  Any denizens of
net.physics care to elucidate?

Michael Sleator
Ann Arbor Terminals
{aatpdx, cbosgd, mb2c, psu-cs, uofm-cv}!aat!sleat