peng@chaource.cs.wisc.edu (PENG) (12/15/90)
I hope this has not been asked before. On the December issue of Byte (p.132), there is a short article on an interesting cooling device, which can be mounted on a CPU and is able to cool the chip down to 0 degree centigrade. This device, according to Byte, is a Peltier effect device, which is "a thermoelectric cooling system based on the principle that passing a current between two physically connected, dissimilar materials produces cooling on one side and heat on the other." I am not quite sure if I know what this sentense is talking about. Can someone knowledgeable elaborate this a little bit? -peng BTW, in case you are interested, this device is used to push the 486 to 50Mhz. I am not sure what the 486 is originally rated? (33Mhz?) -- +-------- |C. J. Peng @ University of Wisconsin - Madison |ARPA: peng@cs.wisc.edu |UUCP: ...!{ihnp4,hplabs,seismo,topaz,etc.}!uwvax!chaource!peng
grege@gold.gvg.tek.com (Greg Ebert) (12/15/90)
In article <1990Dec14.213730.10078@spool.cs.wisc.edu> peng@chaource.cs.wisc.edu (PENG) writes: >I hope this has not been asked before. On the December issue of Byte (p.132), >there is a short article on an interesting cooling device, which can be >mounted on a CPU and is able to cool the chip down to 0 degree centigrade. >This device, according to Byte, is a Peltier effect device, which is >"a thermoelectric cooling system based on the principle that passing a current >between two physically connected, dissimilar materials produces cooling on one >side and heat on the other." I am not quite sure if I know what this >sentense is talking about. Can someone knowledgeable elaborate this a little >bit? > >-peng > >BTW, in case you are interested, this device is used to push the 486 to 50Mhz. >I am not sure what the 486 is originally rated? (33Mhz?) >-- The device is manufactured by Velox Computer Technology, and sells for ~$150 in quantities. See the Article in EDN (Dec 6. , 1990). The Peltier Effect is significant in an Antimony/Bismuth junction. The process is reversible: you can measure an emf when the junction is heated (its a few millivolts), or if a potential is applied across the junction, electrons will absorb (thermal) energy while traversing the junction. I know zilch about solid-state physics, but I think it's the result of differing work-functions of the 2 materials; ie, there is an 'energy gap' at the junction, and the electrons with sufficient thermal energy can cross the gap and get whisked away because of the externally applied emf. As one thermally active electron leaves town, another follows behind which results in a net flux of electrons with higher thermal energy away from the junction, hence cooling is observed. Add a bunch of these together and you can soak-up a few watts of heat. BUT, before all of you go charging out to by one of these electric ice cubes and a crystal, be aware that the logic surrounding to CPU probably wont be able to run at 50Mhz (well, actually 100Mhz). ----- Boycott redwood products ---------------------------- Recycle ----- "Thou shalt abide by The GNU Manifesto" ##### {uunet!tektronix!gold!grege} Register to vote, then ## | ## grege@gold.gvg.tek.com vote responsibly # | # # /|\ # Support high oil prices, waste tax $$ on war, evade domestic #/ | \# problems, and die young on foreign soil- Just say YES to Bush #######
mir@opera.chorus.fr (Adam Mirowski) (12/17/90)
In article <1990Dec14.213730.10078@spool.cs.wisc.edu>, peng@chaource.cs.wisc.edu (PENG) writes:
%% I hope this has not been asked before. On the December issue of Byte (p.132),
%% there is a short article on an interesting cooling device, which can be
%% mounted on a CPU and is able to cool the chip down to 0 degree centigrade.
%% This device, according to Byte, is a Peltier effect device, which is
%% "a thermoelectric cooling system based on the principle that passing a current
%% between two physically connected, dissimilar materials produces cooling on one
%% side and heat on the other." I am not quite sure if I know what this
%% sentense is talking about. Can someone knowledgeable elaborate this a little
%% bit?
That is one of the less known of the 4 thermoelectrical effects.
I remember having done some exercises on it on the second year. As
far as I remember, the Joule/Peltier ratio wasn't very impressive.
The materials must be semi-conductors.
French popular science magazine "Science et Vie" proposed its readers
to construct an experimental cooler based on Peltier principle early
in the eighties. The machine was able to produce ice cubes (< 0C), but
had to include a ventilator (good as a camping cooler).
--
Adam Mirowski, mir@chorus.fr (FRANCE), tel. +33 (1) 30-64-82-00 or 74
Chorus systemes, 6, av.Gustave Eiffel, 78182 Saint-Quentin-en-Yvelines CEDEX
jvincent@bnr.ca (John Vincent) (12/18/90)
In article <1990Dec14.213730.10078@spool.cs.wisc.edu> peng@chaource.cs.wisc.edu (PENG) writes: >I hope this has not been asked before. On the December issue of Byte (p.132), >there is a short article on an interesting cooling device, which can be >mounted on a CPU and is able to cool the chip down to 0 degree centigrade. >This device, according to Byte, is a Peltier effect device, which is >"a thermoelectric cooling system based on the principle that passing a current >between two physically connected, dissimilar materials produces cooling on one >side and heat on the other." I am not quite sure if I know what this >sentense is talking about. Can someone knowledgeable elaborate this a little >bit? > The devices mentioned are fairly common in the opto-electronics world and are usually used to cool laser diodes used in opto transmission systems. They are, in the most simple terms, a reverse thermocouple. These things come in a bunch of sizes. The biggest I have seen is about two inches square. They can freeze water, although they suck a lot of power. -- $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ $ John Vincent $ My opinions are my own, steal them and I'll $ $ Bell- Northern Research $ hunt you down and kill you. $ $ P.O. Box 3511, Station C. $ $ $ Ottawa, Ont , K1Y 4H7 $ Money talks, LOTS of money screams. $ $ (613) 763-8446 $ $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ $ Bitnet/NETNORTH : jvincent@bnr.ca $ $ UUCP : uunet!bnrgate!bcara315!jvincent $ $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
cramer@optilink.UUCP (Clayton Cramer) (12/22/90)
In article <1990Dec14.213730.10078@spool.cs.wisc.edu>, peng@chaource.cs.wisc.edu (PENG) writes: > I hope this has not been asked before. On the December issue of Byte (p.132), > there is a short article on an interesting cooling device, which can be > mounted on a CPU and is able to cool the chip down to 0 degree centigrade. > This device, according to Byte, is a Peltier effect device, which is > "a thermoelectric cooling system based on the principle that passing a current > between two physically connected, dissimilar materials produces cooling on one > side and heat on the other." I am not quite sure if I know what this > sentense is talking about. Can someone knowledgeable elaborate this a little > bit? > > -peng A Peltier effect device is essentially a thermocouple run backwards. (A thermocouple involves heating two dissimilar metals, and getting electron flow -- and therefore electricity -- from one metal to the other. Such devices are used in places where simplicity and reliability are more important than efficiency -- say a spacecraft). I can visualize the process by which heat causes electron flow from one metal to another (since different metals have different electro- negativities) -- for some reason, I can't picture how electron flow causes cold. -- Clayton E. Cramer {pyramid,pixar,tekbspa}!optilink!cramer Self-defense is the most basic of human rights. Lacking the right to defend yourself today can make it very hard to exercise any other rights tomorrow. You must be kidding! No company would hold opinions like mine!
valdes@bgsuvax.UUCP (oscar Valdes) (12/22/90)
In article <5070@optilink.UUCP> cramer@optilink.UUCP (Clayton Cramer) writes: >I can visualize the process by which heat causes electron flow from >one metal to another (since different metals have different electro- >negativities) -- for some reason, I can't picture how electron >flow causes cold. See if the following helps. The electrons, as they flow, carry heat away from the junction. In the thermocouple, heating increases the electron flow. In the Peltier device, the voltage produces the electron flow that carry the heat from one metal to the other. Thus, one metal gets hotter and the other gets colder. ******************************************************************************* Sooner or later reality bites you in the ass *******************************************************************************
whit@milton.u.washington.edu (John Whitmore) (12/22/90)
In article <5070@optilink.UUCP> cramer@optilink.UUCP (Clayton Cramer) writes: > >A Peltier effect device is essentially a thermocouple run backwards. > >I can visualize the process by which heat causes electron flow from >one metal to another (since different metals have different electro- >negativities) -- for some reason, I can't picture how electron >flow causes cold. You're SO close! Stand back and look at what you just said. Heat causes electron flow from one metal (carrier-rich) to another (carrier-poor), by diffusion from high carrier concentration to low carrier concentration. This diffusion, since it spontaneously creates a charge separation, requires energy. The heat that CAUSES this electron flow is used up thereby. Entropy still increases, because the charge carriers are spread out more evenly (diluted, as it were.) Osmosis is a VERY similar phenomenon. It is the difference in charge carrier densities at the two disparate temperatures that drives the complete circuit as a thermocouple, by the way. The back-EMF caused by the charge transfer would otherwise equilibrate at a relatively low voltage, and only if mechanical work (lifting one material off the other) is done on the charge would any significant energy be noticed. That is the principle of rubbing materials together to generate a static charge (and the original 'electricity' phenomenon.) All this is actually very close to common elementary electric experimentation that we all did in our youth. It is common nowadays to show the charge separation in the absence of rubbing, by the way (to clarify the fact that friction has nothing to do with it). The usual technique is to pull a piece of sticky tape off of another (identical) strip of tape. The sticky-side has different charge density from the slick side, so the two strips, once separated, have opposite charges. John Whitmore
userDHAL@mts.ucs.UAlberta.CA (David Halliwell) (12/25/90)
In article <5070@optilink.UUCP>, cramer@optilink.UUCP (Clayton Cramer) writes: >In article <1990Dec14.213730.10078@spool.cs.wisc.edu>, peng@chaource.cs.wisc.edu (PENG) writes: >> I hope this has not been asked before. On the December issue of Byte (p.132), >> there is a short article on an interesting cooling device, which can be >> mounted on a CPU and is able to cool the chip down to 0 degree centigrade. >> This device, according to Byte, is a Peltier effect device, which is >> "a thermoelectric cooling system based on the principle that passing a current >> between two physically connected, dissimilar materials produces cooling on one >> side and heat on the other." I am not quite sure if I know what this > >A Peltier effect device is essentially a thermocouple run backwards. >(A thermocouple involves heating two dissimilar metals, and getting . ^^^^^^^^ . not quite!!!! >electron flow -- and therefore electricity -- from one metal to the >other. Such devices are used in places where simplicity and reliability >are more important than efficiency -- say a spacecraft). > >I can visualize the process by which heat causes electron flow from >one metal to another (since different metals have different electro- >negativities) -- for some reason, I can't picture how electron >flow causes cold. >-- . A thermocouple circuit does involve two dissimilar metals, but the current flow is generated because the junctions are at different temperatures. However, current flow is not the governing electrical characteristic - voltage is. For each unit of temperature difference, the thermocouple will output a certain voltage, in a nearly linear fashion (at least over a small temperature range). For example, a copper- constantan thermocouple puts out about 40 microvolts per Celcius degree difference. Thermocouples are often used for temperature measurement, but because they just give a difference there must be an independent method of measuring the absolute temperature at one junction. The thermocouple then gives a voltage indicating the temperature at the second junction. The reference junction can be placed in something of known temperature - e.g an ice/water mixture - or measured using a thermistor, platinum RTD, etc. . Given this basic understanding of a thermocouple, a Peltier module basically reverses the process. By imposing a voltage in a thermocouple circuit, we can force a temperature difference between the two junctions. This in itself will not create cooling, but if we cool the warm junction we will still have the cold junction at the same offset below it. The heat added to the system through the electrical power consumption must also be removed from the warm side, but the cold side should be cooler than it would be if we didn't have a Peltier module and just tried to cool the component directly. . Years ago I worked in a lab where Peltier modules were used for freezing tests on soil samples. We used a water-cooled warm side, which kept the warm junction around room temperature. It was easy to get sub-freezing temperatures from the system. The module was controlled by a rather large mass of electronics, to get the stability needed for the tests. It made a rather impressive demo during show-and-tell at the open houses, to have a block of ice sitting on a module at room temperature showing no signs of thawing... Dave Halliwell P.S. Merry Christmas to all. Or Season's Greetings, etc. Choose one.
valdes@bgsuvax.UUCP (oscar Valdes) (12/28/90)
In article <2036@mts.ucs.UAlberta.CA> userDHAL@mts.ucs.UAlberta.CA (David Halliwell) writes: >. A thermocouple circuit does involve two dissimilar metals, but the >current flow is generated because the junctions are at different >temperatures. However, current flow is not the governing electrical Check again. You don't need two junctions to make a thermocouple. Heating a single junction will produce a voltage that depends on the temperature of the junction. ******************************************************************************* Sooner or later reality bites you in the ass *******************************************************************************
pierson@ggone.enet.dec.com (Dave Pierson) (12/29/90)
In article <6792@bgsuvax.UUCP>, valdes@bgsuvax.UUCP (oscar Valdes) writes, in part: > Check again. You don't need two junctions to make a thermocouple. >Heating a single junction will produce a voltage that depends on the >temperature of the junction. There are always two junctions. One may be the juction of the TC leads with the measuring instrument terminals. TC rigs always measure the difference between the two junction temperatures. To get an "absolute" reading, the temperature of one junction is measured by independent means. In "lab" TCs, this is typically an RTD type detector, mounted on a comon thermal mass with the screw terminals. The terminal block temp is then "added" either analog fashionm, or digitally. TCs are non linear over large delta t's, but reasonably linear over small ones. TCs are perfectly happy as either voltage or current output devices. Current (8)>>) practice treats them as voltage sources, due to the availability of good, hiZ amps. In the past, the were commonly used with microammeters. (Hi precision work used manually or automatically balanced bridges, which is voltage source mode.) (attempted graphics, Cu/Con couple...) _______Cu_____________o_(Cu)____>Amp hi / | /measuring junction |<-Ref temp \ | \______Con_____________o__(Cu)___>Amp Lo ^-the "other" junction (As i write this, i sit next to a logger with 16 TCs on it...) thanks dave pierson |the facts, as accurately as i can manage, Digital Equipment Corporation |the opinions, my own. 600 Nickerson Rd Marlboro, Mass 01752 pierson@cimnet.enet.dec.com "He has read everything, and, to his credit, written nothing." A J Raffles