Pelt and PSU

cmfic · 08-27-2002, 12:49 AM

I have a 226 watt pelt from the "Danger Den" and their PSU which they say should be perfect for the pelt. I do know that powering pelts @ 75% to 80% is where life begins (pardon the stupid phrase) the pelt is rated @ 226 watts, 15.2 v max and 24 amps. Now it is my understanding that pelts will take as much as they need or can get to work at their best performance, ( I don't know where I read this but I did ). The volts are adjustable up or down to a slight degree. Should I set them for the 12v and leave it at that or what are your suggestions. Thanks,
cmfic (John )
PS. The hot side will be cooled by #ROTO's "Big Brother" chiller which kickes ass. I am using the Danger Den Maze-2 block on a XP 1900+. If I could afford it I would freeze this chip to absolute 0 which is -273' Kelvin(spelling) sorry. With my current setup I beleive I can hit a -30'c on the core. Please correct me if I am wrong. Thanks again,
cmfic (john)

Brians256 · 08-27-2002, 12:17 PM

Running peltiers can be done either with current or with voltage, but it turns out to be the same thing because (although they are non-ohmic devices) controlling voltage means that you control the current.

Having said that, you should focus on controlling voltage because that is what all the specs are based upon. As a rule of thumb, run the pelts at 75% of rated max voltage for optimum efficiency. Running more electricity through the peltiers will generate more waste heat to be dumped, but it won't cool the CPU any better. Of course, a peltier manufacturer will have a voltage-curve that tells where the peak efficiency is, but 75% is a good starting estimate.

Also, don't be misled by the max delta-T spec given by peltier manufacturers. The actual temperature differential between hot and cold side diminishes as more heat needs to be moved. It is just like an car that slows down as you go up a steeper hill.

BTW, peltiers only work down to about -60C. Electron mobility (which goes down as temperature goes down) in the current designs limits their effectiveness. This (along with some other reasons) is why some chips actually start failing to work at low temperatures. On a side note, I just read about how copper vias in modern 0.13um chip designs are starting to pull away from the M1 contacts because of thermal expansion and contraction. Temperature is important, and you can't just assume that things will work well at ultra-low temps.

08-27-2002, 12:49 AM	#1
cmfic Banned Troll Join Date: Aug 2002 Location: North Central Washington State Posts: 26	Pelt and PSU I have a 226 watt pelt from the "Danger Den" and their PSU which they say should be perfect for the pelt. I do know that powering pelts @ 75% to 80% is where life begins (pardon the stupid phrase) the pelt is rated @ 226 watts, 15.2 v max and 24 amps. Now it is my understanding that pelts will take as much as they need or can get to work at their best performance, ( I don't know where I read this but I did ). The volts are adjustable up or down to a slight degree. Should I set them for the 12v and leave it at that or what are your suggestions. Thanks, cmfic (John ) PS. The hot side will be cooled by #ROTO's "Big Brother" chiller which kickes ass. I am using the Danger Den Maze-2 block on a XP 1900+. If I could afford it I would freeze this chip to absolute 0 which is -273' Kelvin(spelling) sorry. With my current setup I beleive I can hit a -30'c on the core. Please correct me if I am wrong. Thanks again, cmfic (john)

08-27-2002, 12:17 PM	#2
Brians256 Pro/Staff Join Date: Oct 2001 Location: Klamath Falls, OR Posts: 1,439	Running peltiers can be done either with current or with voltage, but it turns out to be the same thing because (although they are non-ohmic devices) controlling voltage means that you control the current. Having said that, you should focus on controlling voltage because that is what all the specs are based upon. As a rule of thumb, run the pelts at 75% of rated max voltage for optimum efficiency. Running more electricity through the peltiers will generate more waste heat to be dumped, but it won't cool the CPU any better. Of course, a peltier manufacturer will have a voltage-curve that tells where the peak efficiency is, but 75% is a good starting estimate. Also, don't be misled by the max delta-T spec given by peltier manufacturers. The actual temperature differential between hot and cold side diminishes as more heat needs to be moved. It is just like an car that slows down as you go up a steeper hill. BTW, peltiers only work down to about -60C. Electron mobility (which goes down as temperature goes down) in the current designs limits their effectiveness. This (along with some other reasons) is why some chips actually start failing to work at low temperatures. On a side note, I just read about how copper vias in modern 0.13um chip designs are starting to pull away from the M1 contacts because of thermal expansion and contraction. Temperature is important, and you can't just assume that things will work well at ultra-low temps.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)