No practical experience but
Ups ] [ Post
Followup ] [ CPU-Central Message Board ] [ Home ]
Subject: No practical experience but
Date: 4/29/2001 10:34:18 PM (GMT-7)
IP Address: 18.104.22.168
In Reply to: Peltier posted by
here are some fundamentals:
The Peltier effect simply occurs at a PN-silicon junciton, basically a diode. What happens is that heat is removed from one side and deposited at the other side at a rate that is somewhat proportional to how much power you're running through the Peltier device.
In other words, the Peltier device will create a fairly constant temperature differential.
So, it is possible for one side to get lower than the ambient temperature without violating thermodynamics, becuase that lost heat merely appears on the other side. Actually, though, the effect of dumping a lot of power causes the Peltier to create some heat of its own.
(1) The Peltier ONLY creates a temperature differential. If, to use a few random numbers, the Peltier is capable of creating a 20C temperature differential, and your cooling solution is only keeping the "hot" side at 80C, the "cold" side will run at 60C. Not exactly the desired result.
(2) There will be a lot of heat to remove. A standard heatsink may work with a low-power Peltier, but you will need to get a lot of heat out of your case fast. This = much fannage.
(3) If you keep the "hot" side adequately cooled, the "cold" side can achieve very low temperatures. Thus you may run into condensation problems -- which means the inside of and area around the CPU socket will need to be sealed somehow, some sort of spray-foam for example.
Again, I haven't used Peltiers at all (I've held and examined one, that's all), but I've heard it expressed previously that Peltiers are an ongoing science experiment, and generally only work well in conjunction with water cooling.
FollowUps | Post
E-Mail Sender |
Maximum of 100 messages displayed.
Post a Followup
Ups ] [ Post
Followup ] [ CPU-Central Message Board ]