View Single Post
Unread 02-17-2004, 06:21 PM   #20
Gooserider
Cooling Savant
 
Gooserider's Avatar
 
Join Date: Jun 2003
Location: North Billerica, MA, USA
Posts: 451
Default

Agreed, all the info that I've ever seen says that in a reasonably good flowing system (over about one quart (Liter) per minute) the temperature difference going around the loop is less than 1*C.

If you read some of BillA's articles, you will note that he makes a big deal of this, as this is one of the factors that makes block testing so very difficult. The temp. differences are very small, which makes them hard to measure, and makes any error highly problematic in terms of it's impact on the test numbers.

BillA also says (and I agree) that there is no real difference in block temps caused by component sequence - IOW,
Pump -> Rad -> Blocks will give the exact same CPU temps as Pump -> Blocks -> Rad. This seems counterintuitive, but if you think about it, the logic makes sense.

1. There are two main heat generators in the loop, the pump and the blocks, and one cooler, the rad.
2. Assume for arguements sake that each generator adds one unit of heat to the loop on each pass. Since the system is at equilibrium, the rad must dump two units of heat per pass. (exact numbers aren't critical here)
For each of the following, I'll add or subtract the heat (H) through a cycle
3. Case one - (pump in already has 1H from blocks) -> Pump +1H = 2H -> Rad -2H = 0 -> Block +1H = 1 -> pump in...
4. Case two - (Pump in H = 0) -> Pump + 1H = 1 -> block + 1H = 2 -> Rad -2H = 0 -> pump in...

So, not much difference either way. We do know that given the same components, more flow = better cooling, so it appears that the best way to get the most cooling is to maximize the flow rate. This appears to suggest that the best bet is to arrange the parts in such a way as to give the straightest and shortest loop possible.

Since per the standard flow resistance charts, bends add considerably more resistance than straight runs of the same size, and the sharper the bend the greater the flow resistance, the inevitable balance that has to be drawn between 'short' and 'straight' should probably be tilted somewhat towards straight, but this is a more debateable issue.

Gooserider
__________________
Designing system, will have Tyan S2468UGN Dual Athlon MOBO, SCSI HDDS, other goodies. Will run LINUX only. Want to have silent running, minimal fans, and water cooled. Probably not OC'c
Gooserider is offline   Reply With Quote