View Single Post
Unread 11-30-2002, 02:46 PM   #21
Skulemate
Cooling Savant
 
Skulemate's Avatar
 
Join Date: Jul 2002
Location: Toronto, Canada
Posts: 381
Default

UnloadeD, we are after the same thing. The reason you need to concern yourself with the pump is because it impacts the performance of everything else in the loop. The head loss of any object in the system will increase with increased flow. If you were to test the pressure drop for your block at several different flow rates you'd get a curve that initially started out quite low for small flow rates, but it would increase exponentially as the flow is increased (since the pressure drop is related to the velocity squared). Try to understand that the pressure drop for the waterblock is really a series of drops as the water moves through the block... first a loss at the entrance barb, then new drops for all the tight turns and the channel cross section, then another exit loss. It's the combined effect of all of these losses that determines the pressure drop of a block.

Unfortunately, what you propose is not as simple as it seems. You see, each of the components will have resistance curves similar to the block, and these need to be combined to create a pressure drop curve for the entire system. Only then can you use the pump's head curve to determine what your flow rate is going to be. I've illustrated this concept below (with some made up data, so don't be too critical of my values).
Attached Images
File Type: jpg sample pump and system curves.jpg (41.4 KB, 45 views)
__________________
Michael E. Robbins
M.A.Sc. Candidate, University of Toronto

12.1 GHz of AMD's finest (17.7 GHz total) crunching proudly for the AMDMB.com Killer Frogs
SETI BOINC: Dual Opteron 246s (Iwill DK8N) | XP2800+ (Shuttle SN41G2) | 3x XP2400+ (ASUS A7N266-vm)
SETI BOINC: 2x P4 2.8E (ASUS P4R800-vm) | Crunching 24/7
Skulemate is offline   Reply With Quote