Quote:
Originally posted by Alchemy
Regarding pump sizing, I've been trying to understand how the so-called "sweet spot" is possible for cooling systems. Convective heat transfer resistance *must* reach the minimum limit as the flow rate through the system is brought to infinity.
My best guess of why maximum heat tranfer does not coincide with minimum CPU temperatures is because at very high flow rates (and correspondingly high TDH) the pump duty becomes large enough to overtake the benefits of improved heat transfer.
If you pick the most important part of the system and consider the change in fluid temperature across the radiator as Q=UA(deltaT), UA is increasing very slightly as flow rate increases, but if you have to increase Q more than that to match it, you're going to end up with an even larger temperature gradient. Not good.
What do you think?
|
The "sweet spot" is, IMHO, largely imaginary unless one uses a ridiculously powerful pump. Within reason, more flow will yield temperatures that continue to approach ambient. As you surmise, once the power of the pump becomes a dominant source of the total energy, it'll cause the chip temperatures to begin climbing (unless you also have a ridiculously large radiator). With a very large pump, you
may be better off throttling flow as pump power will drop with flow in spite of the higher TDH. I can see it now, someone is going to begin selling the "Spaceballs" system with ludicrous pump, radiator, and water block. LOL.
Quote:
I assume from your posts you're a mechanical engineer?
|
Yes. And unlike most folks that get the degree, I have chosen not to use it as a stepping stone to other areas. The common joke at most companies that I know (at least the ones that employ engineers) is that "you must check your calculator at the door" when you transfer to any other department.
This alone does not make me any brighter than any other engineer. In fact some would argue I'm thick in the head for staying on the engineering side indefinitely. It has allowed me to continue in my field, learning new things and keeping my hands directly involved with engineering issues on a daily basis. I now manage the mechanical engineering department for a small OEM company and have ~15 people reporting to me. Fortunately, they fly pretty well on their own and I get to spend a decent part of my time thinking about designing and technical issues and not too much on "managing".
I gather you've already got an engineering degree of some fashion or another, too. I recall you said something about a "licensing exam" when you first began posting here.