Quote:
The optimum operating point for the pelt is dependent on the performance of the hot side cooling system. It is not a fixed voltage like 20V. Pelt cooling characteristics are very nonlinear, and they maintain substantial cooling capability to voltages (and/or currents) much lower than Vmax. See Figure 4 of this document for an illustration of this. The power consumption of (and heat dissipated by) the pelt goes up roughly as the square of the voltage. Above Vmax/2, the pelt's cooling ability is increasing less and less for each additional volt. With an underperforming watercooling system and a 16 Amp powersupply, two Drift 0.8's connected in series for 12V apiece, is going to be much more optimal than two in parallel at 20V. |
One single 172w drift will not perform good at all at 12v. The optimal range for ONE 172w drift pelt is between 18v and 20v (I used to run one).
The clamping pressure is what helps determine the amps it can draw. For the best performance, an ampmeter and a variable output PSU should be used. Hook the rig up and put the comp under load. Start the pelt at 17v and watch the ampmeter and your load temps. SLowly start upping the voltage until you see that there aren't anymore amps going to the pelt and/or you aren't able to cool the processor down anymore. I see where you are coming from. You are just running two pelts at under their performance rating as to keep the hot side cooler. That's kinda cheating, but gets the job done. I still say shoot for the moon on it though. ;) With a better performing cooling system for the pelts, you should be able to run them both separately to insure each pelt gets its own fair share of amps. |
Quote:
The clamping pressure recommendations are to ensure good thermal contact between the TEC and the hot and cold plates. It has no significant correlation with the current draw. Quote:
I certainly would agree that checking the CPU temperature over a wide range of pelt voltages would be a good idea. (Need to give the system time to stabilize at each voltage setting though.) Not many people are going to have the supply to be able to do that testing though. |
About clamping pressure. I read that article AGES ago. Like I said, I have a DangerDen Maze 2.2, have a look and SEE what it is:
Specs: http://www.dangerden.com/mall/maze2.asp Just the pic: http://www.dangerden.com/images/MAZE...ront_large.gif As you can see, the coldplate uses the EXACT SAME mounting system. My pelts already HAVE that much pressure. I have mounted the block without the springs to get better pressure from the coldplate to the chip. Not much difference. Maybe 1-2C under load better than with the springs. Don't have time right now to mention the rad, but will check the review out tomorrow. Thanks for all the input. Really like the shroud idea. I should have known to do that :( But thanks! |
For what it's worth...
I checked out the mechanical properties of those nylon bolts, and they should be able to provide the clamping pressure necessary. http://www.plasticnutsandbolts.com/ So you're down to getting that water temp lower. |
You guys don't seem familiar with the DangerDen Maze 2.2. I should have been a little more clear. The nylon bolts hold the coldplate to the CPU. There are steel bolts holding the coldplate and block together with the pelts between them.
Yeah here is a link to my pump, check it out: http://www.pumpworld.net/prod8.htm#3B-500 500gph @ 1 foot, 300gph @5 feet. Quieter than my MagDrive. |
Quote:
|
Quote:
Nice pump. It has an odd curve, where it's most efficient at the limits of its range. This tells me that it's designed for a fountain. Yup, it is. You'll do well with that pump. |
The curve looks odd because they're not using a linear scale on their x-axis... 0 - 300 - 360 - 500 at even spacing? If you plot the individual points that can be read from that graph in Excel it looks much more like a "normal" pump curve... a little flat actually, but not too bad.
|
Quote:
You're still totally wrong about the current leveling off at a certain voltage, and that voltage being the optimum operating point for the pelt. (Are you aware, that a lot of variable output supplies have current limiting circuitry, and they will actually stop increasing the voltage when the current limit is hit, regardless of how much higher you crank the voltage output knob?) How about demonstrating some understanding of how a pelt operates and proving that you're right, rather than giving such a lame response? If my understanding of pelts is lacking and you can explain why, I'd be glad to know it, and humbly apologize for being such an arrogant asshole. Until then, I'm going to consider you a fool giving bad advice. When messing around with pelts, foolishness has significant risks. |
Quote:
|
And I think Mr. Baz has misunderstood DrRom's findings, and believes that because the temps don't change as you add voltage above 18-20V, the amperage does. However, I don't know where clamping pressure affecting current draw came from.
|
Quote:
and believes that because the temps don't change as you add voltage above 18-20V, neitehr does the amperage. |
To quote DrRom, "I'm running the 172w with 18v from a switchmode PSU - higher voltages offered no cooling improvements for me at all."
He didn't say that he found 18V to be the optimal voltage to run at. He only said that running at higher voltages was useless or worse. The simulations I've done, show 16V to 17V to be the optimal place to operate with "good" watercooling. "Extremely good" watercooling might get the optimal operating voltage up to 20V. |
Got my new pump installed. Sure is nice! Temps dropped a bit, the block is cooler, but CPU temps are still too high. ROughly the same as they were.
Would a shroud make a HUGE difference? Am I going to need a new rad/heatercore? What would you suggest as highest performance available? |
Quote:
A shroud won'y make a significant difference. Maybe you could tell us which model fan you're using? BTW, if you've read Bill Adam's radiator roundup article, you should have noticed the airflow graphs: if you can get a fan to create a pressure drop equivalent to 1/4 inch water, you should have plenty of cooling power. I suspect you're using an L1A panaflo fan, which really can't get you the airflow you need. |
Well, there is an article on www.overclockers.com under watercooling, where the author makes a nice shroud and gets a 40% INCREASE in cooling performance from his heatercore.
I have a YS Tech 120MM FD1212387B-2R roughly 135CFM I think? |
I stand corrected!
Link to shrouding article If this fella can drop his CPU temp from 51.8 to 44, You might as well give it a shot. If you want a quick idea of how beneficial it can be, try ducting with tape first. As for the fan, it will deadhead at .25 inch pressure drop. In other words, the actual flow you are getting is probably closer to 75 cfm (max 125) with a pressure drop closer to 1/16 inch. Specs here You might consider a more powefull fan, or multiple fans. |
I don't have room for multiple fans. What would you suggest for a more powerful, yet not DELTA sounding 120MM fan?
|
You can go through this chart and pick a unit that has a low noise to flow rate ration, while maintaining a good cfm (with either a voltmod, or a rheobus, or whatever controlling device you prefer).
[edit] It looks like the Sunon KD1212PMB1-6A is a winner in the 120mm size. |
Which rad/heatercore can cool 2x 172w pelts? BIX can't......
|
Quote:
|
well I have a 120mm hole in the top of my tower which I cut for my BIX. So I can only use a single 120MM fan, but the rad can be bigger than that.
My tower is Antec SX1200 |
I'd try
Code:
Fedco# Height Width Thickness Inlet Outlet Make Model Years They're the smallest cores that are 2 5/8 inch thick. |
What makes you think it would be ANY better than my 100% copper (YES the WHOLE RAD is COPPER, NO OTHER TYPE OF METAL) 2 pass high flow BIX?
I just checked out the Swiftech kits and noticed they include TWO fans with all kits. One blowing air through the rad and one sucking out the otehr side. Would that help at all?? |
All times are GMT -5. The time now is 05:00 PM. |
Powered by vBulletin® Version 3.7.4
Copyright ©2000 - 2025, Jelsoft Enterprises Ltd.
(C) 2005 ProCooling.com If we in some way offend you, insult you or your people, screw your mom, beat up your dad, or poop on your porch... we're sorry... we were probably really drunk... Oh and dont steal our content bitches! Don't give us a reason to pee in your open car window this summer...