CPU Only Cooling / Avg Water Temps (?s)

This is my first post here but I have been around the block a few times. I have been fighting heat issues with my chip and have finally come to the conclusion that I need to get it stabilized at a lower temp before I can pursue any more clocking. The Chilltec just isn't made to handle the temps of an Extreme Ed. Pent D at moderate clocking. Once I get to around 45*C stability becomes an issue when running Prime 95 although I have ran OCCT for over an hour with no problems and have ran Memtest for about 12 hours with no issues. Prime just seems to hit the cpu quite a bit harder.
Anyways , I am planning on just water cooling the chip and have an idea for a custom build. I want to be able to achieve a system that will sustain a cpu temp of no more than 45*C and be able to get my voltage up to 1.4(@1.32v now) if needed. I know that this cpu has a lot left in it and the ram is just getting warmed up. It's just a cooling factor at this point and the fact that it's a Pentium D doesn't help either. I'm just stubborn about moving up to an Extreme C2D and I don't want to drop the cash for one , yet.
Is there an average or method to figure out what the water temp will be at in the radiator/reservoir ? I'm planning on doing a light chilling and need to know about the coolant temps so that I can avoid condensation.
Sorry about being so long winded. Any suggestions or ideas highly appreciated!

You could aproximate it, but not to any real accuracy unless you knew exactly how much heat your CPU is dumping into the waterr the voluem of the water, Thermal conductivity, of all parts of your loops, ambient air temperature, volume of air pushed through the radiator, etc etc. The easiest thing to do would be to make a temp sensor, and put it in the loop.

But just to add, you really shouldn't need a chiller. Unless you are doing that just for the sake of doing so.

Thanks for replying. I think I will use your advice about the temp sensor to see what I'm going to be running. I just wasn't sure what to expect from water cooling and high overclocks. I have plenty of room so I will probably go with a dual fan radiator and a top end cooling block, pump and reservoir. Should I mount the sensor to the water block or in the reservoir for most efficient readings?

liquidfluidity said:

Thanks for replying. I think I will use your advice about the temp sensor to see what I'm going to be running. I just wasn't sure what to expect from water cooling and high overclocks. I have plenty of room so I will probably go with a dual fan radiator and a top end cooling block, pump and reservoir. Should I mount the sensor to the water block or in the reservoir for most efficient readings?

Click to expand...

I would suggest that you consider the res for your thermal probe. The only reason I suggest the res is that in the res, the thermal probe will not reduce coolant flow is it would most anywhere else in the loop.

Yes , I guess that would also give a more constant temp also. Thanks!

Where are you getting those temps from? I ask, because 45*c is pretty low for the Presc"hot" processors. I know, because I had a 3.0E (aka P4/530) doing 4.2Ghz at 1.6v for a pretty considerable amount of time. My water loop consists of a Via Aqua 1300, a D-TEK White Water block, and a generic automobile 260x120mmx45mm radiator. Temps in this loop with two 120mm fans pulling at 7v usually hang around 55*c, except for in the winter when it cools down a bit.

These Prescott's are good for something stupid like 68*c reliably, and I believe are rated for even more. So if your temp is 45*c, then I wonder if your instability is truly thermal or something else instead. I guess what I'm trying to say is, if temps are as "good" as you say they are, then water may not help much more...

These temps are taken with PC Wizard , Speedfan and Smart Guardian. I have also tried Everest Ultimate but It was showing 31C the whole time so I quit using it. I am assuming that these progs are getting these from the bios. Oh yeah , this is a Presler core Pent. D. I've seen them take up to 1.6v but there is no way that the Chilltec can keep up with that. I made the assumption of the heat holding me back because when I reached the 45*c mark is when I started having issues. I even reseated my cooler but that didn't help either so this has lead me to trying watercooling.and maybe adding a small chiller to it If I can figure how to keep away from condensation.
Could it maybe be that my NB needs some more volts? I have noticed that it starts getting warm also. It is using the stock heatsink and running the stock volts. I'm just not sure where to go other than the cooling change.

Even the Presler cores don't flinch at 45*c. You've got FAR more temperature to go until it's truly an issue. If you're unstable at that speed, it's likely due to some other reason -- unless you're talking about some severe-duty overclocking.

Well then I must be doing something wrong. I'm only at 4.25 GHz w/ an FSB of just over 1100. If I drop the multi and raise the fsb, or if I rais the multi then the fsb , I seem to start having issues once temps are around 45*C. I am only running at 1.34v. Any suggestions? I'm sorry this is turning into a "how to overclock a Presler" thread but I thought that the heat was the issue.

liquidfluidity said:

Well then I must be doing something wrong. I'm only at 4.25 GHz w/ an FSB of just over 1100. If I drop the multi and raise the fsb, or if I rais the multi then the fsb , I seem to start having issues once temps are around 45*C. I am only running at 1.34v. Any suggestions? I'm sorry this is turning into a "how to overclock a Presler" thread but I thought that the heat was the issue.

Click to expand...

You simply aren't giving it enough voltage. It is not related to your temps. I had a Pentium D 805 that was clocked at 4.3ghz and temps were 66C and it was stable, so I do not think that temp is your issue....

thanks for the advice. I will give it a shot and see what happens. I can't believe you ran yours that hot! The 955Xe is rated for 68*C and I would never run it that hot , though.

That's a decent overclock to be sure, but you're running very thin on voltage at that level. These chips are power hungry; you have to feed them power to keep them running faster. If temps become out of control while feeding them all the extra voltage, then of course you'll need extra cooling.

Obviously, that's where water comes in. But you've still got a bit more room on air before you need to worry about it...