Just wondering...
 

Hello everyone.. I am brand new.. I still have the tag on my shirt.

I was wondering what benifit a room temperature CPU could do. I understand that you can overclock greatly when bringing below room temperature.. but theoretically, what if the CPU never got above 80 degrees (Fer.) no matter what load was put on it. Would this chip be overclockable as well to a great extent? How much advantage is a TEC. Doesn't this just add the problem of condensation if it brings the temp below room temperature?
:shrug:

Welcome to ProCooling, especially to a fellow Texan!

Condensation is always a problem, and can occur at temps above the freezing point. It depends on the humidity level.

For a fixed temperature, there is a maximum speed that the CPU can reach. It's unfortunately not calculable. What you need to keep in mind is that there is a great difference between a CPU sitting at idle, and one that's hard at work. The latter emits a lot more heat.

If you are able to maintain a constant temperature, regardless of the CPU load, then that's great. It's not simple, but it's been done. Your max overclock will still be whatever the CPU can reach, at full load, while you're maintaining that temp. You have to keep in mind that the CPU has to be tested at "full load", to make sure that your computer is still stable.

Cooling at lower than ambient temperatures brings on a number of problems, the first of which is condensation. Go well below freezing, and you'll need to pre-freeze the CPU, just to get it to start!

There is a benefit in keeping a CPU cool: it will extend its life...

Thanks for your warm welcome.
65 degrees and sunny.. now thats what makes Texas a great place in the Autumn. ;)
Anyhoo.. now the beefy question. Could I get a 1.53 gig athlon to 2.0 gig at just room temperature. Since I really don't wish to mess with the entire consensation issue, I will shoot for a room tempurature cpu cooler. I am looking for a TEC that is variable, and will keep the cpu down, but not below room temp. Now, I know thats the tricky part.... but before I go there.. does anyone think this can be achieved without dipping below room temp? (75)

This way I won't have to freeze the cpu before I start.. and can turn on the cooling system when I turn on my comp.

Is this possible, or am i just a wishful thinker...?

It's most definitely possible.

Someone here made his own chiller, and TEC controller, and keeps fluid temps at about 10 deg C, if I remember correctly. No condensation.

Getting an 1800+ XP to 2.0 GHz depends on a lot more than the temperature of the chip. In general, the overclock is going to depend on 1) the quality of the CPU (most AMD steppings all clock about the same and so stepping rather than initial MHz is a better estimate of overclocability) 2) the voltage you apply to the chip (and the ability to keep the chip cool when you are pumping 2.0V or more through it), and 3) the ability of your components (PSU, mobo, RAM, etc) to handle the overclock. You can unlock the chip if FSB is limiting you.

If you can keep a chip at room temp no matter the heat produces (which isn't actually possible) then you should be able to raise the voltage to an extreme level and overclock pretty high. But with an old 1800+ then it is still unlikey you will hit 2GHz. Example: I have 2 1600+ XPs, an AGKGA stepping and an AGOIA. The AGKGA does 1610MHz watercooled, and the AGOIA does 1900MHz. Both overclocks are pretty typical for their steppings.

Hope that helps.

Which was a fantastic job, but pretty hardcore. He needed control devices to run the pelts and keep a constant temperature. Without that equipment (which is somewhat expensive - bit beyond my budget, anyhow) the temperture would quickly skyrocket or plummet when CPU load changed.

I can't remember who this was. I thought I had his site bookmarked . . .

Anyway, my current plans are an improvement on that idea - a bit less effective, but a lot cheaper. I'll post them when I get a chance to run some calcs and figure out whether my idea would even work.

Alchemy

Simple answers: ;)


1 - The overclock is possible - there are many factors, but just try it (always watching the temps). If it doesn't work, you can either lower room temperature, or go for higher end cooling.


2 - Room temperature is quite variable. A CPU running at room temperature when the room is 150deg wouldn't work too well;). A CPU running at room temperature in an air-conditioned room would work quite better:).

80deg F is a pretty good temp esp if at full load. If you could maintain that you have a good chance at a decent overclock.

I assume you will be running a higher end heatsink with some Arctic Silver 3, etc.. ?

I'm sure I've seen a TEC controller 'station' on sale somewhere(commercial), can't remember where though :shrug: . It controlled the Voltage of the TEC according to a therm' reading of the core/ambiant temp. its purpose was to keep temps from dipping below the dew point...

I found a TEC cooler prototype at www.bluecooling.com The site suggested that it would be for sale soon. They mentioned something about a controller for keeping it within a degree range. They also had a different kind of block.. blueblock or something.

And one more question.. in a latter post they said someone got down to 10 C with no condensation??? that sounds... impossible. I'm not saying he didn't, but I would think he would have had to mount it to the ceiling of his bathtub to drain off all the consensation. :D And, I will admit I don't know much...

All the replys show me that you guys are really cool. Thanks for all your help so far. I have learned alot.

From the reviews I have seen, an Alpha PAL8045 heatsink cools better than that Blue Mountain setup.

People aren't going to get 10C CPU temperatures under load without big pelts if you are talking AMDs.

WineWood: I currently own an XP 1800+ (1.533Ghz) and with my non pelter setup, I am able to OC to 1.748Ghz using stock voltage and keeping the CPU at a nice 30-32C. The only thing holding me back from going further is that my MB doesnt allow for AGP/PCI bus lock so my video and PCI cards are starting to complain AND I have yet to "unlock" my CPU.
Once I unlock my CPU, I fully expect to hit 2.0Ghz+ with a little voltage boost.

Keep in mind though, your CPU heat will greatly increase for every little bit of voltage you force into the CPU. My old 1.4Ghz T-Bird would max out at 1733Ghz but would run at 40-42C.

Hopefully you will have similar results but at least now you know for sure that it CAN be done without pelters and condensation. :)

Good luck.

I should clarify: it's about 10 deg C (can't remember exact figure) for the coolant temp, not the CPU temp. It's just enough to avoid condensation alltogether, except where it's very, very humid.

If you want an easy 2Ghz, wait until next month when you can buy an 1800+ (same as yours) TbredB for $80-90 that you can watercool to 2.1Ghz. If you aren't experienced in watercooling, peltier use is a dangerous road to go down. Get a good WC setup going for a few months and do a ton of research ... then see if you still want to go through the pelt hassle.

Alternatively, if you haven't bought anything for a WC/Pelt setup yet, look at a Prometia. It'll cost more than the WC/Pelt combined, but you'll get your 2Ghz.

Also, look into Volt Mods for your board. On an old 1800+ stepping (they sucked), you'll need some V going through it to get that high, and that will involve some serious cooling. You'd really be better off with even a TbredA than that.

(If you bought your proc within the last four months that may not apply).

True, even the AGOIA only hit 1900 regularly. I got mine(1800+AGOIA) hoping for 2gig, but I'm happy at 1758 (good luck with the 2gig Jtroutma :D) with stock V~core and cool temps. I'll wait for the 2400~2800's to come to earth (in price I meant) for my 2gig+ odessy...