bobbyt2012

Hi All,

I am going to be doing a science fair project on the effects of temperature on electron mobility by cooling my CPU with several different methods of cooling including Dry Ice and LN2. I have a few questions, and I'm sure more will come up. The most pertinent question right now is how low will the built in thermometers go that display the temperature in CoreTemp? I have an AMD Phenom II x4 940 BE. I'm in high school, and I don't really have a whole lot of spending money. I'm buying an evaporator from Koolance (CPU-LN2), unless you guys know of a better one for less money.

I'm working on getting a large array of relatively quick CPU tests for my experiments. The ones that I have come up with so far are: SuperPI to 4mil, Cinebench for Single Core and Multi, 3dMark06 CPU Tests Only, 3dMark Vantage CPU Tests only, and possibly 1 minute in Prime95 for stability only.

Ceiron

I'm highly doubtful the internal thermometers are built to display anything below 0celcius, or 32F (I think?)
You might need to grab an external electric thermometer if you're planning to go that low below 0.

OverShocked

Thats nuts, im doing the EXACT same science fair project....

I wouldnt buy a pot from koolance, you will get bad temp swings, Buy one from bartx or duniek.

The temp sensor that the cpu has will go down to -10 or -15 and below that it will start reading crazy things like 247c.

I reccomend buying one of these.
I bought one of those and it seems pretty nice, it just takes a couple weeks to arrive.

__________________
E8500-4.93Ghz on air
E8500-5.5GHZ on DICE

i7 920- 4.52 Ghz

9800 gx2-820Mhz

bobbyt2012

Thanks for the input. I looked into duniek and bartx, but the Koolance one just seems to be of higher build quality, no offense to bartx or duniek. I would be willing to consider getting one of theirs, but I just don't know enough about LN2 pots.


I'm planning on testing air, water and ice, dry ice (and water), and LN2. That thermometer that you posted looks like a deal, maybe I'll just pick one of those up.

Also, what do you think the chances of destroying my hardware would be?

Thanks,
Bob

billbartuska

Electron mobility is not dependent on temperature. Just cooling something doesn't change how fast the electrons are flowing through it.

Are you assuming that just cooling a CPU will make it run faster?

__________________
EDGE ...if you're not on it, you're taking up too much space.

epitek

I wish I had friends like you guys in high school, my friends used to make tripple chamber bubbler bongs out of jars, glass cutters, pvc, and superglue :\

__________________
"Let us consider that we are all insane. It will explain us to each other; it will unriddle many riddles..." - M.T

GO CELTICS!

OverShocked

Yeah, no prob.

Bartx and duniek have MUCH higher build qualities. The one that koolance makes doesnt have enough base mass (in order to save money). You will get high temp swings under ln2 if you use the koolance one.

Gomeler did a review on the koolance pot... wish i could find it.

EDIT: found it

"So what.. No results? How best to put this.. the CPU-LN2 is not an adequate liquid nitrogen pot. I started my testing on AMD's Phenom II X4 955 and even with the pot full of nitrogen I was still seeing massive(20C+) swings in base temperature through the relatively short 3D06. For example I can run 3DMark 2006 at 5900MHz on my F1EE but I had to drop down to 5650MHz with the CPU-LN2 due to the processor getting too warm through the CPU test and crashing. wPrime 1024M crashed at 5800MHz typically between 65 and 80 seconds."

Source: http://www.legitreviews.com/article/987/6/

Not to mention the crappy alcetal extention.

__________________
E8500-4.93Ghz on air
E8500-5.5GHZ on DICE

i7 920- 4.52 Ghz

9800 gx2-820Mhz

Patch

The only advantage the Koolance pot has is a marketing budget.

While it supposedly works fine for an aluminum pot, it would be inferior to a custom pot with a copper base. Also, the acetal top on the Koolance pot can crack under LN2 if you're not real careful insulating.

That is a bonehead design flaw to save money on manufacturing costs.

Some motherboards can detect temps down to the -50 to -60 degree range, but they are incredibly inaccurate at that point. And eventually you start getting readings >100 C as it flips out. The best indication of temps at below freezing is to use a temp probe inserted into a small hole drilled into the base of the pot. While the actual core temps will be significantly higher than that (20-50 degrees higher), it still gives you some numbers to adjust and make overclocking decisions by.

__________________

____________________________
.:|Official|Volt|Mod|Squad|Member|:.

Don't tell me I "can't" when you really mean I "shouldn't".

5.27 Ghz i7 920 C0 on LN2......5.31 Ghz i7 975 on DICE......6.36 Ghz E8600 on LN2
7.01 Ghz Pentium 4 631 on LN2......7.96 Ghz Celeron 352 on LN2......7.02 Ghz Celeron 360 on LN2

bobbyt2012

Thanks a lot for all the replies. I really appreciate it.

I'm thinking that I will be doing 2 different tests. The first one is to see if exclusively cooling the processor will cause an increase in performance. I really don't know the answer to that, so I think it is a valid experiment. The second experiment will have 2 manipulated variables. The first variable being temperature and the second being clock speed. Hopefully I can keep the voltage at a constant (is that a bad idea? what should I be doing voltage wise?) With the second test, I will be determining how much of a performance increase that temperature can cause.

Luckily, I am still in the early stages of designing an experiment for the 2nd test, so I have time to adapt it. The first test will be at stock settings. I want to decide on a good clock intervals so that I can test on air, (water), water and ice, dry ice, and LN2. For example 3.0Ghz, 3.5Ghz, 4Ghz, 4.5Ghz, 5Ghz, and 5.5Ghz. Once it won't boot for a particular cooling method, or isn't stable enough, I won't try anymore. I know that I probably won't have a lot of time with LN2 because I'm getting a teacher to bring it into school, so I need to be efficient with my tests.

Can anyone recommend a specific pot from Duniek or Bartx? As I said before, I have an AMD Phenom II x4 940 (AM2+).

Once again, I appreciate the input,

Bob

Weedvender

You know what? Ask first before the school allows you to play with LN2. Ask for everything honestly.

Off topic:

I made a 2.5 kJ Rail Gun for my High school project. I also went to a third world high school in a can't/impractical to sue country.