ira-k

Well I put a tape temp probe on the side of my GTX 480 to see how much it warms up when my chip is under load.The top of my rad, its barb side down is the warmest spot on the rad. As most of you know when your rad gets overloaded it will warm up a lot.I was bored today.

I have a 437W TEC in this loop and my chip was running at 3138mhz.

TEC at 22 vdc =367W

CPU=120W or therabouts at 1.568 V-Core.

CPU temp measured by SpeedFan core-temp.

Total heat-load = 487W



Prime 95 large FFT's were run for 30 minutes and recorded at 10 minute
intervals.



click me


................Idle .........Minutes .......10...... 20....... 30


CPU- -5C ....................................15 .......15....... 15

Ambient- 26C ..............................26 ........25.5..... 25

Water In- 31.8C ........................32.4...... 31.9.... 31.5

Water Out- 30.5C ......................31.2 ........30...... 30.3

Side of Rad- 28C .......................28.5........ 28....... 27.5

Average only includes load temp's.


Average CPU temp- 15C

Average room ambient temp- 25.5C

Average water in temp- 31.99C

Average water out temp- 29.56C

Average side of rad temp- 28C

Average rad temp over room ambient- 2.5C

Average temp of water out over room ambient- 4.06C

Remember this is with a heat load of 487W, or 2 highly overclocked quad-cores
loading with Prime 95 large FFT's for 30 minutes.

Water in and out temp's are measured with in-line temp probes.

I have 4-108 cfm Sunon on a 1.25" shroud on pull through the rad. I have 2 more 115 cfm Panaflo in push in the front on no shrouds.I have the materials coming to do a full 1.5" on push and pull for a optimum configuration. I'll do more shroud tests then to see if th extra depth helps cooling any.

This is hardly scientific I and used cheap temp measuring equipment, that being said it is as good most on the web base their claims on.I don't this is just a very basic test using uncalibrated equipment and is meant to be nothing more then a broad guideline of how this rad behaved today in this configuration and in this moon phase.

I can say this with a certainty though even using this testing equipment, 487W of heat dump did not even begin to load up this rad.

The lousy CPU temp's are I believe caused by my broken mobo back-plate, I hate to take it off anymore my plate is comformaled on the mobo, so I will just wait till I upgrade to take it off again.I broke off both studs and had to jury rig it back on.

I was going to do a pressure drop test on this rad last week but one of my pumps was dieing on me and I didn't know it and that invalidated any of my reading.One reading that is correct can be seen in the following pic's which are self-explanitory. I only lost .25-.50 psi going from a BIX 240 rad to a 480 GTX rad.The loops are different but just from this I can say its not very restrictive at all.The psi gauge is placed as close to the outlet of the pump to be as accurate as possible.The guage itself is accurate to +-2% midscale.

BIX 240

click me

GTX 480

click me

Bunch of junk

click me

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section** TEC Guide:Basic Principles of TEC's

**E7200 @ 4545 *9 **5000BE @ 3510*10**LE 1600 @ 3500*10**150 Opty @ 3242*12**

loop0001

hmm..cool beans lol, oh yeah btw...you have quite the mess there!
good to know that the 480 rad with the fans on it is in safe hands :P

__________________

Currently Playing: Crysis, CSS, COD4, GRID
Behold!!! I am a Parodine Shifter!!!

CrazyJoe

Great job there ira ! Reps for all the hard work !

Perhaps the components and hose lengths could be fed into Martin's Flow Rate Estimator to see what kind of numbers that comes up with as compared to your actuals ?

__________________
Old enough to know better ... still young enough not to let that stop me !

dr4gon

Nice, can't wait to see the final results with charts .

So the HW lab claims of 2200W+ still might be accurate but definitely hard to test.

__________________
Project: Orange Box
Join the New OCN Flickr Group - My Flickr
>>> For Sale: <<<
Team Fortress 2 - Add OCN @ 8.9.16.94:27015 to your favorites.

Martinm210

Awesome...400+ watts and less than a 5C delta is great.

So it's probably good for something like 900-1000 watts at a 10C delta

__________________
Flow Rate Estimator | Radiator Sizing Estimators

ira-k

Well Joe as far as I can tell it wont be anymore restrictive then the 360 BIX, its barely more restrictive then the 240 BIX but I have a D/O block and res and V8 on the second reading...So to err on the side of caution say the 360 BIX, which isn't very restrictive anyway...One of my DDC's burned out so that with a single orange impeller 18W DDC and the one with the 240 BIX is to, that reading was taken a long time ago when I was running a single also...


Yeah Martin I would say it could easily handle a 1000W load...It really feels cool to the touch with this much on it, just a couple of C warmer the room ambient...If I can figure out a way to load it up more I will...On my next loop I would like to TEC the GPU also but that would only be a few hundred more watts at the most....I figured I'd do this since Hardware labs wont give us any info other then the wattage they claim it will dissipate....I could go quad I guess but I would rather just do a dual-core so the TEC can give me better temp's..Thanks guys it was fun, when I have my full 1.5" shrouds front and back I'll try it again and see if it gives me any better temp's then the shorter ones did...

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section** TEC Guide:Basic Principles of TEC's

**E7200 @ 4545 *9 **5000BE @ 3510*10**LE 1600 @ 3500*10**150 Opty @ 3242*12**

ChielScape

add a few more volts to the pelt ;p

__________________
Bigger = better, except for transistor size.
480GTX Club

ira-k

I could only get another 70W or so its not enough to bother with...

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section** TEC Guide:Basic Principles of TEC's

**E7200 @ 4545 *9 **5000BE @ 3510*10**LE 1600 @ 3500*10**150 Opty @ 3242*12**

ChielScape

add your other stinger to the loop with 320w pelt with some heatsink on the cold side? should give you the rated 320w+what comes from power consumed by the pelt.

__________________
Bigger = better, except for transistor size.
480GTX Club

ira-k

Yeah I hadn't thought about that I might just do that...It wont be on a die but running it at 12v will give me around another 226W any way...Then add another 70W with this 437W TEC would be around 783W....If it can keep that under a 10C delta its a killer......If its a true 1000W rad it should be around 7-8C delta....

__________________
**Water Cooling Essential Threads** Overclock.net's Downloads Section** TEC Guide:Basic Principles of TEC's

**E7200 @ 4545 *9 **5000BE @ 3510*10**LE 1600 @ 3500*10**150 Opty @ 3242*12**