View Full Version : Swiftech 462UT temps...help?

10-26-2002, 05:25 AM
Thanks to OPP and DUDE and a few others , I got my 226w pelt hooked up to my WB and running. Temps are lower BUT:

1600XP 0213 Y stepping

174 x 11.5 = 2004 MHZ @ 2.02v

Water temp: Slightly warmer than ambient( 76F ? )

Idle 42C , Load 50C

Thats a LOT LOT LOT higher than I thought it would be.

Heres the set up:

2 Gallon Resevior --->Via Aqua 370 GPH -->GFX WB/TEC---->CPU WB/TEC-----> Heater Core ( 10"X8"X3" )----> 2 Gallon Res.

I believe the CPU WB is Seated correctly on the core , that was the first thing I thought of. ANy suggestions?

10-26-2002, 06:30 PM
Ok ,

I pulled mobo and resat the WB/TEC. The Swiftech 462Ut is shipped assembled with AS3 already applied between pelt and cold plate. I did reconfigure:

2 Gallon Res-->Pump--->CPU--->GFx WB--->RAD---> 2 Gallon Res

IM still hot. CPU= 50C-53C loaded. Water temps , 80-85 F now ( warm ) after 2 hours loaded. Really no change at all.

I can geta voltage meter Mon. Im using a 400w Powmax , 12V 25A . ( It was $30 instead of $120). The POwmax has nothing else but the 226w pelt attached.

10-26-2002, 08:32 PM
its either the themal paste or the mixture of the water. Even the thickess matters with the thermal paste. Ive seen lower temps with 25% window washer fluid, 75% water. I would stay away from antifreeze because it is a much thicker type of fluid which slows the flow rate.

10-26-2002, 09:11 PM
window washer fluid is basically diluted methanol, which is probably the most efficient coolant additive. HOWEVER, methanol is poisonous and can cause blindness if too much is inhaled, or fatal if ingested.

antifreeze does increase the viscocity of your coolant, especially in sub-zero temperatures. It does help to keep the critters away, so i'd suggest dropping a couple ounces of glycol antifreeze into your loop.

I actually had similar results from my UHT block, which i was using with a chiller. My problem turned out to be the diode in the center slug area wasn't making contact with the bottom of the cpu, and the cold plate wasn't tightened enough to the block. unfortunately, the TEC wire snapped off, which got me to thinking, if i'm using a chiller, a peltier is really overkill... I got a little swept up in the fervor of extreme cooling ;)

did you ever leak-test the block? while leak testing (basically running the loop without mounting on the cpu), you can test the peltier and see what kinda temps your cold plate gets to. If your cold plate doesn't form condensation on it, something is wrong.

Your water temps are tolerable considering the load, but you can try reducing the size of your reservoir (bigger is not always better) or adding ice cubes to it.

lastly, what kind of radiator are you using? if you have a good radiator/fans, your water temp will get considerably closer to room temp (which greatly affects your radiator's efficiency), which will probably lower your cpu temps.

well, really lastly... make sure you really are making a nice, clean footprint on the bottom of the cold plate. a matter of uber-importance!

ok really really lastly, water-cooling temps will always be higher than they really are, if you're using a motherboard that has a diode in the center slug area, because there isn't the same airflow under the socket as there is with a HSF. i direct you to swiftech (http://www.swiftnets.com/troubleshooting.htm) for more information....

PS, you're using dielectric grease in the socket pin holes, right?

10-27-2002, 03:45 AM
Im reading ffrom BIOS PC Health , and yes that is the thermistor in the center of cpu slug. Dielectric grease of course , and plenty of neoprene to.

I did reseat the block , and the footprint from my original install looked perfect.

MAybe its the PSU? I cant check its output till Mon , fluke is at the office.

10-27-2002, 08:00 AM
you should test your block without it mounted on the cpu, And see how cold the cold plate gets...

10-29-2002, 08:22 AM
hell i want one of these babys too but the mcw478-uht hope to post temps n things soon.

11-18-2002, 08:46 PM
Mine's been running 24/7 for 2 months, I have a promethia on the way, If anyone interested in my 462-ut.

11-24-2002, 05:21 AM

I ve checked the coldplate while its not mounted. It gets very cold. I also have removed the mobo gasket for installation. This puppy is sitting FLAT on the die.

Is Socket die temp readings of 45C Idle - 52 C loaded sound right at 2.05 volts @ 1800 mhz.

Btw , this Cpu is a AGIOA '9' 0213. My "Y" is taking a breather lol.

11-24-2002, 08:39 AM
I see you're using a heatercore? That should be able to handle the load. I get similar temperatures at 1900MHz, 2.15V, but I only have a 156W TEC. At stock speeds, my temps stay less than 10ºC. The only things I can think off is either bad contact between cold-plate and CPU, bad contact between cold-plate and TEC, bad contact between TEC and waterblock, or last of all, that your radiator can't handle the heat load.

11-24-2002, 09:02 AM
I have 2 Rads actually with shroads and good airflow.

Im thinking it is the TEC PSU. I DID have a 13.8v/25A RAdio Shack ( which I returned), but after 2 -3 hours the water got real hot and temps where about 50C loaded. When I would first power up , the temps would be like 15C.

Right now Im using a cheapo PowMaxx 400W Atx psu for the 226 pelt. Its rated 12v/25A. It seems to work ok...as the coldplate does get really cold really fast - no ice though. If I shut off the TEC psu , I can watch the temps quickly climb to the 60's @ idle.

I have tried tightening the coldplate , which was really tight to begin with , and reseating the block several times( each time I find a perfect ASIII imprint. I guess Im gonna have breakdown and order the Meanwell for $118.

Or....does anyone suggest something else??

11-24-2002, 09:07 AM
Yeah, get a dedicated PSU. An ATX PSU is not meant to handle that kind of load. Though it may be possible to still use it. Do you have any load on the 5V line of your PSU right now? if not, you may want to consider loading it down so you get about 1A through it, it may help stabilize the 12V line, if a crappy 12V is the problem.

11-24-2002, 09:17 AM
KE ,

I have a very simliar setup as you , Whats your TEC PSU?

11-24-2002, 09:23 AM

I'm using a 13.8V, 23A SEC1223 bench PSU. At the moment it runs my TEC and my rotrons.

11-24-2002, 05:55 PM
It's hard to belive 2 PS are giving you the same. As for loading the 5 volt it's a must on a ATX. im using a variable 5/17 volt 36 amp supply with gauges.

If it is the PS's. all i can guess is your radio shack model isnt cutting it, and then your ATX isnt loaded without a 5 volt on it and your only getting 10 volts or less.

Cant be your heatercore, cause your getting those temps at startup.

I've had mine on my xp1600 2.15 volts 2000+ months ago. and was holding steady @ 12C with some cold water.

What's the temp of the cold plate unloaded if you have something to measure it. Maybe you got a bad TEC, Let it sit and run, you should get a lot of ice on it, in a while .not just cold as you stated.

11-26-2002, 09:04 AM

After consulting with someone smarter than I , lol.....

What Ive done with the PowMaxx 400 ATX is rigged a 5 ohm 10Watt resistor in the 5volt line to load it.

I now have ice


I dont know how long this PSU is gonna last ( it didnt get hot or smelly after 6 hours) but for $35 its doing pretty good for me atm.

Im having a hell of a time scrapping up the $ for a GOOD variable PSU with this kind of ampage. A fooking $200 speeding ticket the other day doesnt help my cause either.

11-27-2002, 06:11 AM
If the Powermax is holding and not getting hot you'll be ok. Watch the resistor though. It's got to me hot as hell.

I've got an antec TP 550 in my rig, and i've ran the comp, and the TEC for a week just to test it Volts dropped to 11.72, but it ran fine, and didnt get hot.

I like the variable for being able to back it down to 9 volts or so when im not Home. If you noticed the pelt wires do get hot. And hate to burn the joint down.

11-27-2002, 06:32 AM
kchip, What are your temps now?

11-27-2002, 07:00 AM

Thats the thing that bothers the shiat outta me.

Right now Im using the TEC on my EPOX 8K3a+ with the sensor in the socket slug. I sat the block on the cpu , bolted it down , then removed it to check the imprint - PERFECT.

Cleaned , reseated , powered up. 2100mhz @ 1.80v = 48 C idle 52 Load in BIOS and desktop.

Annoyed , I removed , let Ice form again , cleaned again , resat again , removed again , imprint PERFECT again ,reinstalled , Powered up 48C idle , 52 C Load.

Note: While waterproofing the cpu slug , i git some rubber silicone on the sensor. I did remove with a razor knife the excess after it dried.

From everything Ive gathered , thse motherboards with the slug sensor are just " guessing" what the temps should be? Am I right in this assumption?

11-27-2002, 07:14 AM
That is right, several people have stated same thing. My prom is running a 8K5A2+, 2200+@2333MHz, 1.95 Vcore. Cpu temp according to MBM and Bios Diode temp is 25c which I know is not right I reseated that Microfreezer so many times I can do it Blindfolded.My Swiftech Q Pwr with 226W pelt, 1600@1916MHz, 1.95 Vcore. Diode temp 30c. Epox temps are BS.:)

11-27-2002, 08:47 AM
SO its not just me thanx god

11-27-2002, 08:58 AM
As long as you can get a good overclock (which you are) then I wouldn't worry overmuch about your temps. We can be fairly sure, I think, that your TEC is not being overloaded at the speeds/temps you're running at, so don't worry about it too much. Good to know you got it working :toast:.