Helium results!!! Window air con chill box build. - Page 5

Can't believe I'm still waiting on this poxy dvi cable....I've gone about as far as I can without it, got my mobo in the chamber and started her up using the cables I've run to make sure there were no problems.....there were no problems



Uploaded with ImageShack.us

Thanks Loki......just had my DVI cable delivered, on the home stretch now, should be testing by tomorrow

Okay just ran my first live test this morning.....and it didn't go too well.

I'm not sure if this is a temp monitoring problem or reflects an actual physical problem

Firstly the cpu.....I'm monitoring temps with "core temp" is this any good for subzero temp monitoring?

What is happening is on start up core temp was showing low 30's deg C, which seems high for idle considering the chamber temp was -24deg C. But then starts to climb rapidly...went to 128!!!!!!! at one point.

I would have thought the internal cpu thermal cut out would have shut it down by then if that was actual temp? Does the Q6600 have one?

I've altered my cpu fan settings in the bios to manual by duty cycle and set it to 100% and the fan is spinning at 2400rpm according to the bios hardware monitoring.....in the bios it showed cpu temp at 22 deg C but I didn't stay in there long enough to see if it started to climb again..perhaps I can try that later...22 deg C still seems a little high for idle considering the chamber temp.

I'm wondering it is as I speculated a heat pipe issue.....the evaporant in them is I believe distilled water......it may be boiling off and freezing at the far end of the heat pipe and so not be available to cool the cpu.

As regards the gpu I have no idea whats going on there. I'm using MSI Afterburner to monitor, the two gpu temps did briefly appear on start up..low about 3 deg C but then it didn't show a temp value and both fans climbed rapidly to 100% fan speed...at idle!!!

I've got 2 gtx 460's in sli palit 2gb platinum models. I don't think the heat sinks are vapor chamber based, which may suffer the same problems as heatpipes, I think they are soild. Don't think Nvidia introduced vapor chamber cooling on reference cards until the 500 series. My cards are only basic models so I don't think they have specialized 3rd party vapor chamber cooling.

I'm stuck and not sure what to do, I haven't done sub zero cooling before and not sure about the monitoring software...Help please!
Edited by technogiant - 7/1/12 at 4:32am

Just done another test.....started the system with the chamber above freezing and monitored the cpu temp in the bios then started the air con unit.

The cpu temps started dropping with the chamber temp, the cpu went down to 8 deg C at a chamber temp of -11 deg C but then started climbing rapidly......it's as I thought a heat pipe issue I believe.

That posses a dilemma for me. I can get a solid copper cpu heat sink that doesn't use heat pipes. But I'm still not sure what is going on with the gpus.

Even if my current gpu heat sinks are solid and this is only a monitoring issue with those, there is the question of what I would do about future gpu purchases. Most of them now use vapor chamber heat sinks and I don't think I would be able to get a current gen mid/high power gpu card that didn't have vapor chamber cooling...which I fear would suffer from the same problem
Edited by technogiant - 7/1/12 at 5:15am

Sorry Matt, I miss informed you there, mine are the sonic not sonic platinum......certainly not a heat pipe to be seen, and like I said don't think vapor chambers came out till the nvidia 500 series.

I don't want to have to turn the cooling power of the unit down unless absolutely necessary.

Besides I can't make the ac unit cycle on/off with my build as it stands, there's not enough heat capacity stored in there to allow the ac unit to turn on and off, once they turn off they don't turn back on for a set period of time and my system would over heat in that time.

I've just picked up an akasa AK-955 its a solid copper core with aluminium fins.....would have prefered a zalman starflower all copper but think it would not fit on my mobo.....still this one will do to see if it sorts it out.

I'll report back later...one more point do you know if MSI afterburner and CoreTemp work properly reporting temps below zero.....I've never done this sort of thing before.

To rule out one possible problem I've swapped out the cpu heat sink for an Akasa-955.
It has a solid copper core and alu fins. Also I can confirm that the gpu heat sinks are solid. I've set cpu and gpu fans permanently at 100%.

So at least I can now test the system with confidence I'm not going to roast my components.

Things just aren't monitoring properly though cpu temps are locked in core temp at 9,5,5,9 whether I'm running prime95 or not. I've restarted the application and the pc but it is the same. I did a hardware check in the bios and it reported cpu temps of 248 Deg C!!! so obviously erroneous.

I overclock my gpus and monitor them with MSI after burner but while it was originally showing temps when running furmark they then dropped from the graphs and haven't returned? must have been something I said or it just doesn't do sub zero.

So having done all this I'm unable to report temps

One thing I can report is that even while running multi gpu furmark and prime95 ie system totally maxed out the unit was holding the chamber temp at -22 Deg C

But there is a flip side to that also as the ac unit while doing this inexplicably shut down??....perhaps a thermal cut out on the compressor...it was quite hot...not sure if this is normal?


I have to sort out two things....firstly how to monitor my cpu and gpu temps at sub zero.

Secondly I need to know why my ac unit shut itself down.....I have briefly restarted it since and it appears to be working but obviously cant be having that happening all the time.
Edited by technogiant - 7/1/12 at 12:08pm

I think msi afterburner has been messing things up for some reason.

I switched to evga prescision and while that doesn't show sub zero temps, when I run furmark and temps pick up it registers them above zero....also coretemp appears to be responding....again at least under load.

So I can at least post some figures now....at least for full load.

When the system is fully loaded with furmark and prime the chamber temp is holding at -21 deg C....I'm well impressed with that

Gpu temp at full load is +29 deg C (compared to +76 deg C before mod)

Core average temp of cpu full load is +13 deg C (compared to +69 deg C before mod)

Can't post idle temps as they are not registering.

system specs are sli 2x gtx 460 2gb palit's oced from reference of 675 to 820mhz
and Q6600 oced to 3.52ghz at 1.35v

Haven't played with further ocing yet till everything is sorted properly

I'm still in a bit of a dilemma though because now I know temp reporting seems to be a bit hit and miss to say the least I'm not sure if the heat pipes were actually a physical problem or just a reporting issue cause by low temps. I will at some stage have to try it again if I can get the temp reporting sorted properly .

Oh and the ac unit shutting off.....duh had the ac temp probe in a draft