Testing Vapochill contact

[Durzel]

Probably a stupid question this..

Is there any way besides me removing the Vapochill head from the processor to work out whether I've got a good contact on the CPU? Reason I'm reluctant to remove it to check is because I think I've got a really good seal at the moment. The other issue is that I've tightened it so much to ensure an airtight seal that if I unscrewed the 4 thumbscrews I'll most likely pull the guides out of the board with them. :(

When I boot it in the BIOS it shows a temperature of about -50. In Windows however PC Probe is off the chart - showing 200+ degrees C. I have noticed whilst booting that the BIOS will report a temperature lower than the actual temperature on the front of the unit for a short time, until eventually the front display "catches up" to it. So I don't know whether the BIOS temperature is inaccurate, or the Vapo LCD temperature is, or both...

CoreTemp just reads 11C constantly regardless of what I do (that's the lowest Ive ever seen it go).

When I run ORTHOS the temperature on the front of the LCD doesn't change a huge amount, which is a little unexpected. I think it went down from about -58 to -55 with a stock vCore E6700 ES. I would've expected it to go down more than that if under load?

Without having any kind of meaningful temperature guide in Windows because PC Probe and CoreTemp (well, the board) are both unable to deal with negative temperatures properly I don't really know what else to check.

I did notice the temperature go up in PC Probe whilst running ORTHOS briefly (went from about 205 to 211 lol), if that means anything.

[sierra_bound]

I would trust the BIOS more than any monitoring program in Windows. As you have seen, most of them do not read sub-zero temps correctly anyway. I'm not sure the Vapochill display readings are very accurate either. -55 under full load seems a little too good to be true.

-50C in BIOS is pretty good. How much VCore?

FWIW, I've also used a Mach II. The display readings on the Mach are nowhere even close to reality.

[rob725]

At stock vcore, it may not be throwing off too much heat.

[Durzel]

I would trust the BIOS more than any monitoring program in Windows. As you have seen, most of them do not read sub-zero temps correctly anyway. I'm not sure the Vapochill display readings are very accurate either. -55 under full load seems a little too good to be true.

-50C in BIOS is pretty good. How much VCore?

FWIW, I've also used a Mach II. The display readings on the Mach are nowhere even close to reality.Stock vCore on an E6700 ES chip, so a little less than retail. I've just remembered that I've got throttling and EIST etc enabled as well because I just wiped the BIOS recently... I guess that's playing a part too. Doh!

[kayl]

When I boot it in the BIOS it shows a temperature of about -50. In Windows however PC Probe is off the chart - showing 200+ degrees C.
once die temps reach 0c in windows it reads 256c( -1c) so 200 would be -56c, but of course its not really that cold

When I run ORTHOS the temperature on the front of the LCD doesn't change a huge amount, which is a little unexpected. I think it went down from about -58 to -55 with a stock vCore E6700 ES. I would've expected it to go down more than that if under load?
with stock oc and volts that would happen, try 1.7v in bios and OC the chip. If you don’t have good contact with the cpu, the mobo wont let the system boot I have found with max vcore and high OC

Without having any kind of meaningful temperature guide in Windows because PC Probe and CoreTemp (well, the board) are both unable to deal with negative temperatures properly I don't really know what else to check.
I did notice the temperature go up in PC Probe whilst running ORTHOS briefly (went from about 205 to 211 lol), if that means anything
Taking into account that 256c is -1c that would make 205=-51c noload and -45c load. Id say you have good contact, but bump vcore to 1.7v and that will tell ya more.

[Gray Mole]

Keep in mind that due to the location of the temp sensor you can actually see a slow response from the probe.

The cpu however is right there, and if your bios monitoring is reasonable then it's going to show the cold much sooner than the probe as it's got a much larger contact area.

As far as contact goes, just start loading it up and after a while your display temps should get hotter. Try about 10 min at no-load, then read the temps after about 10 min on full load. If you get at least a few degrees difference you know you have some contact, and it should be a much larger difference the more vcore and higher clock you get your chip.

If your Bios is reporting the -50 temp, then you'll be able to get that in windows as well, just have to look for the right sensor. Try speedfan and look through the temps until you see one that's close to the bios temp, then watch it as you put the cpu under load. If it responds the way it should, that's probably the sensor you want to watch.

Cheers, hope that conroe clocks well for you :toast:

Gray

[Durzel]

Thanks a lot, that's really helped :)

Strange how it reads correctly in the BIOS but is completely wrong in PC Probe, as my P5W-DH didn't even read it properly in the BIOS either..

http://www.superficial.net/miscstuff/p5w64_mod/vapols_bios.jpg

[ayrtonsenna]

try speedfan to monitor temps.

[Durzel]

Tried Speedfan, the only temperature it picks up is one of my hard drives. :(

Currently running ORTHOS at 4.3Ghz at 1.4750v vCore. PC Probe is reporting a load temp of 249 degrees C, which Im guessing is about -17 using kayls useful info earlier.

Ive disabled all throttling, EIST, Speedstep, Hyperpath 3, etc.

Would a -17 load on a modded single stage Vapo LS (NL11F compressor, superheater) be about right for a good contact? LCD on front of Vapo is showing -51 at the moment, CoreTemp reports a solid 11C (never moves).

Really don't know how to determine if Ive got a good contact or not :(

[Gray Mole]

I had a freezer on the FX-55 and FX-57's I had before and running full load was around -10 to -15. The Dual core AMD stuff tended to be around -5 to -10 under load with modded Mach's as well.

So it does sound like your contact is good, and with the no load temp being around -69 or so, -51 sounds like it's pulling on the phase system fairly well too.

I'm pretty sure that's close to what I saw on it with the 150w tester (somewhere around there :p ) so looks like things are going well so far.

If you're stable and it's clocking well. and the display temps are stable and warmer than the no load temps, that's some of the best indications of good contact :)

Cheers

Gray

[Durzel]

That's good news. I know it sounds stupid but Im trying to think of everything except actually removing the head to see if Ive got a good contact. It was so much easier on my old IC7-MAX3 - I could tell straight away from looking in the BIOS if I had a good contact or not! With these boards not being able to read negative temperatures properly I dont know where I am :)

I just know if I remove it now Im gonna end up with a worse contact, and I know its gonna pull the guides out of the board because its so tightened.. basically it'll be a pain in the ass to refit it.

Vapo LCD is reading -48 now, PC Probe is up to 253C. Been ORTHOS stable for 40 mins so far... obviously cant say its totally stable yet but its positive :)

Thanks Gray for taking the time to reply :) Putting my mind at ease as usual!

EDIT: Awww ORTHOS bombed out after about 1 hour 30 mins.. guess I'm not using enough vCore, although PC Probe did report up to 254 degrees (-3?). Surely a unit capable of showing -69 on the front shouldn't be dragged down to -3, should it? ....

[Gray Mole]

Looks like your chip is starting to pull a little harder than the tester did.

At -48 display, it's likely you're around -38 or so as far as 'real' measurements go (Vapo tends to be about 10c off) so now it's just the difference between the evap and the core temp.

I don't know what the Conroe's are like but I know the AMD cpu's with IHS tended to be about 35c difference between the evap and the core temp under load, some 30c and some as much as 40c.

So that would put your core temp at around -3 if 35c difference is right for those chips.

Older Intel stuff was more like 20-25c difference but I don't know if something's different with the conroe chips that the difference is more like the AMD chips now perhaps.

If you're running it very close to it's stable limit you're going to see a much larger increase in heatload from the chip. It sounds like it's starting to 'ramp up' in heatload.

I'd recommend to find your highest stable clock, then reduce that by about 50-100mhz, and just a touch less Vcore. Then run it under a consistent high load for a couple of weeks to break the chip in a little bit. That's especially important when it's brand new as it can 'loosen up' after some running time.

Then come back to it and clock it to it's max and see if it's got some more in it.

Cheers


Gray

[Durzel]

Bit of an update - currently running ORTHOS at 4.3Ghz, 1.4875v vCore, 1.6v vMCH, 2.1v vDIMM, everything else AUTO. Manual RAM timings 4-4-4-12 1:1. PCI-E Freq 101, locked at 33.3Mhz.

http://www.superficial.net/miscstuff/conroe5.jpg

I'm guessing from the PC Probe readout of 255 degrees under load that this is -2 actual? Is this about right when the evaporator is reading -46.2? I still can't work out whether that means a good contact or not, when the unit unloaded shows -69 on the front.

TAT, like CoreTemp, is just fixed at 11 degrees - obviously it can't read lower than this.

Tried 4.5Ghz last night but even giving it 1.6v vCore, 1.8v vMCH, 1.4v vFSB and 1.2v ICH it wouldn't run for more than about a minute. Tried bumping the vDIMM to 2.25v as well with no change.

[johann]

The problem is that the vapochill itself is giving you an incorrect reading to start with.

The -69c you see when unloaded is more like -55c real temp. The -46.2 you see is more like -30c to -35c on the evaporator. -2c seems to be pretty normal then

[Durzel]

Thanks for that clarification johann. If only things were more accurate!

Well I stopped "Small FFTs" after 12 1/2 hours and started running "Large FFTs" to stress the memory (2GB G.Skill PC2-6400 HZ, D9 chips) but it failed after about an hour.. would that point to not enough vDimm memory? (2.1v I think currently).

vMCH is still at 1.6v, if this makes any difference.

Sorry to turn this into a general overclocking thread, just looking for advice really.

[Ssilencer]

Thanks for that clarification johann. If only things were more accurate!

Well I stopped "Small FFTs" after 12 1/2 hours and started running "Large FFTs" to stress the memory (2GB G.Skill PC2-6400 HZ, D9 chips) but it failed after about an hour.. would that point to not enough vDimm memory? (2.1v I think currently).

vMCH is still at 1.6v, if this makes any difference.

Sorry to turn this into a general overclocking thread, just looking for advice really.

I think you will never stress your computer that much in normal use, play some games and check stability.