QX6700 O/C expectations?

[Idontcare]

OK, here's what I basically got:

1. QX6700 (B3, I think)
   Striker mobo
   Vapo LS

Here's where I am at so far for O/C (orthos testing, less than 1hr generally)

1. 2.67/2.93/3.2GHz all stable at "stock" voltage (10x, 11x, and 12x multi)...Asus probe reports 1.23V idle in XP and 1.20V at full Orthos load (4 instances of orthos)
   3.47/3.73GHz require Vcore bump to 1.45V in BIOS to be Orthos stable (13x and 14x multi)...Asus probe reports 1.34V idle and 1.29V orthos loaded
   4GHz requires Vcore bump to 1.55 in BIOS to boot into windows and still not orthos stable at this voltage (15x multi)...Asus probe reports 1.45V idle and 1.4V orthos loaded

So here is my current quandary, would love to secure 4GHz orthos stable for 24/7 usage, but am not sure if ramping the voltage up any higher than currently tested is going to get me there.

Are these voltages and quad-core speeds expected? Is this a "lousy" chip based on the O/C results above?

Consider these are all under Vapo LS. I got it maxed cooling for now while I probe the O/C curves for this chip. At 3.73GHz and 1.45V (BIOS setting on Striker) speedfan tells me my chip is running at about -5C with 4 instances of Orthos small running.

Also the BIOS set voltage is goofy, not sure what feedback to trust. I set the BIOS Vcore to 1.55V, the Vcore hardware monitor in the BIOS reports Vcore is actually running at 1.40V while in Windows XP with Asus probe software it reports Vcore is 1.45 to 1.44V at idle and 1.40V at load. Which one of these measures of Vcore do I need to be concerned with exceeding Intel's spec?

 

[Duvie]

God it would be nice to have phase cooling......

 

[cmdrdredd]

you should probably head over to XS and ask this stuff...people here usually aren't interested in WR overclocks.

 

[Rubycon]

I have to set 1.6V in BIOS just to get 1.5V under full load. Quad is sucking the regulators dry. :Q

I'd say your chip is holding you back. If I can do 4GHz at 80C under load I'd hate to imagine what it would do with phase cooling.

However it seems like core cpu's hit a brick wall where additional cooling does little to help them get further. LN2 doesn't count.

 

[The-Noid]

Speedfan (Nvidia Bios) reads about 7-9 degrees cooler then what it should be. Only mild overclocks can be allowed on the QX6700. With Quadcore you are overclocking a processor with a TDP of 135 to start. The Vapochill is only able to handle 175w of heatload. I would recommend that you check the evap temperature. Your processor should never cross 0C as this will cause condensation under the IHS. If your evap temperature is below -37 to -35 on the vapochill manager you are overloading your vapo. At -33 your system is in the range of where it will completely reset and cause permanentdamage to the evap.

 

[Cooler]

I think the wr for QX6700 ~4.6ghz

 

[Idontcare]

Originally posted by: Yoxxy
Speedfan (Nvidia Bios) reads about 7-9 degrees cooler then what it should be. Only mild overclocks can be allowed on the QX6700. With Quadcore you are overclocking a processor with a TDP of 135 to start. The Vapochill is only able to handle 175w of heatload. I would recommend that you check the evap temperature. Your processor should never cross 0C as this will cause condensation under the IHS. If your evap temperature is below -37 to -35 on the vapochill manager you are overloading your vapo. At -33 your system is in the range of where it will completely reset and cause permanentdamage to the evap.

At 3.73GHz/1.5V (BIOS) my vapoLS evap temp typically runs -45C to -50C at idle and climbs to high as -30C under full load.

I am not sure I am following you though in regards to your logic of how and why this could ever be bad.

Also I do not follow why a processor under 0C is a bad thing, my understanding is that with the heat-tapes in place there is no opportunity for freezing the humidity onto the board or CPU.

You have much experience though with the LS's so I am very much interested in understanding what you mean here, please help me understand.

Also, regarding voltages, how high is to high and what report on voltages should I be using as my gage for this voltage ceiling?

Regarding temperatures, what should I use to measure temp? Coretemp just pegs the numbers at 20C, 17C, 15C, 15C no matter the temperature and I am assuming this is because of the phase cooling. What program is good when using phase?

 

[Idontcare]

Originally posted by: MS Dawn
I have to set 1.6V in BIOS just to get 1.5V under full load. Quad is sucking the regulators dry. :Q

I'd say your chip is holding you back. If I can do 4GHz at 80C under load I'd hate to imagine what it would do with phase cooling.

However it seems like core cpu's hit a brick wall where additional cooling does little to help them get further. LN2 doesn't count.

Thanks for the read-point MS Dawn, I will shoot for 1.5V under load and see if 4GHz is stable.

Are you using Striker or eVGA or other?

 

[The-Noid]

Its over 0C that is the problem. You can go to about 5-6 degrees before your processor starts to have problems. The problem is that with the heaters and also the dielctric grease you can not get anything under the actual IHS of the processor.

If your processor temperature reaches the dew point, which depending on the area can be lower then others you start to develop condensation inside the actual CPU itself. It is the cold that is saving your cpu from problems.

As for the evaporator is only made to handle a certain amount of heatload (180w max). After that it has no where to take the heat and will end up overloading and causing permanent damage. If you take the side panel off with a hex screwdriver you will see how hot this thing really is. I have literally burned my hand on the compressor when it was running quad (after it being off for 1 hour). If you see an EVAP temperature of lower then -33 I would worry as that is too high a temperature. The way phase works is it creates a certain amount of pressure to change the gas into liquid. If you keep increasing heat pressure has to keep rising to cause the vapor to turn into liquid. After a certain point the compressor can not hold this much heat load and will completely shut off. This is also not great for the compressor as it is working double time to what it should be.

For the quad and actual temperature reading I noticed the delta to be about 40 degrees from Evaporator temperature to what the cpu actaully was. There is not really an accurate way to measure temperatures on phase so you mostly have to look at your evaporator and make sure that contact is good.

 

[The-Noid]

From what I remember from Chem 120 a few years ago (P1V1)/T1 == (P2V2)/T2. (Pressure, Volume, Temperature) Volume should be constant, so as temperature goes up, pressure must also rise inside the compressor to compensate. That is why it is not recommended to hit such high temperatures on the evap as you are causing the internal compressor to work overtime, and in th end it just can't handle it.

 

[Rubycon]

Originally posted by: Idontcare

Are you using Striker or eVGA or other?

P5WDG2 WS Professional