Q6600 overclocking problem

Ok so the mosfets are the small black chips as seen in the pic below.

gaep45ud3lrmosfets.png



The heatsink which I used on mosfets was a small blue one with thermal pad as shown below which I took off from my brother's dud graphics card.

imag0126so.jpg



Luckily it fitted fine and was small enough to be placed on top mosfets aswell between chokes.

imag0124qw.jpg



imag0125v.jpg



However with bios vcore still at 1.4375V, my PC ran IBT fine for 2-3 runs before it crashed again.
I am just gonna have to live with the fact that in terms of IBT stress testing, my motherboard has reached it's limit for [email protected]. Anything above and my motherboard can't handle the stress and crashes.

I am running my [email protected] anyway knowing I am Prime95 stable and for me this is fine :).

However today I have learnt something new about mosfets cooling.

Cheers mick and the rest:)
 
Last edited:
Mick-1965 said:
I take it you still cant up the core voltage to help stabilise it, now you have the heat sinks on.
Click to expand...

Yup.

I increased bios vcore to 1.45v giving me load voltage of 1.408v. Ran IBT and after 20-25s, PC crashed.

I then increased bios vcore to 1.50V giving me load voltage of 1.456v. Ran IBT and PC crashed after 1 sec.

So as stated before my motherboard can't handle higher voltages and I think this is again directly related to VRM (4 phase, mofsets, chokes, capacitors etc). Need much superior VRM.

Also during IBT, I can hear the squealing/hissing noises coming from the motherboard area.
 
WingZero30 said:
Would you still have those settings with you to give some rough idea?

I am going notch by nocth of bios vcore until I can see where I face serious problem then start changing some other values like PLL, VTT etc.
Click to expand...
The settings i had on the above voltages at 3.8ghz were.

PLL, 1.58
NB, 1.34
VTT, 1.34

Im almost certain that i used the same settings at 3.6 as well, (was quite a while ago), vcore at 3.6 was 1.400, 1.51 ono @3.8.
 
It certainly is.:) That was pre lapping and upping to 3.8ghz, temps were still the same even though i had upped the voltage considerably.

3.8 ss
untitled-2.jpg


When that 3.6 screen was taken i had the voltage at 1.41 something in bios, i was actually able to get it lowered to 1.400 exactly, dont know if i saved the screenshot or not though.

EDIT, found the screenshot with lower voltage at load, 1.408.

tempsLarge.jpg
 
Last edited:
setter

Your Q6600 was an impressive cpu. Seems to be even better than mine in temperatures department even though your Q6600 VID was slightly higher than mine.

Your lapped results are even more impressive.

My Q6600 has advantages in 2 out 3 main factors.

It is a low VID cpu
It consumes less power/voltage to run at high clockspeeds compared to majority of Q6600s. (Most Q6600s seem to hit 3.6Ghz between 1.45-1.50V load voltages from googling)

However it's main disadvantage is that somehow it seems to run hotter especially compared to your Q6600.

All the results I have posted in this thread are after lapping TR IFX-14 and Q6600 and they seem to be on par with your unlapped setup.
 
I heard a lot of low vid chip users reporting the same, less volts for a given clockspeed, but hotter running. The fans used on the ultra were, arctic cooling 12025 pwm's @3.6 and sharkoon silent eagle 2000's @3.8. Push/pull setups on both. The p5q deluxe really helped, the cpu wouldnt go over 3.0 on a supposed quad friendly evga 680i a1 rev, no matter what settings i tried.
 
Asus P5Q deluxe I think is the best 775 board.

My motehrboard is the stripped down version of the famous Gigabyte GA-EP45-UD3P which is supposed to have similar performance capabilities compared to Asus P5Q Deluxe.

Unfortunately when I was buying motherboard last year, UD3P wasn't on the market anymore especially in UK while UD3LR was available in plenty. So decided to go with the latter.
At that time I didn't have enough knowledge about motherboards especially with regards to cooling solutions and power-phase voltage regulators.

Oh well I am happy that atleast my mobo can withstand prime95 loading for [email protected] and I am happy with that:).

Many times I have heard that 10-20 passes of IBT equal 40 hours of Prime95. I don't agree with that for the simple fact that both tests are totally different.

IBT (LinPack) uses Gaussian Elimination method to solve matrices.

Prime95 makes use of Fourier series if I am correct hence you get small/ large FFts which stand for Fast Fourier transforms whose base is the fourier series and involve calculus and mathematical series.

Now I have studied both methods during my engineering degree and I must say I found Fourier transforms to be more complex than Gaussian elimination method.

IBT is cyclic in nature where it places immense load on the cpu and when outputting results, it momentarily slows down and then goes back up. I think it is the way LinPack is programmed that makes IBT/LinX so harsh.

Prime95 on the other hand is more constant on loading and is definitely ideal for long term stability.

So imo they can't be directly compared although both are very useful for overclocking purposes as they are stressing the FPU (Floating Point Unit) of the CPU or the math co-processor as xsistor put it:)

I found a nice article on Prime95 which gives more insight :

http://www.playtool.com/pages/prime95/prime95.html
 
The p5q dlx can be had fairly cheap now if you were looking for one. Cheers for the link on p95, good point about comparisons between prime/ibt. Ive always found it best to use both for stability testing, ive seen me passing x amount of runs of one, yet the other can fail in a few minutes.
 
Guys

Back again

I wanted to check the 12V reading during load.

PSU is OCZ Modxstream 600W. It has two 12V rails each capable of delivering 25A.

Now I am finding it very strange that in case of Prime it is around 3.6-3.7V. For IBT it is slightly higher at around 4.3V. As stated before IBT still crashes.

I would have thought that during loading, it is supposed to display value around 12V:confused:.

Could it be that both HWmonitor and Speedfan are wrong or that it is either 4 phase VRM or the PSU itself?


psuvoltagetestprime.png


psuvoltagetestspeedfan2.png



psuvoltagetestibt.png


psuvoltagetestspeedfan.png
 
hello WingZero30

I got same problem.. but mine is this:

got a Q6600 G0 . mobo MSI G41M-S01 (same as G41M-P26)

the bios dont have vcore settings.. so i checked in AIDA64/CPU-Z and got this.. VID 1.32 (too high) ...

running on stock speed @ 2.4ghz..got problem in OCCT/IBT ... error or DROPs (cpu speed go to 6xx MHZ after certain load) and the time in IBT increase too much.. so the cpu really drop the speed.. Ive checked again using ThrottleStop and the result:

CPU CLock modulation: Always 100% , so no problem with Speedstep throttling or anything similar

Chipset Modulation . drop to 25% , so this is the throttling problem , Ive checked that is the OCP ( over current protection ) of VRM´s .

so I set Chipset Modulation and CPu Modulation to Always run at 100% , the results . no drops! stable times in IBT ( 26secs each test)

so I jump to overclock my cpu to 2.8 , 1.32 vcore , run IBT again and no drops (with chipset mod 100% in throttlestop)
I close Throttlestop and IBT drops again , so Ive found that chipset modulation is my problem...

Try with ThrottleStop and report the results.. hope I would help you :)



Uploaded with ImageShack.us
 
You must log in or register to reply here.
Back
Top Bottom