HalloweenWeed

rafale

There is only one thermal sensor too (I don't know where it is located) and some VREG...

The sensor is imbedded in the Vreg package itself, just like the CPU, it would be fairly useless otherwise.

 
It makes sense but why would we only have one read out and which one is it reading?

HalloweenWeed

The sensor is imbedded in the Vreg package itself, just like the CPU, it would be fairly useless otherwise.

Proof that they exist:
http://www.nxp.com/documents/application_note/AN10137.pdf
 
"An N-channel power MOSFET with monolithically-integrated temperature sensing and clamping diodes that internally monitor the temperature of the MOSFET chip. Designed for high current applications, ..."
 
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=5158040
 
"Abstract
In this work, difference - temperature (DeltaT) sensors based on the Seebeck effect integrated into a common drain smart power MOSFET technology are presented. The sensors generate a voltage signal proportional to the DeltaT. The highest Seebeck coefficient measured was 0.92 mV/K. This result was achieved with a p - doped silicon and n⁺ - doped poly silicon Seebeck DeltaT sensor. Power MOSFETs with embedded Seebeck DeltaT sensors are, to our knowledge, demonstrated for the first time."

HalloweenWeed

HalloweenWeed

The sensor is imbedded in the Vreg package itself, just like the CPU, it would be fairly useless otherwise.

Proof that they exist:
http://www.nxp.com/documents/application_note/AN10137.pdf

"An N-channel power MOSFET with monolithically-integrated temperature sensing and clamping diodes that internally monitor the temperature of the MOSFET chip. Designed for high current applications, ..."

http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=5158040

"Abstract
In this work, difference - temperature (DeltaT) sensors based on the Seebeck effect integrated into a common drain smart power MOSFET technology are presented. The sensors generate a voltage signal proportional to the DeltaT. The highest Seebeck coefficient measured was 0.92 mV/K. This result was achieved with a p - doped silicon and n⁺ - doped poly silicon Seebeck DeltaT sensor. Power MOSFETs with embedded Seebeck DeltaT sensors are, to our knowledge, demonstrated for the first time."

 
I never doubted you... 

v1ral

ok i plugged in hweeds setting with some help of drey got it to 199 base clock.
boots and everything but doesnt go to the second set of linx..temps are around 84.. i chamge vtt to 100.

Raise CPU Vcore to 1.325. If that don't work, then try raise vtt just to 150 @ 1.3125Vcore. You may hit 89C. If this works, you may be able to raise Bclock to 201 or 203 @ 89C max.

okay guys i tried the 1.325/+150vtt got to 2nd pass then failed temps are 85 gonna bump vcore

Okay update..
It failed..however.. same gflops.. what should I goose up?
edit: temps were the same 84c

you dont? why?

I think someone suggested this already, but you should try all this with all your ram timings set to auto.  Yes, it will set your timings high, but at least you'll know it's not that limiting you.
 
I tried that when I hit a "wall" at around bclock 208.  I was then able to go up to 216 and work on the timings AFTER I got the bclock stable.
 
Just a suggestion.

Because your temps will soar when you have to raise your Bclock and/or VTT so high to reach 201/203, you will not have any wiggle room left to make 205.
 
Loosening your mem timings may help, won't hurt to try, but for me it didn't, it resulted in loss of available mem. But personally, if it were me, I would plug in numbers instead of going W/auto. I would add about 15% and round up when it wasn't close to an integer.

What do you by it "failed" BSOD stop responding etc? and if it BSOD which one?  To reach 4.2ghz i actually had to reduce some voltages.  its not always VTT or Vcore sometimes CPU PLL or QPI PLL need to be adjusted up or down.  Also i notice on your template your using the 49 BIOS have you tried to update to the newest? 

HalloweenWeed

Because your temps will soar when you have to raise your Bclock and/or VTT so high to reach 201/203, you will not have any wiggle room left to make 205.

Loosening your mem timings may help, won't hurt to try, but for me it didn't, it resulted in loss of available mem. But personally, if it were me, I would plug in numbers instead of going W/auto. I would add about 15% and round up when it wasn't close to an integer.

Or that ^

Oh. That's disappointing. At this point, I would defer to other's suggestions, until you try another BIOS. As crusheddream said, your BIOS revision is lousy. BIOS 77 was good, for later BIOSes, but I would suggest you use:
 
EVGA X58 Classified (E760/E761/E769) - S61J BIOS Released
 
Many ppls have touted this as the bestest BIOS for OCing a 920. They also said their temps were higher when they tried newer BIOSes. But I suggest you give it a few days to see if anyone else has a better idea, see if anyone else just has a quick 'plug this number in' suggestion first. Changing your BIOS at high OC is a big deal, you need to back your OC down a bit before trying it with the other BIOS, as you will find your OC works just a bit differently as far as specific voltages, and hopefully temps.
 
And remember, don't load a previously saved BIOS profile after flashing to a diff BIOS, and just to make sure immediately afterward (after you plug in your mem numbers) save all 8 BIOS profiles to clear them out. Standard BIOS update protocol (reminder): After flashing, with the PSU on and computer off, press and hold the CMOS reset button for at least 10 seconds. Write down all your BIOS settings before flashing, of course.
 
I thought maybe we could work with your BIOS revision, wanted to try it, but apparently not. Sorry I didn't request this sooner.
 
Oh yeah, and if you ever are going to upgrade your CPU to something other than a 920, at that time you should update your BIOS to something much more current. You may not even be able to get into BIOS, nor flash, with S61J and a 960 or 990.
 

update:
i changed timings to 9-9-9-24..1t. been running a little over an hourmax temps 84. vcore is below 1.3
bummer about your ram but i am hoping and praying my ram isnt bad.....
also what would it mean if my desktop gadgets stop working.? at 199 base i kept getting these.

yes.. its weird.. if/when i loosen timings it runs cooler..
idk how it is now i am at work.. but i set it to.run 20 pases had my girlfriend plug.in the settings

v1ral

yes.. its weird.. if/when i loosen timings it runs cooler..
idk how it is now i am at work.. but i set it to.run 20 pases had my girlfriend plug.in the settings

LOL!  Remote overclocking.