FP IN9-SLI - New CPU instantly overheats

[KenboCalrissian]

Hello all! Obviously, I'm here because abit's site no longer offers customer support, and I'm finding myself burned by a purchase of a new processor for my old LGA 775 rig. I was hoping to make a cheap upgrade and then donate this machine to my girlfriend after building a brand new rig, but instead it's looking like I might have just thrown a wad of cash down the drain...

I recently purchased a new processor (Intel Core 2 Quad Q9400 Yorkfield 2.66GHz LGA 775 95W Quad-Core Desktop Processor AT80580PJ0676M), and had to update the BIOS before installing - I was previously on BIOS 12, updated to 16, admittedly from a third-party site since the drivers listed on abit's page all appeared to be corrupted. Within moments of turning on the computer with the new processor, if I enter the BIOS it's already reporting a CPU temp of 95C - which, of course, auto-shuts down the computer after it starts booting to Windows.

I believe I have ruled out thermal paste/improper seating due to the following reasons:
-My old processor (Intel Core2 Duo E4500 @ 2.20 GHz) still functions fine, idling at 45C after cold boot. Even after repeatedly swapping out the processors between failures, my results are consistent - I only see overheating when the new quad-core processor is installed.
-I've placed my hand near the CPU heat sink while the BIOS was reporting a temperature of 95C, and did not feel any excess heat. Also, after seeing this temperature report I immediately turned off the computer, unplugged everything, removed the heat sink and fan, and touched the CPU - at this time, it was still cold to the touch. Even though it was impossible for me to reach the CPU immediately for this test, I believe I was still fast enough to at least feel a little warmth if the CPU were indeed overheating.
-I've applied, removed, and re-applied thermal paste to the new CPU three times, each with the same results. Again, the old CPU never had any overheating problems.

This leads me to believe either the CPU is reporting the wrong temperature, or the BIOS is reading the reported temperature incorrectly. The latest BIOS version notes report that a fix was made for incorrectly reading the temperature on a 45nm Yorkfield processor, but this appears to be the problem I'm seeing here. Is there anything I can do to fix this?

Thank you,
Ken

[hesham1516]

Bios problem, quick google search reported lots of people having the exact same problem with that mobo+cpu combo as well as other C2Quads. Consider selling off the motherboard + E4500 and getting a different board since your one is no longer supported.
An alternative is to risk flashing a modded bios if you can find one.

[technodean]

Your E4500 is the old 65nm chip as apposed to the q9400 which is 45nm. So it sounds like your motherboard does not accept the newer 45nm chips. It may work but not properly. Had the exact opposite problem to you where my old 775 board would only use the 45nm cpu's.

[KenboCalrissian]

Found a work-around to this problem. Basically, all I had to do was enter the BIOS, disable Auto-Shutdown, and raise the CPU Overheat Alarm Temp to its maximum of 120C. Additionally, I downloaded CoreTemp to get an accurate reading of my processor temperature after Windows has loaded. CoreTemp is showing all four cores idling at 38C. Since I can no longer rely on the BIOS to warn me of an overheat, I set up overheat protection with CoreTemp (though given the difference between the idle temp of 38C and the BIOS thinking it's 92C, that max of 120C for the alarm might be conveniently set to go off when the CPU is actually around 75C - no way to really tell, I'm just making a wild estimate based on the difference in reported temp vs. actual temp).

One last thing I did was set the CPU fan to target 70C (another max value). I did this thinking it might reduce fan noise a little, but it's probably still running at max speed - fortunately, it's still very quiet. It's not the prettiest fix in the world, but a little research has shown that while BIOS beta 18 might fix my problem, it might also cause blue screens with Windows 7. Given that I plan on donating this machine in a year or so anyway, it's preferable over buying a new LGA 775 mobo.

Thanks for the quick replies! Hope this trick helps somebody out in the future.

[technodean]

Wish you luck, But remember your max temp for that cpu is 71deg.

[KenboCalrissian]

Thanks for the tip. Did some testing today and peaked at around 79C for what I'm trying to do with it, so I'll definitely be investing in a new cooler.

Interesting to note... my alarm did not go off when CoreTemp reported I reached 79C. This is above the actual temperature that I thought "120C" according to BIOS should have been. It turns out the inaccuracy is not merely adding a set number of degrees to the actual reading, so making any guesses based on a range between CoreTemp and the BIOS reported temperature is completely useless.

[hesham1516]

Thanks for the tip. Did some testing today and peaked at around 79C for what I'm trying to do with it, so I'll definitely be investing in a new cooler.

Interesting to note... my alarm did not go off when CoreTemp reported I reached 79C. This is above the actual temperature that I thought "120C" according to BIOS should have been. It turns out the inaccuracy is not merely adding a set number of degrees to the actual reading, so making any guesses based on a range between CoreTemp and the BIOS reported temperature is completely useless.

About time you picked up one of those fan controllers with thermal sensors is it not? It will be the only way to accuratly monitor what is going on.

[Bambooz]

Last Beta BIOS for the IN9 SLI is Version 18 (1.8?). Those ridiculously high fake temps seem to be a general "45nm CPU vs. Abit Board" problem.

BIOS on my Abit FTP mirror:
ftp://91.121.194.115/pub/download/bios/in9sli/

Version 18 is bare, i.e. only the .bin file. You'll have to download an older version to get the DOS flash tool and edit the .bat, or replace the old bios .bin with the version 18 one and rename it to the old name..

[KenboCalrissian]

Thank you all for your responses! It feels a little early to call it, but I think everything's working now. Below is some information that could be useful for others with the same problem later (or, worst case scenario, a reference in case it's not over yet).

So, the fight wasn't over after turning off the overheat protection and installing a new fan. I started getting blue screens (stop code 0x000000007F, usually param D but occasionally 8) pretty frequently after performing the steps in my last post. It took me a while to figure out it would only occur while using the peripherals - I could let it sit overnight running virus scans, but if I was typing, clicking, or playing on a game pad for more than 20-30 minutes, this BSOD was bound to happen.

I started uninstalling things one at a time, thinking it was driver related (more like praying it wasn't processor related), to no avail. Even CoreTemp - but just before I uninstalled that, I noticed CoreTemp was reporting the clock at 1.99 (and while running Verifier, it was fluctuating frequently between 2.65, 2.33, and 1.99). That's when I realized what was going on (or at least, I had a theory); because BIOS 16 thought my temp was so high, it was cutting the voltage to the processor - possibly to the point where it was under-powered, rendering it unstable (if not that, then the frequent fluctuations couldn't have been good for it either)!

That's when I finally took my chances with the Beta 18 BIOS provided by Bambooz (with my BIOS 16 disk waiting on standby in case things went very south) - even though I knew it would correct the heat readings, I was afraid to try it because I'd read some rumors that Windows 7 users blue screened a lot with that version (in other words, I'd just have a different flavor of the same problem), but on another forum, someone had guessed this might be because 18 has a weird habit of supplying roughly 16% less power to the CPU VTT than you actually tell it to. I verified this to be true for the VTT as well as the core. So, I set the voltages to a little above the mid-range where my processor voltages ought to be, and confirmed that the actual power supplied was landing right in the middle.

As of now, I've been unable to reproduce the blue screen with a method I had become certain would reproduce it (it seemed fighting "The Guy" in I Wanna Be The Guy was both an effective and appropriate trigger for that BSOD), and I've been using it normally for a total of 6 hours now (with constant user input). If I go the rest of tonight without seeing another BSOD, I'll post my voltages (what I set, and what's actually coming out according to BIOS) and anything else I set manually in case it'll help others seeing problems with BIOS 18.

[nzspst0]

Thank you all for your responses! It feels a little early to call it, but I think everything's working now. Below is some information that could be useful for others with the same problem later (or, worst case scenario, a reference in case it's not over yet).

So, the fight wasn't over after turning off the overheat protection and installing a new fan. I started getting blue screens (stop code 0x000000007F, usually param D but occasionally 8) pretty frequently after performing the steps in my last post. It took me a while to figure out it would only occur while using the peripherals - I could let it sit overnight running virus scans, but if I was typing, clicking, or playing on a game pad for more than 20-30 minutes, this BSOD was bound to happen.

I started uninstalling things one at a time, thinking it was driver related (more like praying it wasn't processor related), to no avail. Even CoreTemp - but just before I uninstalled that, I noticed CoreTemp was reporting the clock at 1.99 (and while running Verifier, it was fluctuating frequently between 2.65, 2.33, and 1.99). That's when I realized what was going on (or at least, I had a theory); because BIOS 16 thought my temp was so high, it was cutting the voltage to the processor - possibly to the point where it was under-powered, rendering it unstable (if not that, then the frequent fluctuations couldn't have been good for it either)!

That's when I finally took my chances with the Beta 18 BIOS provided by Bambooz (with my BIOS 16 disk waiting on standby in case things went very south) - even though I knew it would correct the heat readings, I was afraid to try it because I'd read some rumors that Windows 7 users blue screened a lot with that version (in other words, I'd just have a different flavor of the same problem), but on another forum, someone had guessed this might be because 18 has a weird habit of supplying roughly 16% less power to the CPU VTT than you actually tell it to. I verified this to be true for the VTT as well as the core. So, I set the voltages to a little above the mid-range where my processor voltages ought to be, and confirmed that the actual power supplied was landing right in the middle.

As of now, I've been unable to reproduce the blue screen with a method I had become certain would reproduce it (it seemed fighting "The Guy" in I Wanna Be The Guy was both an effective and appropriate trigger for that BSOD), and I've been using it normally for a total of 6 hours now (with constant user input). If I go the rest of tonight without seeing another BSOD, I'll post my voltages (what I set, and what's actually coming out according to BIOS) and anything else I set manually in case it'll help others seeing problems with BIOS 18.

dude if u are running good please let me know ur V cuz i gettint crazy with this !!i burn a ip35pro so i decide to put it back the fatality fp in9 and i'm getting freaky with the V 2 o 3 hours good them crash! reset craash reset even i cant format

[grigori316]

Hi Kenbo

I have the same mobo, where did you get the bios updates from?

[OilSheikh]

Q9400 is a quad core chip - it runs hotter! Your HSF seems to be the problem.
Most likely, it's ideal to cool a C2D chip but not aquad core chip.