Temperature stuck on E8500

L31mr0d

Ok the temperature is stuck on my E8500, it stays at a constant cpu diode temp of 33'c and 33'c, 30'c for each core respectively.

I've confirmed this with various temperature monitoring programs, plus it doesn't change one degree under load or idle.

From reading around this seems to be a common issue with E8*** Intel CPU's. So what should I do? Should I just continue to overclock it and monitor stability?

I bought it off adverts a good few months back but haven't got around to using it since my old motherboard died shortly before being able to use it.

Silenceisbliss

ooh, sensor broken...sucks.

if it were me, I'd overclock anyway but i wouldn't OVERclock the chip beyond what it seems to be capable of. Id see what other people with an e8500 and a tuniq are getting as overclocks and do the same. I wouldnt push it beyond what seems too high.

Hand on heatpipes aswell when stressing, it will give you a vague idea of what kind of temperatures your getting...

warrenty? or has it already been clocked? ah...sig...

L31mr0d

thing is my Tuniq has NEVER gotten warm to the touch. The case would have to have a fairly bad airflow for that to happen. I find the temperature of CPUs under the Tuniq just spike when under load and then quickly drop off.

I might just take your advice and read around on what others are getting. I'll limit my max vcore to ~1.45v. Plus hopefully stability will become an issue long before the CPU burns itself out.

I'm hoping for 4.5Ghz but without knowing temps I might not push it that far. I'd be happy with 4.2Ghz, 1Ghz above stock is nothing to be sniffed at. Plus apart from synthethic tests I don't know if I'll really see the benefit of the faster CPU speed.

Silenceisbliss

an idea just occured to me but i'll have to investigate it a bit to see if it's at all feasable.

at work here, we put temperature indicating tape on the A330/340 landing gear axels. It serves as a visual inspection to show how much heat the friction or stress has produced and transfered to the axel.

I know you can buy some pretty accurate thermal indicating paint and stickers like http://www.korthals.nl/e/Product/TOSe.html (only more accurate)

Im just wondering now would it be worth the while looking into this. (yes, i'm bored at work today...so ill have a gawk)

even to put a sticker on a copper tube and compare it to a temperature you can read off an older chip with the tuniq....

Ill check it out...why not

Silenceisbliss

just dawned on me;


My dad once had to borrow an infrared thermometer from work before, point the laser at something, it will tell you what temperature it is. pretty cool really.

if you could get your hands on one, set up a chip whos temp you can read, point laser at top of copper plate after like 10 minutes of orthos.

clock the known chip to different levels to get varieties in temperatures, and plot a table of CORE TEMP vs INFRARED readings.... see if it adds up and use that...

L31mr0d

what's weird is that I always thought there was a temperature probe directly below the cpu in the socket? That one shouldn't be stuck?

PogMoThoin

The motherboard should give a reading for the sensor in the socket, although not as accurate as on die, but it'll at least be something to work out your temperature delta from. Allow 10' or so difference, measure the temp at idle (you'll have a rough idea how hot it is) and work out the difference

awhir

most of the first e8**** batch's had faulty sensors , search back a few months on xs and you might find some ussful info.

Flaccus

Some still do, even the 9x series. Usually though the sticking is temporary and can unstick underload. I wouldn't worry too much. Programs like Coretemp, Realtemp etc.. are trying to read the distance to tjunc max, and tjunc max is an unpublished figure by Intel (though these programs assume it's the value in one of the registers on the chip) who say such ways for measuring temps are not recommended when overclocking anyway and are purely designed from the point of view of measuring idle temps and provding information to the cpu regarding when to throttle back. Like others have said you could use other monitoring software utils out there (which use the Tcase value, and max tcase values are published by intel) to give you a general idea of how the chips temps are behaving.

Also 1.3625v is the max volts intel spec on the 8500, so if on air as you are, isn't 1.45 a bit high (depite what people on xs say). Shouldn't you be on water at that voltage ?