• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

e8400 Tjmax can't be 100

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

don'tknow

Member
Joined
Sep 1, 2008
At 100 Tjmax, my idle temps are 44C and load temps 55C with RealTemp 2.70.... this can't be right. Yet supposedly Intel confirmed that this is the Tjmax? On a C0 stepping?

With a Xigmatek-S1283 on the CPU, at stock clocks and 1.25v, and with push/pull fan configuration. TIM used is AS5 and I applied it as follows; in between the heatpipes where the gaps are, and one thin line on the middle of each heatpipe, about 60-70% the length of the chip. Possibly slightly too much to make sure that it's all covered, but not excessive. The "Crossbow" retaining bracket from Xigmatek is also used, all 4 screws are tightened.

Yet my temps are higher than my older 90nm CPU fully overclocked to its stable slocks... and I'm using the same cooler, except I put too little TIM on the 90nm which I found out after I took the cooler off to install this new CPU on the new mobo. It wasn't covered by TIM very well. Also, it wasn't using the push/pull fan configuration either, because of the way the method for installing AMD chip works (the heatsink is set horizontal instead of vertical). It was an Opteron180 OC'd to 2.8 at 1.4v, with temps being Idle ~35C, load ~53C.

So my temps on a 45nm with a better cooling method and stock speed, are higher than my temps on a 90nm with worse cooling and OC'd to max stable levels? What am I doing wrong? Or is Intel lying?
 
the CPU might be warped. Try do a razor test.

Get a razor-blade and put the "cutting edge" on top of the CPU as if you were going to cut it in half, but dont apply any pressure of course. put it up to the light and see if you can see any light. If the CPU is warped, you will prob need to lap it.

I get my temperatures from everest, and i trust the hottest one. Its a known issue with the 45nm processors that the temperatures are not reading correctly. This is true for all 45nm processors, your not alone. You kinda have to use judgement, but as a rule, i trust the hottest reading to be on the safe side, just make sure the temp changes, cos my first core temp dont change from 50 at all, but the second one does.
 
I think the E8400 temps with 100 tmax are too high also. With a 23C room temp, my system idles at about 42C-45C, which seems kind off high despite the 4100 MHz o/c. I liked the 37C-39C idle readings with the old 95 tmax, which seems more realistic.

The cpu temp in the bios reads at 42C, while it is 29C in Windows. Why would the cores idle at 15C higher than the cpu? the cores should be more like 5-7C higher at idle than the cpu.
 
here's my temps in SpeedFan and RealTemp right now:




The 'min' and 'max' ones were when it was at 95 Tjmax, then I changed it to 100 and screenshotted.
The flaming core @ 51C is my gtx 260 GPU. Ignore the false 12v reading of '11.40', I checked it in bios and it shows as 12.096.
 
i also think temps aret reading right, my q9550 E0 @ 1.28v... the first two cores idel @ 50ish which cant be right. i dont trust lapping the cpu but i might need to lap my sink. but still i think something is up, my old x3210 (q6600) was cooler
 
Intel said that Tjmax was 100*C, but they also said these sensors are incredibly inaccurate and get worse the further you go from Tjmax.

The RealTemp documentation outlines how you can calibrate your sensors to get more accurate idle temps.

Ask yourself this...does this chip heat up your room like that older 65nm CPU?
 
Yep, DTS reading in idle are far off. Anyway, idle doesn't matter, try to load it with LinX/IBT and see distance to TJMax. If you're >30 you're OK.
BTW use latest beta build RealTemp.
 
this is with the newst realtemp. seems kinda high to me. but like i said my heatsink needs a lap, and prob cpu too

test1.jpg
 
My e8400 Idles @ 35 and 28 and loads @ 50 and 49 right around where I would expect it to be .

I used core temp 99.3
 
Just did the calibration thing from the RealTemp documentation and it showed same temps... my sensors are stuck. 266x6, 1600 mhz and 1.025v .. 44C idle. Putting full load on both cpu's would not move the temps, probably because the load temp is still less than the 44 it's stuck at.

Guess I'll have to assume it's always showing higher than it really is until it gets to around 60C+, which is where it should supposedly be accurate.
 
Back