Way Too High Temps on E6700

[BGVirtual]

I have bought boxed E6700(L629A831), noticed that with the boxed HSF it was running at like 44C(Idle)/70+(Load) after running TAT for 5min, at stock 2.66Ghz and Auto Vcore, which is 1.38 Idle, 1.37 Load since my P5W64-WS Pro overvolts. Changed cooling to Tuniq and AS5, during the Tuniqs installation process though one of the screws broke, decided to give it a go with 3 screws untill i hopefully get a replacement of the screw. To my surprise temperatures we again like 43/65(too little AS5 was applied 1st time), after 5 remounts best temps i got were like 41/58 (10min Orthos). My retailer didnt answer my request for a replacement so i decided to fix it by myself, did a new remount and fixed the 4th position with 2 case-to-motherboard screws, the pressure that was applied was roughly the same as the screws, maybe just a little bit less:

http://img143.imageshack.us/img143/5901/00221cimg3508xe0122258lpy6.th.jpg (http://img143.imageshack.us/my.php?image=00221cimg3508xe0122258lpy6.jpg)

Though the temps were again crappy ~41/56. I think its not an issue of an uneven pressure since temps got marginally lower compared to the 3 screws mount. I have inspected the IHS and the coolers base for concaveness and they seem flat to me, there is no contact between the base and something else around the socket. Temp sensors also seem to report correctly, for example 100% load at 3.8Ghz, 1.5-1.55 Actuall Vcore and 60C+ in Idle leads to restart :) .

Finally i changed the Vcore in BIOS to 1.27 so i was getting "the stock" 1.32 under 100% Load, thats what i got on 2.66Ghz and 10min of Orthos:

http://img96.imageshack.us/img96/9828/cpuxa1vl5.th.png (http://img96.imageshack.us/my.php?image=cpuxa1vl5.png)

41/56... lol, thats only like only 2-3C better then what an Intel HSF should do.

Little OC attempts, 3.6Ghz, 1.41Vcore, ~1.5hrs of Orthos:

http://img185.imageshack.us/img185/6207/1hrgl5yt2.th.png (http://img185.imageshack.us/my.php?image=1hrgl5yt2.png)

50/70 hot

and 4.05Ghz, 1.50 Vcore, 1Mi PI only :

http://img339.imageshack.us/img339/7924/405ghzta7.th.jpg (http://img339.imageshack.us/my.php?image=405ghzta7.jpg)

57C+ idle


I have tried with the case opened and lying on the floor(assuming that it might be because of the screw) but temps dropped just like 1-2C. I think that i either cant spot my IHS/base concaveness or my CPU had got some really bad contact between the IHS and the crystal, because even after 1hr of Orthos the Tuniqs heatpipes arent warm at all, only got "obvious warmness" when i turned off all fans and let the CPU @ 100% Load to see if its gonna throttle.

So should i lap this one and hope that my temp issues will gone away, or RMA-it to Intel saying something like "It stopped functioning at once, i havent clocked it" ;) and hope for a better replcement ?

edit: yes, i do have a rear exhaust fan, thats the case :

http://img412.imageshack.us/img412/3766/cimg3519xs5ov7.th.jpg (http://img412.imageshack.us/my.php?image=cimg3519xs5ov7.jpg)

[smsmasters]

Have you got a rear exhaust fan?

I get 35 idle, 70C load with orthos @ 3.6ghz 1.43V with my e6600 + Arctic Cooling 7 Pro + AS5.

In real world applications, by load temps rarely reach 60C+

[adamsleath]

i would say that i'm about the same as you....57c idle, i dont think you'll be running that.

i lapped my cpu; it was convex as hell when i got it. as5 aswell here (a very thin layer) and i have tightened my tuniq screws as much as i dare...they are tight.

im 47idle; load in orthos is up to around 60c @ 1.46v (at 1.5v my load went up to nearly 70c)

id say at 1.41v i'd probably be near your temps...maybe even higher.

at 1.3v my cpu is about 26c idle. coretemp....i wish my cpu clocked high at 1.3 :)

[Blackoberst]

Make sure you have PECI enabled in BIOS, and make sure that CPU temp doesn't go over 60C, reported by ASUS Probe, or Everest. That's Tcase, and max for that is 60C. As for the temps reported by CT, they shouldn't reach 85C.

[adamsleath]

yeah my cpu craps out at 65c asusprobe...just crashes.

[BGVirtual]

i have also tightened the screws to the max, thats why one of them broke :P . Everest, TAD and CT are all reporting the same +/- 2C, PECI is on, its simpy going up to 60C with normal frequency and the only settings changed in BIOS are vcore,1.32 real, vDIMM 2.1, and the ram timings. It simply refuses to cool down :(

My ambient temps are normal , even abit lower since my heater isnt working properly :D, airflow is perfect, opening the window after 22:00 makes temps drop with 8-9C for 5min :D

[dekruyter]

It's kind of hard to follow all your stock and overclock settings/temps,
but, I'll throw in that lower to mid 40's idle and lower to mid. 50's load w/ a mild overclock is pretty good for e6600's or e6700's.

[adamsleath]

what is PECI?????

anyway is your room hot or not enuf airflow...anything like that?

you sure that cpu ihs of yours is flat?...dunno maybe the tuniq is just an overrated pos :rofl:??

but for sure 41c idle at 1.32v is unusually high.....i'm inclined to think that your temp sensors/readings are wrong...

(or mine are?)

[Brama]

what is PECI?????




It is 1 wire serial bus that carries the temperature informations of new Intel chipsets and cpus. I was not able to find more infos.

By the way my E6600 cpu starts thermal throttling at 80 C°, measured by internal digital sensors, even if some utilities say that it would start at 85 C°.

[adamsleath]

mk thanks....

[BGVirtual]

usually when PECI is active PCProbe reports like 10C lower temps then TAT and CT.

[adamsleath]

i see...well thats what my asusprobe does.
37c probe = 47c ct here; i can confirm that.

so... if your temp readings are accurate then wtf is going on? does your tuniq feel warm to the touch?? at idle or under load?...mine stays cool aand at load only mildly luke warm...

if your mounting is tight and your cpu ihs is flat and has good contact with the tuniq base then that only leaves airflow and room temperatures, unless something else is making the tuniq hot/warm.

[BGVirtual]

IHS seemed ok, measured it with a flat pencil on it, seemed to be even at the same level from the left to the right edge, didnt saw anytihng wrong(though edges could have been 1mm higher and i didnt noticed that).Im not 100% sure of the temp readings are correct but when it idles in 60C+ and i try to load it, it restarts within a second at various voltages. On load Tuniqs heatpipies are just "a little warmer" then in idle. Room temp is like 22C.

[adamsleath]

anyway...if you can live with 3.6ghz 50/70 is alright.

60c idle at 3.8...dats not going to happen either is it? i was running 3.7ghz at 1.57v at 55c idle for a while...it crashed ultimately...i got a 'cpu init' failure...and this was after id' run orthos and been using the computer for 2 days. i think it just got too hot.

whats your room temp.?

[Lestat]

my scythe Infinity gave me the same crappy results, so i got rid of it.
nothing i did made any difference.
so i went back to water.

[adamsleath]

well there you go if you want 200 more mhz you'll need water :) maybe you'd even get 4ghz with water..who knows.?

well if the cpu sink isnt hot...and your room temp is 22c (much lower than mine at 26c) then either your cpu is overheating for some reason or the heat just isnt transferring to the heatsink.

sorry i can't help...but i tell ya i've seen dudes with lower temps than mine using the "same" cpu sink and similar voltages...so maybe diff cpus just act differently cos they are diferent batches etc.

[BGVirtual]

So i suppose Tuniq is simply overrated piece of metal and its only for the "cool" CPUs, i should have never bought it :D Yes maybe, i can boot into Windows at 4ghz with 1.47 real and opened window:D Yes, after all maybe its just some bad luck :mad: . :slapass: Intel . Hmm another heatsink, as i said Intel HSF 44/61C + after 2-3min of 50%load in TAT and 70C+ after 1-2min of TAT, though at 1.37 Vcore, dunno if you can say straight that CPU should be concave, Load temps diff. btw 1.32 and 1.37 is like 5-6C.

[adamsleath]

my first choice was tt big typhoon...then scythe ninja...
scythe imine/infinity...

i've heard good things about ttbt, but honestly there's no way i'll muck around with other heatsinks unlesss something radically better comes out...the next step is water at the moment....but if you try another heatsink let us know how it goes. :)

[ineedaname]

I'm having similar temps with my TTBT and E6600. I've reseated it like 5 times and i get about 45C idle and 61C load at 3.6ghz.

Its kinda pissing me off cuz i got a 150CFM Delta blowing on it and it's not helping at all. I'm considering lapping my cpu. Like you I also find it really weird cuz the top of the heat sink base and heat pipes are not even warm. My heat sink base is lapped and i'm using liquid metal for my tim.

Even when the cpu is at stock speeds it is at 4xC at idle......

[smsmasters]

When I enabled PECI on my P5B Deluxe speedfan reports my temps about 10C lower than tat or core temp. Is this ok?

[Revv23]

have you checked contact? i mean by putting an exteremly thin layer of thermal paste on the cpu, mounting cooler with moderate pressure, then pull the cooler, and see where the paste is stuck.

My bet is that the center of the sink has no paste on it, because your IHS isnt flat.

Two things you can do is either adjust the amount of thermal paste you use or lap, lapping would be significantly better.

Intel IHS's in my experience almost HAVE to be lapped to get semi decent temps.

i got a dramatic increase in overclocks and a big decrease in temps from lapping my 6400..

[Lestat]

one thing ya gotta remember guys is once you hit around 1.45v these chips start getting REALLY hot. they put out alot of wattage and require damn good cooling.
its why you see me openly challenging people who are claiming such high clocks and volts on air, in order to do that you need the mobo sitting on a table with the A/C blowing on it or by an open window with the cold winter air blowing in.

air @ 1.45v and up is easy but you gotta have your cooling set up perfectly.

your setup now has way more, and i mean way more air flow than the average user will have, those lian li V2000 cases have an insane amount of free air space and room.

why its running so warm can be only a couple of things.

can you verify that those temps are correct ? is the air coming off it actually pretty warm ? are the heatsink fins getting that warm ?
do you have a thermal probe to install ?

[BGVirtual]

Well like one guy said, I think the low temps on the coolers base is due to the fact that Conroe seems to have problems with the heatspreading, because of the bad contact between the heatspreader and the crystal. When you combine high volts with high frequency it simply overheats because it cant disperse the heat properly, and thats why the temps are jumping drasticaly, and the temp on the top of the heatspreader is pretty normal(and the coolers base), but inside the cores temp is like 80C+. And probably they are 2 extremums, a CPU with very bad thermall contact btw the heatspreader and the crystal, like mine, and a CPU with a really good contact, like this one :

http://img73.imageshack.us/img73/4104/10562xc9ay3.th.jpg (http://img73.imageshack.us/my.php?image=10562xc9ay3.jpg)

and he is using Tuniq too. And thats a E6600 with a high-end water cooling:

http://img148.imageshack.us/img148/3396/coretemperaturetrotllu9.th.jpg (http://img148.imageshack.us/my.php?image=coretemperaturetrotllu9.jpg)

So again its probably a luck pot :)

back on topic: From what i can remember the thermal paste was evenly distributed in the centre with some micro areas being not covered with a paste. That was after a mount with a very-thin layer, and temps of 60C+ idle in BIOS.With a thicker layer i got better temps and best temps with the "blop in the middle" method. Seems that more paste gives better temps which might really means that IHS is not even somewhere on the surface.

The Air Flow coming from Tuniqs fan is aways cold no metter if the CPU idles at stock or benches at 3.6Ghz, fins are also aways cold. I dont have a thermal probe at the moment but i can load the CPU with all fans disconnected and see at what temp it will start throttling, also only so the fins and the heat pipes are getting hot.

[Lestat]

yes this has been an ongoing issue with the conroes.
the stupid solder they are using between the core and the IHS is not allowing proper contact to be made, and its definately not the best thermal conductor.

there was a screenshot a while back showing the solder only on 1 corner of the cor, thus making the rest of the cpu radiate an enormous amount of heat.

the whole solder thing they are doing is simply nuts, it never should have been done this way.

[Millyons]

Guys anyone that has higher temps just needs to lap the chip, at one point i tested 4 e6600 and one was way hotter than the rest like 10c more at idle and even more on load, the thing is it looked really flat i also started thinking its the core to ihs conntact, i lapped it in 30 min and my idle temp dropped over 12c and load even more.....although i still do believe some chips dont have as good contact between the core and ihs but also think that it isnt as much of an issue as concave/convex ihs

[topboy]

FAO BGVirtual

I also have a Lian-Li 2000 case and was getting really high temps.
Its because there is not enought air being pushed into the case from the front and then out of the back exhaust fan.

Let me explain.

Im on watercooling and only had my cpu block in the loop at the time. I also only had 1 120mm fan on back like you to exhaust the air out.

I found that the chipset was getting real hot and the heat was getting trapped around the chipset and the cpu and being trapped by the Graphics card. The heat simply couldnt go anywhere cause of the lack of airflow.

I then placed a 120 mm fan in the front of the case in the bottom 3 5/14 bays pulling air into the case and being exhausted out the back 120mm fan.

This dropped my temps down by 12c.

Give it a go on the v2000 case. you need to get the air flow going through these cases with the Intel chipsets and the C2D Proc due to the design of the case being a BTX style thing.

[serialk11r]

yes this has been an ongoing issue with the conroes.
the stupid solder they are using between the core and the IHS is not allowing proper contact to be made, and its definately not the best thermal conductor.

there was a screenshot a while back showing the solder only on 1 corner of the cor, thus making the rest of the cpu radiate an enormous amount of heat.

the whole solder thing they are doing is simply nuts, it never should have been done this way.
Hmmm... But solder is tin by the time the flux and lead (if there is any) burns out and tin has a thermal conductivity of 70W m^-1/K^-1 as opposed to around 5 with AS5/ceramique/alumina. So unless they were being idiots and had a huge amount of solder then there shouldn't be a problem. But then we run into the second problem, which is the IHS being screwed up and the solder not making good contact. I am lucky, my IHS is perfectly flat :P but I believe there are a lot of people with very concave IHSs, and when you have an IHS like that you either lap it or live with the crappy temps.