Thuban Temperature discussions

[Titan7171]

Im having a hard time keeping this cpu at the same temperature as my deneb at the same voltage.......it reminds me of when I moved from my 6400x2 to my 9950. The heatsink could not handle the heat load.

While running stability tests with prime and others and the voltage at 1.57v on this 1090T Im seeing temps reaching up to 68c at this point I get a bsod and the system reboots. Where normally with Deneb I could hold the temps at 50c at 1.57v.

Obviously the extra 2 cores are having a huge effect on my watercooling

Its crazy because this means that I cant add the graphics card to the loop I have (what I thought) was a top end watercooling setup and Thuban is pushing the limits.

On a side note I probably have a dud as almost the majority of people are using lower voltage with higher clocks than I can attain with this cpu. The cooling is its salvation......but Id love to know what I could do with a better cpu....... Lucky its not mine

for those that dont remember the setup

4x18cm fans.........compared to a 12cm can



thats the rad with a 38cm fan on it



And the way I have it setup



So I think I have to add another rad!

[JumpingJack]

I am having a hard time just getting a physically meaningful temperature reading ... meaning idle is ~ 8 degrees lower than room temperature.

[Titan7171]

I am having a hard time just getting a physically meaningful temperature reading ... meaning idle is ~ 8 degrees lower than room temperature.

what motherboard?

[wuttz]

titan,
position that rad to freestanding and for the fan to blow into the other side of the rad without an obstruction (like the top of your case).
rad needs unobstructed airflow to maximize efficiency, might be time to look at another cpu block, koolance or heatkillers.

[ridney]

last time i ran prime on my x6 at 4.2ghz 1.575v for 30mins, temp reading from hwmonitor never exceeded 50c, i think it was about 48c IIRC at some 29c room temp. well i didn't have a screenshot as i fried the board for too much volts but temps were really good. i'm using a monsta with 6 fans in push+pull config and 2 small vgas under water too.

here's a screenshot of cinebench prior to burning my 790fx-gd70 mobo, check the coretemp reading, it's about 41c max at that run at 1.575 vcore




i agree with what wez is suggesting, maybe it could allow you to drop down the temps further down than 68c.

[PatRaceTin]

50C vs 68C

i think +2cores will not produce +18C hmmmm

[Titan7171]

titan,
position that rad to freestanding and for the fan to blow into the other side of the rad without an obstruction (like the top of your case).
rad needs unobstructed airflow to maximize efficiency, might be time to look at another cpu block, koolance or heatkillers.

Thanks, the rad rests on 4 legs about 3 inches high so its not laying on the case top......it was only like that when I was setting it up and testing it out. The air blows straight down and through the rad..........I didnt think it could be the waterblock! Hmm something with more mass to it? I idle at 27-30c.....its only after 20-30 minutes that the heat seems to overwhelm the cooling system. Seems like its not drawing heat away fast enough so it builds up over time with te Thuban where as the Deneb rose to a certain point then couldnt deal with the cooling and level out at 50-52c. This is at 1.575v.

[Titan7171]

last time i ran prime on my x6 at 4.2ghz 1.575v for 30mins, temp reading from hwmonitor never exceeded 50c, i think it was about 48c IIRC at some 29c room temp. well i didn't have a screenshot as i fried the board for too much volts but temps were really good. i'm using a monsta with 6 fans in push+pull config and 2 small vgas under water too.

here's a screenshot of cinebench prior to burning my 790fx-gd70 mobo, check the coretemp reading, it's about 41c max at that run at 1.575 vcore


i agree with what wez is suggesting, maybe it could allow you to drop down the temps further down than 68c.

let me get coretemp and try the same tests and see what I get compared to you......thanks for the pic

50C vs 68C

i think +2cores will not produce +18C hmmmm

Maybe a bad mount on the block or even bad contact maybe..........my last 2 965's were lapped thuban is not......the waterblock has also been lapped to match the flat surface of a lapped cpu.......so maybe this is the reason??

Gonna have to pull the block and see what kind of contact is being made on the surface of the cpu.

[CrimInalA]

coretemp gives individual CORE temps . the temperature you should look at (and probably Titan is looking at that) is CPU temp .
check this with rc tweakit or amd overdrive .

[Titan7171]

this is what Im getting for now

[wuttz]

titan, are those temps accurate? i've read somewhere here(posted by informal or particle i think) temp sensor is borked(again.)

[Lightman]

My temps according to PC Probe and motherboard sensor are 10C higher than what AOD / CoreTemp is reporting.
Saying that I don't see any difference in temps at given clock between 940BE (C-2) and 1090T (H-0).
Notice I was sane enough not to try 1.6V on AIR

[Forsaken1]

Pc probe appears to be in the ball park with temps.Offset coretemp to match.

Which water block?

I went back to swiftech GT awating AM3 bracket for swiftech XT.GT ran poor for me.

http://www.xtremesystems.org/forums/...&postcount=545

[chew*]

These things are heat monsters over 1.55v just some fyi. They are not deneb. On phase just booting and idling at 1.575 I lose 3c from 1.55v.

I personally would not reccomend much over 1.475 on air. and max 1.525 on water or degradation could occur.

[charged3800z24]

The temp is not acurate, look at bios reading. It is actual. I have been cruchig for a long time and @ 4ghz, the thing never goes into the 50. My water setup is not that great . Idle is well below my ambient room temp also. Even with air cooling will low speed fan.

[chew*]

The temp is not acurate, look at bios reading. It is actual. I have been cruchig for a long time and @ 4ghz, the thing never goes into the 50. My water setup is not that great . Idle is well below my ambient room temp also. Even with air cooling will low speed fan.

Bios readings are off to........My phase measures temps preety far away from evap....and says my max temps is 32C........bios reports 33c at die....

Not possible.

There is only one defacto standard to measure temps truly.

We notched an IHS and installed a probe ran under load and got a serious reality check.

[charged3800z24]

Bios readings are off to........My phase measures temps preety far away from evap....and says my max temps is 32C........bios reports 33c at die....

Not possible.

There is only one defacto standard to measure temps truly.

We notched an IHS and installed a probe ran under load and got a serious reality check.

Hmm, interesting. How did you notch the heat sink? I'd like to see where this thing is sitting temp wise.

[chew*]

Hmm, interesting. How did you notch the heat sink? I'd like to see where this thing is sitting temp wise.

I't sitting in michingan atm for further testing. We put in a mill and machined it, took the high edges off and milled for an accurate thermocouple.

We were taking temps delta readings on ln2.

[Zeus]

Hw monitor reports 16ºC idle for my cpu...with 21ºC ambient temp...

Titan, something must be wrong with your mounting i think, with your rad you shoud see even better temps than i get.

[Hey Zeus]

Getting 25C Idle's and around 45C-48C Prime's at this point at 1.52v. Pleasantly surprised at these temps.

[Dumo]

these things are heat monsters over 1.55v just some fyi. They are not deneb. On phase just booting and idling at 1.575 i lose 3c from 1.55v.

I personally would not reccomend much over 1.475 on air. And max 1.525 on water or degradation could occur.

qft

[Titan7171]

I want to say thanks to everyone who is posting info and insight into this thread

I have had a revelation, be it an idiotic one on my part and is almost offtopic to what this thread is really trying to be about.......apparently through all my testing and excitement, I seem to have overlooked one item in the bios that Ive pretty much always disabled.......spread spectrum! through all my tests this has been enabled in the bios and probably is one reason why Ive been needing more voltage than necessary. Now I dont know what spread spectrum does but Ive been disabling it since forever. So I disabled it this morning and have managed to knock off a full .1 off the voltage needed to run prime at 4.2ghz as shown in the below screenshot. Originally in the screenshot above using Asus pcprobe you can see 1.63v when using prime and you can see below in the new screenshot Im chugging along at 1.53v loaded 4.2ghz. In addition to the lower voltage comes lower temps. When I took the screen above it was much cooler in the room than it is when I took the new screen. So as of right now in the new screen Im at a higher ambiant with a lower cpu temp during a longer period of time. most of this is due to the .1 that has come off the cpu v

It should be noted that the original 1.63v is a product of the motherboard overvolting with 100% load, voltage in the bios was 1.59 this was set by stepping up .01 until I reached the point where prime would not bsod or reboot the system. starting from that point.......I went in down in voltage until prime would reboot or bsod then took it .01 up and retested so this is the result.

Still have to see about the mounting and other suggestions above from various people.



Bottom line......disable spread spectrum if you havent or forgot to dont be an idiot like me

[afireinside]

These things are heat monsters over 1.55v just some fyi. They are not deneb. On phase just booting and idling at 1.575 I lose 3c from 1.55v.

I personally would not reccomend much over 1.475 on air. and max 1.525 on water or degradation could occur.

So Thuban is more fragile than Deneb?

[Vapor]

These temp sensors are awful.

34.4C water temperatures and my CPU is loading at 35-35.6C at 3850MHz, 1.5v

[BeepBeep2]

These temp sensors are awful.

34.4C water temperatures and my CPU is loading at 35-35.6C at 3850MHz, 1.5v

Just act like you are only cooling a 15w heatload and say AMD has the most efficient CPU's.