Overclock.net banner

fx-8120 + sabertooth 990fx

12K views 56 replies 19 participants last post by  mattlb0619 
#1 ·
idle
cpu - 46c
vcore1 - 50c
vcore2 - 49c
vcore 1.176v~

load (prime95 - large ffts, 10 minutes)
cpu - 60c
vcore1 - 65c
vcore2 - 58c
vcore - 1.140v

using the coolermaster hyper 212 evo with cougar vortex pwm fans. oc in bios set to manual awith everthing on auto but the memory, set to 1866, the ram default. what is the deal here?

edit: using MASSCOOL G751 Shin-Etsu Thermal Interface Material.... i applied four small(1/4 pea size) dots of paste at the four points of a square onto the cpu. i apllied i once and removed the hs and checked it, looked good. applied it again and put it together havent checked it since then, out of paste. seems to me that the vcore is hot too though so ... ?
 
#2 ·
voltage is too high, 1.55 is the highest i have heard of on water cooling.
i have a zalman performa air cooler, and its idle is 40c ,but the core temps are saying much less,at 3.8ghz iam hiting the thermal barrier of 61c.
if you go into amd overdrive look at board status and tmpin1,thats the highest temp and the ione to watch in my book.
 
#4 ·
I have to lean towards a improperly seated cooler. remove, clean, reapply tim, and re-install the cooler.

I have seemed to over look the last part of your post due to lack of sleep apply the tim in a pea size shape i the middle of the cpu chip. As far as i know this is the most efficient way of doing it also do you have the flat bottom 212 or the one with the gaps along the pipe flat bottom follow above the pipe gap way fill the pipe gaps a little bit before following the pea method. forgive any errors in my grammar .
 
#6 ·
@cannon - i didnt apply more.. i applied the norm but so that it would cover a larger area, thinner.

@irish - it is direct pipe but i wouldnt say that it isnt flat.
 
#7 ·
sorry for the dbl post but i have another point/?...

if it is just bad mounting or tim application, then why are the 2 vcores(as seen in the asus software) so high as well?

keep it coming guys.... love the assist.
 
#8 ·
Quote:
Originally Posted by mattlb0619 View Post

sorry for the dbl post but i have another point/?...
if it is just bad mounting or tim application, then why are the 2 vcores(as seen in the asus software) so high as well?
keep it coming guys.... love the assist.
if you are using the asus thermal monitoring software, vcore1 and vcore2 are you vrm/mos fets on the motherboard, nothing to do with the cpu. check out the thermal radar.

that being said, your cpu temps aren't horrible considering the cooler. while the 212 has a decent reputation for being a bargain, it is not a high end cooler, and the 8120 will be a handful for it if you plan on overclocking.

my preference on a heatpipe cooler is to apply the TIM and spread it very thin, making sure to get coverage. it has given me the best results with both the 212 and Xigs.
 
#9 ·
Look at all your temps thru the ASUS Thermal Radar and see if the high temps are located around the CPU. I have a 8150@4.4 running 24/7 with full load and getting the following temps:

CPU - 51c
Vcore1 - 52c
Vcore2 - 51c
MB - 37c
 
#10 ·
Quote:
Originally Posted by Hawtoothpk View Post

Look at all your temps thru the ASUS Thermal Radar and see if the high temps are located around the CPU. I have a 8150@4.4 running 24/7 with full load and getting the following temps:
CPU - 51c
Vcore1 - 52c
Vcore2 - 51c
MB - 37c
Those temps are not bad, but I do question why you have 100F MB temp? That should be much closer to ambient's since the sensor is at the bottom right corner of the board, airflow issue? My MB temps are 23-25C with a 5.0Ghz OC at 1.56V. As for the V-core they are good up to 85C before it will self throttle.
 
#12 ·
ok, 1, you are using a rather low end air cooler on a chip that is known to be very "hot headed". 2nd, you are applying the paste wrong. on the 212+ you need to apply thin lines of the TIM along the cracks of the 212+ HS. that is the best way to apply TIM on a 212+. on other, non direct pipe heatsinks, the single dot in the center of tim is the best way. but if you have ever applied a dot on the 212+ then removed it you will see where that is flawed. Do what i suggested and your temps will drop somewhat but not much.
 
#13 ·
I am also running two video cards at full load doing 2 BOINC GPU jobs and producing higher temps for the motherboard.

PCIE1 - 56c
PCIE2 - 52c

GPU1 - 79c
GPU2 - 82c

To support my OC, I used the DIGI+VRM to up my CPU Load Line Calibration to Ultra High and the CPU/NB Load Line Calibration to High
 
#14 ·
sorry ,miss read your post,thinking you put 1.7 volts,which is nuts.
 
#15 ·
Quote:
Originally Posted by Papas View Post

ok, 1, you are using a rather low end air cooler on a chip that is known to be very "hot headed". 2nd, you are applying the paste wrong. on the 212+ you need to apply thin lines of the TIM along the cracks of the 212+ HS. that is the best way to apply TIM on a 212+.
Hes using the 212+ EVO Difference being that the evo has an even copper base (no cracks to fill). At least thats what Ive gathered from all the evo pictures out there. And I wouldn't consider that a low end cooler by any means. The only low thing about it, is the price!

500

Its been my experience that the shin-etsu paste has horrible spread-ability. I would take one of the members advice from an earlier post and manually spread it on thin.
 
#16 ·
im pretty much at idle now... temp straight from thermal radar.

379 379

653

papas - 2advanced is correct there are no gaps, its flat.

2advanced - as i stated before i tested how it was gonna spread before i started

mrpreforations - at 1.7v this thing would blow up as it is now, lol

Hawtoothpk - enabling those features counter acts the voltage drop thats happening but its also raising the temps even higher. in the asus turbov software voltage is at 1.21250. setting LLC to high raises the vcore to this.
 
#18 ·
UPDATE:

upgraded to a Fractal Design Arc Midi. Only using the included 3x 140 fans right now but 5x cougar 140s are on the way. I have also remounted the cooler, yes the tim was lacking, not enough over the area. mx-2 now, plenty applied.

front
450

back
450

idle @ 4ghz
338

load @ 4ghz
339

ambient temp is about 76f

to me the temps are still too high exp with this case and tim at this voltage.... thoughts?

it is stable at this voltage though at 4ghz... have done all prime t tests over 1hr and 3dmark and pcmark without fail.

im obviously not water cooling but i didnt want a case without the option, i will be upgrading. question is, is the cpu defective or am i missing something? cause imo its still too hot!?!
 
#19 ·
its just dam hot matt.
i cant get above 4ghz due to heat.

i just tried to test mine at 3800 and hit 61 in 3 minutes and just proved how good my heat sink is ,in about 30 seconds the temp droped by 10c.
if yours is the same its not the paste or sink.

just testing out your lower voltage idea,and it seems to be working,thanks
thumb.gif


i got my 3800 @ 1.25volts @ 58c and prime small fft for 1 hour,seems stable.
 
#20 ·
5x cougars fans in....

2 fractal in front pwm 12v
1 cougar in bottom 7v
1 cougar in side pwm 12v (this one and back are only two unfiltered, been hard finding a good 140mm filter, suggestions?)
2 cougar out top 7v
1 cougar out back pwm 7v

idle
337

load
338

ambient temps are the same. i am guessing this is normal now or i have a defective cpu....?

ps
for the record ide recommend these fans to anyone... exp at $12/each. ive never owned a noctua but idk how they could be that much better at this price.
 
#21 ·
That's a bit weird i'm using the stock cooler @ 4014Mhz / 1.125 and i hardly reach 45 + when i'm using boinc over night or 58 ~ when i ran Prime95.
I have the corsair R500 case series maybe it's making that much of a difference? (I'm using Noctuas Thermal Grease)
My room is like 19-20 / (Warmer days) 24 C
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top