A computer components & hardware forum. HardwareBanter

If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below.

Go Back   Home » HardwareBanter forum » Video Cards » Nvidia Videocards
Site Map Home Register Authors List Search Today's Posts Mark Forums Read Web Partners

HOT GPU



 
 
Thread Tools Display Modes
  #1  
Old April 11th 07, 11:51 AM posted to alt.comp.periphs.videocards.nvidia
Charles C. Shyu
external usenet poster
 
Posts: 18
Default HOT GPU

A couple months ago I posted my 6800GT (inside a Dell XPS Gen 3, WinXP)
temps (according to NVIDIA Control Panel):
Ambient temperatu 56 deg C
GPU Co 81 deg C
Slowdown threshold: 135 deg C

Now I've found that if even a simple program like the 3D Pipes screensaver
comes on, the temps jump to:
Ambient temperatu 72 deg C
GPU Co 112 deg C
Slowdown threshold: 135 deg C

The numbers will quickly drop to "baseline" levels once the screensaver is
stopped, and will spike again if the screensaver is activated. In the
NVIDIA Control Panel, when the temp goes above 100 deg C or so it will be
shown in yellow instead of green, presumably indicating that this is not a
good temp to be running at.

I haven't noticed any symptoms with my system (locking up, reboots) at all,
but surely this can't be normal operating temp, right? Perhaps the fan on
the video card is malfunctioning? I did open up my case and de-dust the
thing, but the temps haven't budged. What gives?

Thanks!
--
Charles C. Shyu
http://home.earthlink.net/~shyuc/shyu.html


  #2  
Old April 11th 07, 03:28 PM posted to alt.comp.periphs.videocards.nvidia
Mr.E Solved!
external usenet poster
 
Posts: 888
Default HOT GPU

Charles C. Shyu wrote:
A couple months ago I posted my 6800GT (inside a Dell XPS Gen 3, WinXP)
temps (according to NVIDIA Control Panel):
Ambient temperatu 56 deg C
GPU Co 81 deg C
Slowdown threshold: 135 deg C

Now I've found that if even a simple program like the 3D Pipes screensaver
comes on, the temps jump to:
Ambient temperatu 72 deg C
GPU Co 112 deg C
Slowdown threshold: 135 deg C

The numbers will quickly drop to "baseline" levels once the screensaver is
stopped, and will spike again if the screensaver is activated. In the
NVIDIA Control Panel, when the temp goes above 100 deg C or so it will be
shown in yellow instead of green, presumably indicating that this is not a
good temp to be running at.

I haven't noticed any symptoms with my system (locking up, reboots) at all,
but surely this can't be normal operating temp, right? Perhaps the fan on
the video card is malfunctioning? I did open up my case and de-dust the
thing, but the temps haven't budged. What gives?

Thanks!



Run the same test with the PC case open, and see if the temperatures change.

That the temperature drops quickly is by design, due to the differing
speeds of the card's components depending on what the application is
doing. (Rendering or Rastering)

That you are reporting 112c under load on a 6800GT means something is
not right. While it might work fine today, eventually that temp (and the
ambient temp of 72c) will dry out capacitors and possibly damage traces
from expansion/contraction.
  #3  
Old April 11th 07, 05:35 PM posted to alt.comp.periphs.videocards.nvidia
ShutEye
external usenet poster
 
Posts: 95
Default HOT GPU

My 6800gt ran too hot.
I took off the plate covering the fan and removed the WALL of dust blocking
airflow.


  #4  
Old April 11th 07, 05:58 PM posted to alt.comp.periphs.videocards.nvidia
Phil Weldon
external usenet poster
 
Posts: 276
Default HOT GPU

'Charles. C. Shyu' wrote, in part:
| Now I've found that if even a simple program like the 3D Pipes screensaver
| comes on, the temps jump to:
| Ambient temperatu 72 deg C
| GPU Co 112 deg C
| Slowdown threshold: 135 deg C
_____

The ambient temperature says it all. It is too high. The cooling solution
result depends on the ambient temperature. An ambient temperature (I assume
you mean temperature within the system case) is 30 to 40 C higher than it
should be (unless you are located deep in a South African gold mine.) It is
therefore only to be expected that the GPU temperature should be 30 to 40 C
higher than it should be. For your under load example, 112 C - (30 to 40 C)
= 82 to 72 C. Still high, but within reason. As 'ShutEye' posted, house
cleaning is in order (and perhaps a check to see if the case fans are
actually turning.

Phil Weldon

"Charles C. Shyu" wrote in message
link.net...
|A couple months ago I posted my 6800GT (inside a Dell XPS Gen 3, WinXP)
| temps (according to NVIDIA Control Panel):
| Ambient temperatu 56 deg C
| GPU Co 81 deg C
| Slowdown threshold: 135 deg C
|
| Now I've found that if even a simple program like the 3D Pipes screensaver
| comes on, the temps jump to:
| Ambient temperatu 72 deg C
| GPU Co 112 deg C
| Slowdown threshold: 135 deg C
|
| The numbers will quickly drop to "baseline" levels once the screensaver is
| stopped, and will spike again if the screensaver is activated. In the
| NVIDIA Control Panel, when the temp goes above 100 deg C or so it will be
| shown in yellow instead of green, presumably indicating that this is not a
| good temp to be running at.
|
| I haven't noticed any symptoms with my system (locking up, reboots) at
all,
| but surely this can't be normal operating temp, right? Perhaps the fan on
| the video card is malfunctioning? I did open up my case and de-dust the
| thing, but the temps haven't budged. What gives?
|
| Thanks!
| --
| Charles C. Shyu
| http://home.earthlink.net/~shyuc/shyu.html
|
|


  #5  
Old April 14th 07, 02:51 PM posted to alt.comp.periphs.videocards.nvidia
Charles C. Shyu
external usenet poster
 
Posts: 18
Default HOT GPU

"Mr.E Solved!" wrote in message
. ..
Charles C. Shyu wrote:
A couple months ago I posted my 6800GT (inside a Dell XPS Gen 3, WinXP)
temps (according to NVIDIA Control Panel):
Ambient temperatu 56 deg C
GPU Co 81 deg C
Slowdown threshold: 135 deg C

Now I've found that if even a simple program like the 3D Pipes
screensaver comes on, the temps jump to:
Ambient temperatu 72 deg C
GPU Co 112 deg C
Slowdown threshold: 135 deg C

The numbers will quickly drop to "baseline" levels once the screensaver
is stopped, and will spike again if the screensaver is activated. In the
NVIDIA Control Panel, when the temp goes above 100 deg C or so it will be
shown in yellow instead of green, presumably indicating that this is not
a good temp to be running at.

I haven't noticed any symptoms with my system (locking up, reboots) at
all, but surely this can't be normal operating temp, right? Perhaps the
fan on the video card is malfunctioning? I did open up my case and
de-dust the thing, but the temps haven't budged. What gives?

Thanks!



Run the same test with the PC case open, and see if the temperatures
change.

That the temperature drops quickly is by design, due to the differing
speeds of the card's components depending on what the application is
doing. (Rendering or Rastering)

That you are reporting 112c under load on a 6800GT means something is not
right. While it might work fine today, eventually that temp (and the
ambient temp of 72c) will dry out capacitors and possibly damage traces
from expansion/contraction.


Thanks for the suggestion. I ran the computer with the case fully opened,
and first found that the "idle" temps seem to be the same:
Ambient temperatu 56 deg C
GPU Co 78 deg C

Mind you, my apartment temperature is 73 deg F = 23 deg C.

Upon loading the 3D Pipes screensaver, the temp went to:
Ambient temperatu 66 deg C
GPU Co 94 deg C

So the "idle" temp doesn't seem to change much with the case open, but the
"load" temp does seem to significantly drop.

Firstly, I'm wondering if the temperature shown by NVIDIA Control Panel is
even correct. How the heck can the "Ambient temperature" be 57 deg C when
my room is only 23 deg C? I put my hand in the case and it only feels
slightly warmer than outside the case.

Secondly, is there another way to verify these temperature numbers? If the
reading is incorrect to begin with, I may be trying to diagnose something
that's not even there.
--
Charles C. Shyu
http://home.earthlink.net/~shyuc/shyu.html


  #6  
Old April 14th 07, 02:53 PM posted to alt.comp.periphs.videocards.nvidia
Charles C. Shyu
external usenet poster
 
Posts: 18
Default HOT GPU

"Phil Weldon" wrote in message
link.net...
'Charles. C. Shyu' wrote, in part:
| Now I've found that if even a simple program like the 3D Pipes
screensaver
| comes on, the temps jump to:
| Ambient temperatu 72 deg C
| GPU Co 112 deg C
| Slowdown threshold: 135 deg C
_____

The ambient temperature says it all. It is too high. The cooling
solution
result depends on the ambient temperature. An ambient temperature (I
assume
you mean temperature within the system case) is 30 to 40 C higher than it
should be (unless you are located deep in a South African gold mine.) It
is
therefore only to be expected that the GPU temperature should be 30 to 40
C
higher than it should be. For your under load example, 112 C - (30 to 40
C)
= 82 to 72 C. Still high, but within reason. As 'ShutEye' posted, house
cleaning is in order (and perhaps a check to see if the case fans are
actually turning.

Phil Weldon


Phil, thanks for the input. I've run my system with the case open and
posted the temp results earlier in this thread. I've cleaned up the dust as
much as I've can, and yes, the case fans as well as the video card fans seem
to be working. I'm now wondering if the temperature reading by NVIDIA
Control Panel is (way) off for some reason...
--
Charles C. Shyu
http://home.earthlink.net/~shyuc/shyu.html


  #7  
Old April 15th 07, 03:05 PM posted to alt.comp.periphs.videocards.nvidia
RF
external usenet poster
 
Posts: 75
Default HOT GPU


"Charles C. Shyu" wrote in message
.net...
Phil, thanks for the input. I've run my system with the case open and
posted the temp results earlier in this thread. I've cleaned up the dust
as much as I've can, and yes, the case fans as well as the video card fans
seem to be working. I'm now wondering if the temperature reading by
NVIDIA Control Panel is (way) off for some reason...
--
Charles C. Shyu
http://home.earthlink.net/~shyuc/shyu.html


Just running your system with the case open might actually raise
temperatures because there's no fluid airflow through the case. You'll get
pockets of air sitting in places. If you run with the case off, blow a
small house fan into the open side of the case to ensure that air is flowing
everywhere properly.

RF.


  #8  
Old April 15th 07, 05:47 PM posted to alt.comp.periphs.videocards.nvidia
Mr.E Solved!
external usenet poster
 
Posts: 888
Default HOT GPU

Charles C. Shyu wrote:

A couple months ago I posted my 6800GT (inside a Dell XPS Gen 3, WinXP)
temps (according to NVIDIA Control Panel):
Ambient temperatu 56 deg C
GPU Co 81 deg C
Slowdown threshold: 135 deg C

Now I've found that if even a simple program like the 3D Pipes
screensaver comes on, the temps jump to:
Ambient temperatu 72 deg C
GPU Co 112 deg C
Slowdown threshold: 135 deg C

from expansion/contraction.

Thanks for the suggestion. I ran the computer with the case fully opened,
and first found that the "idle" temps seem to be the same:
Ambient temperatu 56 deg C
GPU Co 78 deg C

Mind you, my apartment temperature is 73 deg F = 23 deg C.

Upon loading the 3D Pipes screensaver, the temp went to:
Ambient temperatu 66 deg C
GPU Co 94 deg C

So the "idle" temp doesn't seem to change much with the case open, but the
"load" temp does seem to significantly drop.



Firstly, I'm wondering if the temperature shown by NVIDIA Control Panel is
even correct. How the heck can the "Ambient temperature" be 57 deg C when
my room is only 23 deg C? I put my hand in the case and it only feels
slightly warmer than outside the case.


The "ambient temp" readout is the temperature in the immediate vicinity
of the motherboard mounted temperature sensor, which will be
significantly higher than the temperature ambient to your person,
assuming you aren't sitting inside your case.

Also called "System Temperature" it controlled by the Winbond or other
brand specific sensor chip which is an integral part of your system board.


Secondly, is there another way to verify these temperature numbers? If the
reading is incorrect to begin with, I may be trying to diagnose something
that's not even there.


You can use any number of third party temperature sensor monitors: MBM,
Speedfan (which never worked for me), Rivatuner(?), Powerstrip(?). Often
times motherboard compatibility is specific to the monitoring
application (and version) used. Of course, ntune and nvmonitor work with
nvidia chipsets. It is likely you can run multiple monitoring programs
side by side, to compare results.

Be aware, the temperatures reported are not actual "temperatures" in the
sense that you might think. The sensor's circuits reports a number from
1-255 depending on the resistance in the sensor, and that result is
translated to a familiar temperature scale. They are accurate only to
the resolution of 3 degrees Fahrenheit, by design. Some do have greater
resolution, down to 0.1 degree Fahrenheit, by virtue of reporting a
value of 1 to 4K or some other larger number.

While absolute temperatures are relevant, since resolution is poor, you
must not rely on snapshots, you must inspect temperature logs to see
what the peaks are, when they are, and how the system responds to
increased thermal load by checking the slope of the temperature over
time over load, that way you more accurate determine if your problem is
rooted in:

1) convection
2) conduction
3) radiation


Radiation of heat from the components will heat up other components,
making even more heat, and convection of air currents is the primary way
to move that heat away from the hot spots and out the case. I know this
sounds basic, but you have to make sure that not only do you have case
fans working in tandem, but that the resultant air current is actually
occurring where you need it most: back of you video card, over your
mosfets, over and through your CPU HSF, over and through your PSU.

You could have the best case fans made, but if they are only moving a
portion of the air in side the case, and not that portion you need, when
additional thermal load is added, the GPU temp and System temp will
spike dramatically, but stay normal at idle.

While GPU outboard thermal exhaust fans are excellent (by local
conduction and convection built into the unit, not adding to the case
thermal load), they only solve half the problem, since the back of the
video card does not benefit from the gpu exhaust fan's convection,
regardless if there are heat pies and other fancy ram coolers on the
back of a card.

In most PC cases, that void between the PCI-E slot and the CPU HSF is a
hot spot with little convection and much radiation, that the power
moving mosfets are generally close to that location, adds to the thermal
load. Pay particular attention to that area on your mobo, especially if
your PSU does not have a down-side opening with a fan to assist in
convecting that air up and out the back of the PSU.

Long and short of it is more fans does not necessarily mean lower temps,
but more convection in your specific trouble spots, common to most pcs,
will lower your load temps.

btw, the sensor on your video card is likely a thermistor mounted
directly under the chip itself, though I seem to recall some sort of
actual in die sensor mentioned somewhere, but not on the GeForce 6
series? Someone might know the facts on which sensor design is in which
models.










  #9  
Old April 16th 07, 04:12 AM posted to alt.comp.periphs.videocards.nvidia
Charles C. Shyu
external usenet poster
 
Posts: 18
Default HOT GPU

Long and short of it is more fans does not necessarily mean lower temps,
but more convection in your specific trouble spots, common to most pcs,
will lower your load temps.

btw, the sensor on your video card is likely a thermistor mounted directly
under the chip itself, though I seem to recall some sort of actual in die
sensor mentioned somewhere, but not on the GeForce 6 series? Someone might
know the facts on which sensor design is in which models.


Lots of good info, but I'm still not convinced that I'm actually dealing
with a problem here. I downloaded nTune to see if it gave a different GPU
temp reading, but it gives the same as NVIDIA Control Panel, which I guess
is to be expected? So I'm still left wondering if the temp reading is
correct or not...
--
Charles C. Shyu
http://home.earthlink.net/~shyuc/shyu.html


 




Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump


All times are GMT +1. The time now is 02:28 PM.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 HardwareBanter.
The comments are property of their posters.