4870 randomly throwing overheat protection.

[Mystik]

just recently my card's been randomly throwing itself into overheat mode.

i could be doing anything:

playing a game
browsing the net
photoshopping
watching a movie
idling with music
doing absolutely nothing.

this can happen with the fan config of:
automatic
40%
100%

GPU-Z records temps at:
gpu idle: 50-80 (depending on fan)
gpu load: 60-80 (depending on fan)

vrm idle: ~70c (average of all sensors)
vrm load: ~105c (average)

this happens with the software configs of:
less than one year old windows
fresh install of windows

==

anyone else have any ideas of what i should try or is my card just dying?

i can't reproduce it at all and it seems to happen at the most inopportune
moments.. like, for example, forgetting to save my photoshopping after
an hour.

[Blín D'ñero]

Check for dust clogged in your videocard's ventilator, on the CPU cooler and any fan in your PC's case.

[Qualitier]

Besides that, maybe the card's HSF is a bit loose. If so maybe you can tighten it a little with a screwdriver.

[Mystik]

blindnero::
unfortunately, i wish it were that simple.
i try to keep my case relatively clean so not much can build up in it.

qualitier::
i've checked that as well but everything seems fine with it - not loose at all.

unfortunately, there is almost zero chance for me to reproduce it at will so
getting the card rma'd is out of the question -- unless it just declines so far
enough that it just dies.

thanks for the replies guys. gonna keep trying stuff with it.

[Mystik]

i started logging with gpuz to see if something was up. after a couple of days
it finally threw the code again.

2009-01-01 20:39:31 ,              815.0   ,               46.0   ,          0   ,           26.4   ,                         62.0   ,

2009-01-01 20:39:32 ,              500.0   ,               46.0   ,          0   ,           25.7   ,                         62.0   ,

2009-01-01 20:39:33 ,              815.0   ,               46.0   ,          0   ,           26.4   ,                         62.0   ,

2009-01-01 20:39:34 ,              815.0   ,               46.0   ,          0   ,           26.4   ,                         62.0   ,

2009-01-01 20:39:48 ,              815.0   ,               46.0   ,         22   ,           26.4   ,                         62.0   ,

2009-01-01 20:39:48 ,              200.0   ,               -1.0   ,        200   ,            0.0   ,                          0.0   ,

2009-01-01 20:39:49 ,              200.0   ,               -1.0   ,        200   ,            0.0   ,                          0.0   ,

2009-01-01 20:39:50 ,              200.0   ,               -1.0   ,        200   ,            0.0   ,                          0.0   ,

2009-01-01 20:39:51 ,              200.0   ,               -1.0   ,        200   ,            0.0   ,                          0.0   ,

core speed, gpu usage, fanspeed, vddc current and vrm2 (highest of all 3
vrm sensors for me).

i was watching a movie at the time of the crash.

i'm going to wait for 09.1s to come out and see if that does anything (i only
started noticing this happening with 08.12s - but i'm finding it hard to believe
that drivers are causing the cards overheat sensor to engage). otherwise,
i might have to try and rma the card and hope i don't get a non-reference
4870 from asus (their d5 version).

bah.

[AGibbon]

For some reason it seems my temps have gone up about 5c with the 8.12s. I don't have any crashes, but I have to increase my fan speed by 20% to get close to the same temps I got with the 8.11s. Even then they never get down to the previous cats temps. Fluke? Not sure, but was wondering for a while myself. May try and revert back and take a glimpse.

[Mystik]

yeah, i'm gonna be reverting back to the 08.11s because my one year in-store
warranty ends early february.

hopefully it really is just a driver thing for either of us because i really don't
want to end up rma'ing the card and waiting 2-3 weeks for a replacement.

2-3 weeks without a computer.

how will i fap.. ?

[Mystik]

quick update for anyone who cares;

08.11 drivers seem to be fine.
since the only time i've ever seen the overheat protection come up
consistantly is when i'm watching i movie (without fail, although sometimes it
will come up in 30 minutes, sometimes a couple hours, but it will fail), i was
able to leave it running on movies/tv_eps since saturday night and it has not
failed yet.

i guess i wasn't going crazy when i was thinking it had to be the drivers but,
at the same time, i couldn't understand how the drivers could throw a
hardware protection switch off.