Sign in with
Sign up | Sign in
Your question

Prolink FX5900 128MB heat issues

Last response: in Graphics & Displays
Share
Anonymous
April 2, 2004 4:16:10 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

Prolink FX5900 128MB card MVGA-NVG35BA runs at 80C while in 2D mode
and the GPU is set at 300Mhz. In 3D mode at default speeds it gets up
( and likely beyond) 111C. The ambient temp reading is at 36C. I no
longer get Sentinel warning messages after I updated the system to
nvidia's latest drivers, and adding a PCI dual fan card to blow
directly on the FX5900 card, but should it be this hot? I have two
exhaust fans, a Volcano11+ on the CPU, and one intake not counting the
PCI dual fan card.

Can you speed up the fan on the card???? What can I do?

Asus A7N8X Deluxe v 2.0
Windows XP Pro SP1
AMD 2500+ Barton ( 11@ 333Mhz)
Antec TruePower 430W
Coolermaster ATC 210 case
512MB Kingston HyperX DDR (dual channel)
Anonymous
April 2, 2004 8:10:10 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

In article <5e388633.0404021216.39308eba@posting.google.com>, greg_
2day@hotmail.com says...
> Prolink FX5900 128MB card MVGA-NVG35BA runs at 80C while in 2D mode
> and the GPU is set at 300Mhz. In 3D mode at default speeds it gets up
> ( and likely beyond) 111C. The ambient temp reading is at 36C. I no
> longer get Sentinel warning messages after I updated the system to
> nvidia's latest drivers, and adding a PCI dual fan card to blow
> directly on the FX5900 card, but should it be this hot? I have two
> exhaust fans, a Volcano11+ on the CPU, and one intake not counting the
> PCI dual fan card.
>
> Can you speed up the fan on the card???? What can I do?
>
Your temps are way too high! Make sure the fan on the card is actually
connected to the GPU. Also, be sure the fan is turning.

My Leadtek FX5900XT (same GPU, slightly slower memory) stays at 40C or
less in 2D mode and reaches about 55C during a UT2004 session. Similar
ambient temperature.
Anonymous
April 3, 2004 3:15:28 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

My Prolink FX5900 128mb runs at 49c with the 53.03 drivers..!

"wgd" <wgd@nomail.com> wrote in message
news:MPG.1ad7a0ba4acf12f198969f@news.pa.comcast.giganews.com...
> In article <5e388633.0404021216.39308eba@posting.google.com>, greg_
> 2day@hotmail.com says...
> > Prolink FX5900 128MB card MVGA-NVG35BA runs at 80C while in 2D mode
> > and the GPU is set at 300Mhz. In 3D mode at default speeds it gets up
> > ( and likely beyond) 111C. The ambient temp reading is at 36C. I no
> > longer get Sentinel warning messages after I updated the system to
> > nvidia's latest drivers, and adding a PCI dual fan card to blow
> > directly on the FX5900 card, but should it be this hot? I have two
> > exhaust fans, a Volcano11+ on the CPU, and one intake not counting the
> > PCI dual fan card.
> >
> > Can you speed up the fan on the card???? What can I do?
> >
> Your temps are way too high! Make sure the fan on the card is actually
> connected to the GPU. Also, be sure the fan is turning.
>
> My Leadtek FX5900XT (same GPU, slightly slower memory) stays at 40C or
> less in 2D mode and reaches about 55C during a UT2004 session. Similar
> ambient temperature.
Related resources
April 4, 2004 9:21:34 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

First, download a Leadtek or other company's bios for your card.


You can try the 5950 mod (just get a 5950 bios and update it with
that). The 5950's increase the voltage on the graphics processor and
the fan, so you can overclock at a higher stability.

Increasing the fan voltage lowers the life of the fan slightly, but
also makes it spin at a higher RPM, thus cooling at better performance.




If you don't want to experiment, go with the 5900 bios. (Make sure you
check the bios numbers, we don't want to get an older revision)



If this doesn't solve your issue, remove the card.


Then get some tweezers or finepoint needlenose pliers and pinch the
plastic standoffs holding the heatsink of the card on.

They should come out after working with them for a while, mine were
stuck on there pretty snug.


After you've removed the standoffs, remove the heatsink, but NOT BY
PULLING UPWARDS! If you pull upwards away from the card, you risk
damaging the GPU or the fan cord you probably forgot to unplug from the
card.

So, first remove the fan cord and tuck it in the heatsink.

Twist the heatsink slightly, until it starts to loosen. From there you
should be able to pull the heatsink away from the GPU at a 45? angle
(opposed to a 90? angle, directly up from the GPU).

If you have any trouble, you can put a little rubbing alcohol between
the gpu and heatsink, it'll loosen the bond.


After you get it off clean up the heatsink and GPU with rubbing
alcohol. You should remove every last bit of thermal grease.


The thermal grease manufacturers and electricians use is terrible for
processors, memory, etc. So go out and buy some XTC or ArcticSilver,
just make sure you don't get Arctic Alumina, as that is a ceramic
adhesive... you won't be able to get the heatsink off ever again.

If you notice that your heatsink is not smooth(as I noticed on mine),
you could opt to get a package of 80-200 grit sandpaper and stand it to
a nice smooth shine.

This will improve the conductivity between the heatsink and the GPU.


This isn't necessary if you apply the Arctic Silver correctly though.


Stick a pea sized drop on the GPU, using a credit card to sort of
plaster it on smoothly and thinly. Stick a slightly larger drop on the
center of the heatsink, making sure to fill in any abraisions or
crevices that are in approximation to where the heatsink will meet the
GPU.

Any touchups should be done with a foam brush, soft sponge, or similar
non-filamental item, but never use your fingers unless they're behind
gloves or plastic bags.


If you notice any residue from the Arctic Silver (ie you got some on
your finger and touched the card) clean it immediately!

Arctic Silver uses multiple types of silver particles, which conduct
electricity.


Reapply the heatsink, making sure not to move it around much.


I'd recommend not turning on your computer for about an hour, so that
the thermal grease can cure.
It would be more beneficial if you left the card out overnight so the
curing process can work successfully.
If you run your system right after applying the thermal grease, the
heat causes some instability with the grease and it won't conduct heat
well.




This should get your temps down to a stable 35? at idle.






Apoo wrote:

> My Prolink FX5900 128mb runs at 49c with the 53.03 drivers..!
>
> "wgd" <wgd@nomail.com> wrote in message
> news:MPG.1ad7a0ba4acf12f198969f@news.pa.comcast.giganews.com...
> > In article <5e388633.0404021216.39308eba@posting.google.com>, greg_
> > 2day@hotmail.com says...
> > > Prolink FX5900 128MB card MVGA-NVG35BA runs at 80C while in 2D
> > > mode and the GPU is set at 300Mhz. In 3D mode at default speeds
> > > it gets up ( and likely beyond) 111C. The ambient temp reading
> > > is at 36C. I no longer get Sentinel warning messages after I
> > > updated the system to nvidia's latest drivers, and adding a PCI
> > > dual fan card to blow directly on the FX5900 card, but should it
> > > be this hot? I have two exhaust fans, a Volcano11+ on the CPU,
> > > and one intake not counting the PCI dual fan card.
> > >
> > > Can you speed up the fan on the card???? What can I do?
> > >
> > Your temps are way too high! Make sure the fan on the card is
> > actually connected to the GPU. Also, be sure the fan is turning.
> >
> > My Leadtek FX5900XT (same GPU, slightly slower memory) stays at 40C
> > or less in 2D mode and reaches about 55C during a UT2004 session.
> > Similar ambient temperature.
April 5, 2004 5:47:42 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

greg_2day@hotmail.com (G-imme) wrote:
> Prolink FX5900 128MB card MVGA-NVG35BA runs at 80C while in 2D mode
> and the GPU is set at 300Mhz. In 3D mode at default speeds it gets up
> ( and likely beyond) 111C. The ambient temp reading is at 36C. I no
> longer get Sentinel warning messages after I updated the system to
> nvidia's latest drivers, and adding a PCI dual fan card to blow
> directly on the FX5900 card, but should it be this hot? I have two
> exhaust fans, a Volcano11+ on the CPU, and one intake not counting the
> PCI dual fan card.

80 celsius is quite hot, but might be normal; however, since it's
during Windows environment -- something might be wrong. Check this
article's temperature measurements out:

http://www.anandtech.com/video/showdoc.html?i=1955&p=18

> Can you speed up the fan on the card???? What can I do?
>
> Asus A7N8X Deluxe v 2.0
> Windows XP Pro SP1
> AMD 2500+ Barton ( 11@ 333Mhz)
> Antec TruePower 430W
> Coolermaster ATC 210 case
> 512MB Kingston HyperX DDR (dual channel)

Pray and stop eating for 40 days. It's Easter!
Anonymous
April 5, 2004 11:29:36 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

Well, the Vantec PCI fan doesn't seem to be helping very much.
Although when I touch the heatsink after about 60 secs after pwoering
off everything. It is NOT very warm at all. Ambient temps of 35C
should be sufficient. (Asus Probe also has the case interior's temp
around 35C. Over the weekend, I took off the heatsink and fan to try
Artic Silver Ceramique. I am doubtful that it will drop the temps
greatly. This is considering that I have temps at 78- 80C when in 2D
mode with the GPU set @ 300Mhz. May be Prolink should have used
seperate heat sinks for the ram.

I would try the flahsing the bios, but I can't obtain a Prolink BIOS.
April 6, 2004 5:21:14 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

greg_2day@hotmail.com (G-imme) wrote:
> Well, the Vantec PCI fan doesn't seem to be helping very much.
> Although when I touch the heatsink after about 60 secs after pwoering
> off everything. It is NOT very warm at all. Ambient temps of 35C
> should be sufficient. (Asus Probe also has the case interior's temp
> around 35C. Over the weekend, I took off the heatsink and fan to try
> Artic Silver Ceramique. I am doubtful that it will drop the temps
> greatly. This is considering that I have temps at 78- 80C when in 2D
> mode with the GPU set @ 300Mhz. May be Prolink should have used
> seperate heat sinks for the ram.

The Forceware drivers' temperature measurement usually shows an okay
52-55 Celsius after heavy benchmarks and games on my Leadtek 5900XT.
Do you have Prolink's own temperature software installed or do you get
the temperature with the Forceware drivers? It's bizarre if it truly
is that hot, but if your card wasn't _that_ hot when touching the fan,
I believe it's very likely the temperature readings are not correct --
and it's lower than what it says. If your games and benchmarks work
fine without lockups, then everything should be okay.

> I would try the flahsing the bios, but I can't obtain a Prolink BIOS.

Flashing BIOS is always risky, so I wouldn't recommend doing that
unless your warranty covers it.
April 6, 2004 10:18:12 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

G-imme wrote:

> Well, the Vantec PCI fan doesn't seem to be helping very much.
> Although when I touch the heatsink after about 60 secs after pwoering
> off everything. It is NOT very warm at all. Ambient temps of 35C
> should be sufficient. (Asus Probe also has the case interior's temp
> around 35C. Over the weekend, I took off the heatsink and fan to try
> Artic Silver Ceramique. I am doubtful that it will drop the temps
> greatly. This is considering that I have temps at 78- 80C when in 2D
> mode with the GPU set @ 300Mhz. May be Prolink should have used
> seperate heat sinks for the ram.
>
> I would try the flahsing the bios, but I can't obtain a Prolink BIOS.




I told you not to use the Ceramic! This stuff is PERMANENT, you can't
remove the heatsink after it cures....



Use a Leadtek bios, it will work.
Anonymous
April 6, 2004 10:18:13 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

"teqguy" <teqguy@techie.com> wrote in message news:<EYrcc.2274$TS3.1392@nwrddc02.gnilink.net>...
> G-imme wrote:
>
> > Well, the Vantec PCI fan doesn't seem to be helping very much.
> > Although when I touch the heatsink after about 60 secs after pwoering
> > off everything. It is NOT very warm at all. Ambient temps of 35C
> > should be sufficient. (Asus Probe also has the case interior's temp
> > around 35C. Over the weekend, I took off the heatsink and fan to try
> > Artic Silver Ceramique. I am doubtful that it will drop the temps
> > greatly. This is considering that I have temps at 78- 80C when in 2D
> > mode with the GPU set @ 300Mhz. May be Prolink should have used
> > seperate heat sinks for the ram.
> >
> > I would try the flahsing the bios, but I can't obtain a Prolink BIOS.
>
>
>
>
> I told you not to use the Ceramic! This stuff is PERMANENT, you can't
> remove the heatsink after it cures....
>
>
>
> Use a Leadtek bios, it will work.

Ceramic is NOT permanent..it can be removed ... but yes..its capable
to "glue" together the heatsink and the gpu.

Artic Silver 5 though is the best option for GPU's and CPU's!
Anonymous
April 6, 2004 10:39:00 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

> I told you not to use the Ceramic! This stuff is PERMANENT, you can't
> remove the heatsink after it cures....
>
>
>
> Use a Leadtek bios, it will work.

Hahaa. Thanks, but I never have any problem removing the fan/sinks
from Athlon cpus. I think I'll be okay as long as I remove it when it
is warm.

My temps have NOT dropped from baseline. Still at 75 - 80C in 2D Mode
with GPU set @ 300Mhz. I am able to play most of my games without any
hiccups or problems. Just not sure if I SHOULD be playing any games
for extended periods of time. I may go back to earlier drivers for
the chipset.

How much do you think the Leadtek bios will help???
Anonymous
April 6, 2004 6:31:02 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

> The Forceware drivers' temperature measurement usually shows an okay
> 52-55 Celsius after heavy benchmarks and games on my Leadtek 5900XT.
> Do you have Prolink's own temperature software installed or do you get
> the temperature with the Forceware drivers? It's bizarre if it truly
> is that hot, but if your card wasn't _that_ hot when touching the fan,
> I believe it's very likely the temperature readings are not correct --
> and it's lower than what it says. If your games and benchmarks work
> fine without lockups, then everything should be okay.
>
> > I would try the flahsing the bios, but I can't obtain a Prolink BIOS.
>
> Flashing BIOS is always risky, so I wouldn't recommend doing that
> unless your warranty covers it.

Prolink uses nvidia's older driver set, not their own.
April 6, 2004 6:50:02 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

G-imme wrote:

> > I told you not to use the Ceramic! This stuff is PERMANENT, you
> > can't remove the heatsink after it cures....
> >
> >
> >
> > Use a Leadtek bios, it will work.
>
> Hahaa. Thanks, but I never have any problem removing the fan/sinks
> from Athlon cpus. I think I'll be okay as long as I remove it when it
> is warm.
>
> My temps have NOT dropped from baseline. Still at 75 - 80C in 2D Mode
> with GPU set @ 300Mhz. I am able to play most of my games without any
> hiccups or problems. Just not sure if I SHOULD be playing any games
> for extended periods of time. I may go back to earlier drivers for
> the chipset.
>
> How much do you think the Leadtek bios will help???





Dark, my experiences with Arctic Alumina have proved to be a major risk
when trying to remove it.


I had to install a heatsink on a Northbridge that didn't have any
retention clips or pinholes... a couple days later, the thing fell off.

All that was left was a huge crater in the middle of my chipset.



The ceramic alone should be fine, but that adhesive is a bitch.




Those products don't yield equal performance of the Arctic Silver, so
that's probably why you didn't notice a temp adjustment G-imme.


Ceramic is an insulator, the silver flakes and grease in Arctic Silver
act as a transient.


The heat given off is residing in the ceramic too long, and that's your
problem.





I recommend doing the 5900 >> 5950 mod G-imme.


It'll cool it down because it raises the voltage on the fan.
Anonymous
April 6, 2004 8:43:09 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

On 5 Apr 2004 07:29:36 -0700, greg_2day@hotmail.com (G-imme) wrote:

>Well, the Vantec PCI fan doesn't seem to be helping very much.
>Although when I touch the heatsink after about 60 secs after pwoering
>off everything. It is NOT very warm at all. Ambient temps of 35C
>should be sufficient. (Asus Probe also has the case interior's temp
>around 35C. Over the weekend, I took off the heatsink and fan to try
>Artic Silver Ceramique. I am doubtful that it will drop the temps
>greatly. This is considering that I have temps at 78- 80C when in 2D
>mode with the GPU set @ 300Mhz. May be Prolink should have used
>seperate heat sinks for the ram.
>
>I would try the flahsing the bios, but I can't obtain a Prolink BIOS.


No sign of any systematic troubleshooting here... Bang, bang....

Did you use Driver Cleaner to clear out all your video driver junk and

install the nVidia reference driver 53.03 or above ?

You may have faulty thermal sensing due to a corrupted driver.
It is vitally important to ensure that your driver and system software

are above reproach before messing with hardware, particularly
if the board otherwise appears to be working normally and the
heatsink is not glowing cherry-red....

There is a very, very remote possibility that the thermal sensing
internal to the GPU or its board interface electronics is faulty.

Anyway, I assume that the board is pretty new. Did you contemplate
a RMA before the major surgery that you have performed ?

Apparent gross overheating in 2D or quiescent-3D mode, without
any obvious physical overheating at the heat-sink is sufficient
grounds for a RMA from any reasonable board-vendor.

John Lewis
April 7, 2004 5:00:14 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

john.dsl@verizon.net (John Lewis) wrote:
> On 5 Apr 2004 07:29:36 -0700, greg_2day@hotmail.com (G-imme) wrote:
>
> >Well, the Vantec PCI fan doesn't seem to be helping very much.
> >Although when I touch the heatsink after about 60 secs after pwoering
> >off everything. It is NOT very warm at all. Ambient temps of 35C
> >should be sufficient. (Asus Probe also has the case interior's temp
> >around 35C. Over the weekend, I took off the heatsink and fan to try
> >Artic Silver Ceramique. I am doubtful that it will drop the temps
> >greatly. This is considering that I have temps at 78- 80C when in 2D
> >mode with the GPU set @ 300Mhz. May be Prolink should have used
> >seperate heat sinks for the ram.
> >
> >I would try the flahsing the bios, but I can't obtain a Prolink BIOS.
>
>
> No sign of any systematic troubleshooting here... Bang, bang....
>
> Did you use Driver Cleaner to clear out all your video driver junk and
>
> install the nVidia reference driver 53.03 or above ?
>
> You may have faulty thermal sensing due to a corrupted driver.
> It is vitally important to ensure that your driver and system software
>
> are above reproach before messing with hardware, particularly
> if the board otherwise appears to be working normally and the
> heatsink is not glowing cherry-red....
>
> There is a very, very remote possibility that the thermal sensing
> internal to the GPU or its board interface electronics is faulty.
>
> Anyway, I assume that the board is pretty new. Did you contemplate
> a RMA before the major surgery that you have performed ?
>
> Apparent gross overheating in 2D or quiescent-3D mode, without
> any obvious physical overheating at the heat-sink is sufficient
> grounds for a RMA from any reasonable board-vendor.
>
> John Lewis

Listen to John, G-imme, he's probably onto something here.
Anonymous
April 7, 2004 6:26:49 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

Well, I did use Driver Cleaner for the display drivers and nforce2
chipset. Nope.
Tried setting the AGP to 4x. Nope. Added another fan . Nope. Tested
running with the case open. No difference observed. Tried different
versions of drivers after 'cleaning' the previous installed one from
the system. Nada. Nothing else is overheating. Oh yeah, I bought this
card via Ebay. so if I must run it at 300MHz and the memeory at 650Mhz
(325) then so be it.
April 8, 2004 6:03:01 AM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

greg_2day@hotmail.com (G-imme) wrote:
> Well, I did use Driver Cleaner for the display drivers and nforce2
> chipset. Nope.
> Tried setting the AGP to 4x. Nope. Added another fan . Nope. Tested
> running with the case open. No difference observed. Tried different
> versions of drivers after 'cleaning' the previous installed one from
> the system. Nada. Nothing else is overheating. Oh yeah, I bought this
> card via Ebay. so if I must run it at 300MHz and the memeory at 650Mhz
> (325) then so be it.

The most important report that I'd like to hear: do your heavy games
and software work without lockups? It's like John said earlier, your
card's inner thermal sensor might be faulty, or then again not. If
the applications work fine without the freezes and crashes, then you
shouldn't pay too much attention to the thermal sensors and
measurements. There's been motherboards where I've gotten reports of
having CPU temperature below 50 Celsius and then after installing it
to another nForce 2, it says 72 Celsius (which is probably the more
accurate core temp in this case). I have no lockups, no game
freezing.
April 8, 2004 10:11:35 PM

Archived from groups: alt.comp.periphs.videocards.nvidia (More info?)

Nada wrote:

> greg_2day@hotmail.com (G-imme) wrote:
> > Well, I did use Driver Cleaner for the display drivers and nforce2
> > chipset. Nope.
> > Tried setting the AGP to 4x. Nope. Added another fan . Nope.
> > Tested running with the case open. No difference observed. Tried
> > different versions of drivers after 'cleaning' the previous
> > installed one from the system. Nada. Nothing else is overheating.
> > Oh yeah, I bought this card via Ebay. so if I must run it at 300MHz
> > and the memeory at 650Mhz (325) then so be it.
>
> The most important report that I'd like to hear: do your heavy games
> and software work without lockups? It's like John said earlier, your
> card's inner thermal sensor might be faulty, or then again not. If
> the applications work fine without the freezes and crashes, then you
> shouldn't pay too much attention to the thermal sensors and
> measurements. There's been motherboards where I've gotten reports of
> having CPU temperature below 50 Celsius and then after installing it
> to another nForce 2, it says 72 Celsius (which is probably the more
> accurate core temp in this case). I have no lockups, no game
> freezing.





It's not a software issue... that's why I totally avoided the driver bs.


The card would eventually hault to a crawl if it got hot enough,
despite what the thermal probe reads.

It uses the same proprietary thermal clock reduction used in the
Pentium 4.

The clock speed is lowered gradually as heat increases, and this is all
done through hardware.


If you bought it on e-bay it could possibly be that someone already
attempted to flash the BIOS, and did so without research.




Go with flashing the bios, like I said.



Prolink won't bitch at you for flashing the BIOS, but might scrutinize
you if you use another manufacturer's BIOS.



What's the worst that could happen?

Oh wait, it's already happening =P.
!