I think I know what the issue is and how to solve it!
Like a lot of other folks I have had issues with this card.
The first problems showed when playing Bioshock. Regular crashes to desktop or total system shutdowns occured within a few minutes of gameplay. I went through a lot of testing and fault finding and eventually came up with a working solution.
1. The card is power hungry and a 12V rail of 30A minimum is recommended.
A. I swapped the PSU and gained a limited amount of stability.
2. The card runs very hot and the heatsink design, particularly on the Sapphire Radeon card, does not allow a proper contact with the VR chips. In addition, if you remove the original heatsink you may not have any cooling at all on the VR chips using an aftermarket VGA cooler. The Arctic Cooling X2 is an exception to this, with a heat spreader supplied.
A. I cut off the part of the original heatsink with the VR cooling pads on and thermal glued it in place. The Zalman cooler does not have a provision for VR chip cooling.
3. There is some sort of chip almost directly behind the GPU on the reverse of the card, near the AGP slot. This also gets very hot and came under suspicion of being a problem.
A. Stuck a heatsink on it.
Now some of these 'solutions' seem a bit DIY, but they do work. The card now runs rock solid in Bioshock and 3DMark06.
If you are thinking of upgrading the cooling, then I would advise on making sure you get a VR chip heat spreader, a la AC X2. I have also seen a Radeon O/C'd X1950 PRO with a gold VR chip heat spreader fitted.
If you have a stock card and you have issues, the chances are that its heat or power related. First try another PSU that's man enough for the job, if no luck, then consider RMAing the card for another type.
The PSU gave more stability, but it was the VR cooling that really solved the problem. The Radeon card stock cooler does not fit correctly onto the VRs and as a consequence causes problems when the going gets tough, graphically speaking. My card had no VR cooling, as the person who fitted the card with the Zalman VF900-Cu, did not fit any, if indeed any was supplied, as this VR cooling seems to be overlooked in a lot of cases.
The two main issues with this card are:
1. A lack of power supplied to this card. Consideration needs to be given to existing power supply capability before upgrading.
2. Inefficient, if any cooling of the VR chips, causing system instability.
I have the Connect3d x1950 PRO 256mb card, idles at around 46-50c, when playing call of duty 4 it gets up to around 70 before I get "system hang", locking my computer which requires a restart by holding power button. Is this normal at all for a card like this? Shouldn't it be able to perform at temps higher than 70?
Most graphics cards won't hang or throttle until they hit 100 celcius. I know this is the case with my x1950xt pci-e card, and it should not differ from the agp card since we're talking about stress on the gpu itself. Despite pretty extensive efforts, my card will get up to 93 degrees playing bioshock and UT3, but has no stability problems/artifacts.
It may be a power supply problem, being that 1950's are pretty inefficient cards and need as many amps as they can get. If you're running a psu that doesn't meet the specs of the x1950 (30 amps on the 12 watt rail) you'll probably have problems.
This might sound like a silly way to do it. Take out your card, bring your mouth to the grill and blow really hard. Watch all the dust come out of the fan section. I did this my card went form 85-90 down to 59.
I myself have bought the sapphire x1950 pro and at idle the temp will go up to 70 C and with games up to 80-85 i's. The card won't crash and does not gve me and problems so far but i think these are risky temp to have, if im not mistaken. I'm not sure but i think it's my Psu that's the problem. Here is my comp specs
I am having the same problem with overheating on my Sapphire X1950 Pro. Sapphire agreed to RMA the card, but now I have to decide which makes the most sense.
1) $15 to RMA the card, but I will be without a computer until I get it back (no other PCI-Ex. card)
2) $40 Buy a cooler (TT CL-G0102) supposedly works well.
3) Buy a new card entirely. HD 4670 ($85) or 4830 ($130).
I would rather not spend the money as the X1950 works well when it actually runs cool enough. Any advice?
I have an HIS x1950pro which runs up to ~80 deg when running WoW. Unfortunately when it goes above ~60 deg I get all sorts of graphics artifacts. Even tried cleaning and re-seating the HSF with some Artic Silver, but no improvement. I found using ATI Tool that even at 80 deg, the fan was only running at 46%. My solution was to use ATI Tool and run the fan at 85% whenever I'm playing WoW. That usually keeps the temp down enough to avoid the artifacts. Noisy, but it works. Personally, I'm just going to do this until I can find the money for a 4670. I'm sorry I ever got the x1950.
I am in the same boat. Upgraded my 4-year-old Compaq system for the heftier games coming out...namely Empire: Total War. The motherboard has a 754 socket, so the best AMD processor I could find was the AMD Athlon 64 3400 2.4ghz...which runs at 1.8ghz at idle (the BIOS will NOT allow ANY overclock of the CPU ). Upgraded to my motherboard limit of 2gb of RAM...added half a tera of hard drive...and added the x1950pro Ice Q turbo card, which ended up in me having to upgrade my power supply. I basically have a modern computer without the headache of Vista which was the whole idea of the upgrade. Anywho...
I first experienced the artifacts in Medieval II: Total War. Then in SimCity Societies/Destinations...which is a pretty greedy little memory/processor hog. Did the whole driver song and dance...tried originals and brand news to retro upgrades...still was artifacting. I ran across this page and learned about the ATI Tool which has been a bit of a savior as of yet. I am running the fan at 100%...dropping temp down to 38 degrees celcius (100 farenheit). I noticed my artifact starts at 56/57 degrees celcius..roughly 130 degrees farenheit. I am going to run SimCity Societies on various graphics settings and see what I can do as far as pushing this thing without overheating. I am mostly worried about slow, exponential growth in heat which will determine how long I can play a game with certain settings before the GPU overheats. It would seem that the idle is at about 45 degrees celcius/113 farenheit. I will say that my crappy old Ati Radeon (can't remember model...it was a mediocre card back in 2005/2006/) 256mb/128bit wasnt quite as visually spendid, but at least it ran efficiently. There is no reason for this card to get so hot/for the fan to not run properly without us having to customize the settings. I will admit that the cooler is relatively quiet, so that doesnt really bother me...it is just the concept that if the internet community didn't exist, I would know where to get the ATI Tool in order to force the fan to run....
Ran SimCity Societies on High settings in windowed mode...clocked to 590/770 Mhz. Stayed at about 43-46 degrees celcius (109.4-118.8 farenheit) with CPU fan kicking on and off. I guess I will just turn the fan on 100% every time I play a game. Seems to be the only lead.
August 29, 2009 10:26:00 PM
I have never replied on a forum before with a solution...but I was having such problems with my card and this worked so well:
As soleft said earlier in the post, I took my card out and blew all the dust off it, but the worst bit was towards the back of the card, where there is a rectangular inlet. I didn't even know that air was supposed to flow through there as there was so much dust covering it up! All I had to do was blow all the dust off of it!
I used to have major problems with the fan going full speed for every single game (very noisy) and it would restart my PC all the time if I tried to run something too graphics heavy! After this I ran 3DMark 03 and 05 with absolutely no problems and the fan didn't even spin up to anywhere near full speed! (it normally always crashed when I tried to run those)