Kadgar34

Hi There,

I've had this ATI Radeon 9700 pro for a few years now, and all of a sudden it's been giving me a lot of noise and problems during boot up, and during normal usage.

Here is a screenshot:

http://www.student.cs.uwaterloo.ca/~...ugged_gfx1.jpg

Note that this is from World of Warcraft. A few things you should notice: Under the portrait of the person in the top left there is a yellow bar, on the bar are some numbers (which show the current experience needed to level). There are two digits which are black, it's like they just disappeared leaving the black part behind. Also, there are a few NPCs on the screen, and it seems a few polygons on some of them got stretched to the top right of the screen. There are also a lot of other things, like the artifacts all over the terrain and things like this.

What would be the cause of all these errors?

EDIT: Couple other pieces of information: These all started after I thought changing the AGP Aperture size would be a cool thing to do, since I read that changing it can increase performance. I use Windows XP SP1, and this machine is a Dell Dimension 8200. 512MB ram too.

Thanks.

Sarkast

Maybe you could describe the "noise and problems during bootup" a little more in detail aswell as from what to what you have changed AGP aperture size and if you have changed it back to whatever it originally was.

Usually when corruptions like that occur it could be the card is overheating, underpowered (or in the worst case damaged).

For now i'd try to open the case and improve airflow (forcing air in the case with a huge fan) to see if the problems go away. (By the way does this corruption happen from the start in a game or does it take a while to appear?). Also try reseating your card and while you have it out make sure the heatsink is properly touching the GPU and Ramsinks (if designed to do so).

It could also be useful to download ATITOOL and run the 3D cube to see if you have artifacts.

EDIT: Are you using the latest drivers ?

Kadgar34

Sorry, I'm not currently at my computer so I can't get any screen captures. The errors that happen during boot are a little hard to explain, but there is another problem. It doesn't always happen in just 3D applications and games. It can happen at any time, generally when the computer is under a lot of stress, but it occasionally will happen when the computer is fine.

This is how it goes down, basically:

I boot up my computer, and when it gets into windows and starts to load my applications at start up 5 lines will appear on the display in an even fashion.

Here is a screenshot:

http://www.student.cs.uwaterloo.ca/~k3ali/lines.jpg

When these lines appear, the 3D distortions in games like World of Warcraft occur. Also, the problems were occurring when I had the newest drivers (Catalyst 5.3) and recently i downgraded to Catalyst 4.9 and it still happens.

I will try to force more air into the computer when I get home. Thanks for the help.

Skizzik

I have the same type of corrupted display as Kadgar34's second screenshot. I also have an ATI Radeon 9700 PRO 128mb in my machine:

CPU: AMD Athlon XP 3200+
Motherboard: ASUS A7N8X Deluxe
PSU: Enermax EG465P-VE (465W)
Memory: Corsair XMS TwinX 1GB (2x512mb)

I am currently running off an old Riva TNT2 32mb PCI card until I find a solution to the problem or I get a new card.

The first odd error I got was yesterday, during a game of Counter-Strike. The game crashed and I got an ATI popup saying something about VPU recovery and that I needed to reboot my computer. So I did and was able to get back into playing the game again. A couple of hours later, during a game of Lineage 2, random shapes and lines appeared all over the screen. I was able to close it down and restart it, then I got all the random lines as in Kadgar34's second screenshot. From then on, whenever I start up my computer, the ASUS boot-up screen for my motherboard was covered in small squares and random characters of random colors appeared in the DOS screens that followed.

I tried updating my Catalyst drivers to the newest, flashing my motherboard BIOS to the newest, and cleaning the fans and such inside my computer tower, especially the one on my video card (which is just what came built on it from the factory). It seems like my card is overheating (it does seem to feel a bit warm for only starting up Windows XP).

If it did overheat, is it as good as dead now? Or can I still fix it somehow? Or, if overheating is not the problem, what is it?

Any help is appreciated. Thanks in advance.

UncleMacro

To Skizzik:

Since it happens in the BIOS screen you know it has nothing to do with drivers. About the only things you could possibly do is flash your motherboard BIOS (which you've already done), flash your video board BIOS (which is probably pointless), or fiddle with some of the lesser known AGP settings in the BIOS. I doubt any of those would accomplish much because that second screenshot usually indicates bad video RAM.

Most of the time bad video RAM has just been too hot for too long and has finally given up the ghost. The only "good news" is that once in a blue moon, bad video RAM which appears suddenly is caused by too much dust (and associated gunk) building up on the pins of the GPU or RAM chips. You might just take out the video board and try to blow all the dust out of everything very thoroughly including around the pins on all the chips.

You could also download ATITool and underclock the video RAM and GPU as described here. There is some chance that reducing the clock rate of your GPU or video RAM would solve the problem although that won't help you while you are booting because the underclocks just take affect after Windows starts. You can probably actually flash your video BIOS to tell it that it's a Radeon 9700 non-pro to slow it down a little but there's no point in doing that until you see if it works properly at non-pro speeds (GPU=276MHz ,RAM=270MHz).

You can test the overheating theory by opening the case, getting a desk fan, and aiming it at the video board. You have to make sure the air also gets the back of the video board because four of the RAM chips are on the back of the board.

Skizzik

I tried out UncleMacro's suggestions, but I still have the same problem. I cleaned the little dust I found on the board around all the pins of all chips. I laid the tower flat on its side and left the side panel open to allow as much cool air to blow on it as I can (had the window to my room open and the room's ceiling fan on, etc.). I attempted underclocking the card with ATITool, but apparently the symptoms got worse before I could. My computer used to be able to stay in WinXP long enough for me to try out some things before turning off my monitor (at least that's what I think it is doing, my monitor flicks to a completely black screen). Now it goes completely black as soon as it hits the WinXP Welcome screen, so I could not underclock my video card. However, I was able to underclock it in Safe-Mode. I am not sure if this gives the same effect, since, from what I understand, the video card driver gets disabled in Safe-Mode.

Could it be possible that my AGP slot is damaged? I do not have another AGP card to test it out with (like I said in my previous post, I am using an old 32mb PCI card at the moment).

I am considering buying a new card, a cheap but decent card. I am also leaning towards an nVidia card, for better cooling and longer life. I can get an MSI FX5700LE TD256 for a pretty good price. Would this card be a good choice for someone in my position? Or does anyone have any other suggestions (for fixing my current video card or getting a new one)? And how sure could I be that this will not happen to my new card if I were to get one?

Skizzik

Sorry for double-posting, after browsing through the boards a bit, it seems that the FX5700LE is a bad choice. I can also get an MSI FX5700 VTD128 for a bit cheaper than the LE one.

UncleMacro

Given that you're getting artifacts in 2D mode and even on the BIOS screen sometimes, it seems unlikely that the AGP slot is the problem. You can slow the AGP slot down in your BIOS by disabling AGP 8X and disabling fast writes if you haven't already. That will slow the AGP slot down to a relatively pokey AGP 4X. It would be nice to be able to test another AGP card in the slot just to be sure but your symptoms sure sound like the video card and not the AGP slot on the motherboard. If you could get the thing working for a while in Windows with the ATI driver loaded you could use SMARTGART to set AGP speed to off. That essentially turns your AGP slot into a plain old 66MHz PCI slot. If you can boot in safe mode, you should be able to tell windows you have a dumpy old VGA monitor so the next time Windows will try to boot in 640 x 480. Maybe that will give you a chance to fiddle with it while not in safe mode.

It sounds like you may not be too positively predisposed towards ATI at the moment (which is understandable) but it's worth pointing out that an FX5700 is about the same speed as a $95 Radeon 9600 Pro. See here and here for large scale comparisons of video cards. I visit quite a few forums and I see roughly equal numbers of "I'll never buy NVIDIA again!" and "I'll never buy ATI again!". Unfortunately, NVIDIA's offerings really aren't that great until you start hitting the $190 6600GTs. The 6600s and 6200s 128MB are not bad choices in the under $140 range especially if you're willing to do some relatively easy GPU overclocking. But none of them will really compare with your (probably terminally ill) 9700 Pro.

MD_Willington

I'm thinking a couple things here:

Whats the true capacity of your power supply?

What I mean by that is not whats written on the side like "400W" I mean what are the amperage (A) specs for each voltage on the side of the power supply for:
3.3V = ?A
+5V =?A
+12V =?A
+5VSB =?A

MD's Temperature related standard checklist. (it's standardized for AMD processors but the same stuff holds true for an intel system.)

Check the following for overheating or perceived overheating:

- Is the heatsink correctly installed ?
--> Verify it is correctly installed.

- Is the heatsink filled with dust and crud ?
--> Clean with compressed air in a can.

- Is the heatsink fan operating ?
--> Replace defective fan with similar or better unit.

- Is the heatsink adequate for that application/CPU ?
--> Verify it meet the requirements for the CPU being used.
--> RETAIL heatsink? May require lapping and better TIM.

- Is thermal interface material used between heatsink and cpu ?
--> Verify it is used, and is applied according to the manufacturers guidelines.
--> Recommended PCM's, thermal pads from "AMD Thermal, Mechanical,and Chassis Cooling Design Guide #23794.PDF.
- Bergquist HF225UT Phase Change
- Chomerics T725 Phase Change
- Honeywell PCM45 Phase Change
- Power Devices Powerfilm Phase Change
- ShinEtsu PCS-TC-11T-13 Phase Change
- Thermagon T-pcm905C Phase Change
--> Recommended thermal interface material aka: grease, gel etc.
- Arctic Ceramique.
- Arctic Alumina.
- Arctic Silver, 3, 5 etc.

- Is the TIM pad cover removed from heatsink before installation, if using supplied TIM pad ?
--> Remove the cover before installation, failure to do so will insulate the cpu from the heatsink.

- Is there adequate airflow or case ventialation ?
--> Verify airflow, add case fans if required.



- Is the Vcore set higher than specified by the manufacturer ?
--> Verify the Vcore is set to the manufacturers specification in the mainboards bios.

- Is the temperature read via the internal temperature diode. Is the tremperature read via the socket thermistor ?
--> Temperatures read at the diode can and will be higher than an external reading.

- Is the temperature sensor properly calibrated ?
--> Several boards have this problem, check manufacturers bios revision history for calibration fix.
--> Does the heatsink feel warm or HOT, if it's warm it could be a miscalibration error.

- Outside temperature high ?
--> If the external temperature is high, then the case fans will be drawing in hot air, especially during summer months.

Video card overheating?

Do the finger test on the video card heatsink, if it's too hot to touch, then it's generally too hot to run properly.

MD

gaioshin

Hi all,

I had the dreaded 9700 bug with my card.

I read this:-

https://support.ati.com/ics/support/...task=knowledge

I actually tried swapping my power supplies about when that happened... surprise surprise, it made a difference - a cheap and nasty PSU could not keep my graphics card alive!

On the other hand, my card has a weird habit of changing clock rates when that happens. I often use ATITool to make sure that the clock rates are as close to the default as possible - as the default is underclocked for mine :(


Hope that helps.


Gaioshin