Who's to blame: The Asus A7A266 debacle

[Prien]

I recently purchased the above lemon and (other than the HD LED not shutting off), it seems to crash whenever I try to use a 3D application (Unreal Tournament, et al. I've tested the GFX card on another board, it works fine), it crashes after ~10 minutes. The bios drivers are updated as are the ones for the card. I've even installed Win2000 on the machine to see if it was OS dependent; my reading of the review on this site confirms my suspicions (http://www.tomshardware.com/mainboard/01q2/010509/index.htm). My question is two-fold.

First, does anyone have any suggestions concerning this problem? (Other relevent information: Card is VisionTek GeForce2, it has the IRQ 11 and shares with no other device)

Secondly, can I get a refund on a product because faulty design does not allow it to meet design specifications? (Which would probably put microsoft out of bussines=))

 

[Crashman]

I think your overheating. I seriously doubt that the motherboard is to blame as I am only dealing with one person experiencing problems with that board, and his problems appear to be in software. The BIOS revision 1004 fixed any instabilities with the motherboard itself. What version of the Geforce2 are you using, the original? What driver revision?

Cast not thine pearls before the swine

 

[Prien]

If it were overheating, first off, the entire machine would've reset, or so I'd think. Secondly, according to the board, the temperature is running at about 53 degrees celsius in the CPU (athlon 1.3 Ghz 266 FSB) at one point during testing. The tempersture was less on the mainboard. Not that hot at all. I've also tested the board after it's been running a while and after a rested boot. Neither the board nor the card is overclocked.

As for driver versions, I'm using BIOS revision 1004 (unless they've come out with an update today). As for the GeForce, it's a VisionTek GeForce2 GTS w/64 megs of DDR RAM. I've using the denator 3 v. 12.41. I also updated the AGP controller from the ali chipset homepage. DirectX version in 8.0. All this information should be true for both OSes I tested on (Win200 and Win98 SE).

 

[Crashman]

Drop the new driver, I keep hearing of problems with it. Go to www.guru3d.com and get an earlier Detonator3 version. I like version 6.31, even though it is old, because it seems to be the most stable. I have NOT been able to get increased performance from newer revision, only less stability, or even less performance with certain revisions. The double-digit drivers were optomised for the GeForce 3, not the GTS, and are not needed with your system, probably contributing to instability. If you do not know how to completely remove the 12.41 driver, they have an article on the main page. I would suggest the 6.31 once again as I have experienced other revisions unsuccessfully.

Cast not thine pearls before the swine

 

[Prien]

Ok, I'll try that. But would it really cause my card to be *that* unstable? I'll post the results here tomorrow after I get some rest. I'll also probably test the drivers I'm using right now inside another machine, just to be sure.

 

[Crashman]

Remember that certain drivers affect certain chipset differently, least we forget the SB Live/VIA debacle.

Cast not thine pearls before the swine

 

[Prien]

The plot thickens. I donwloaded the new driver (or rather, the old driver, 6.31) and installed it...and it's still crashing. I also decided to check to see if it's a specifically direct3D thing and no, it's not. It crashes under Age of Empires 2 as well about as frequently as during Unreal Tournament. 2-10 minutes and *BLAM*. The game freezes or I'm given some funky error message. This only seems to happen in DirectX (i.e. windows itself can run for hours without issue). It seems like a board issue to me. Anyone help? Or do you think the board's dead? I think I'm going to reinstall directx now....

<P ID="edit"><FONT SIZE=-1><EM>Edited by prien on 06/29/01 04:24 PM.</EM></FONT></P>

 

[Crashman]

You loaded ALi drivers and not VIA drivers correct? Since I have never seen this problem with this board, I can only assume that it is either a defective unit, misconfigured, defective video card, or inadequate power supply. I would try all the parts in another unit with the same power supply

Cast not thine pearls before the swine

 

[Prien]

Yep, it's the Ali drivers. Reinstalled Directx8.0 for no effect. As for being misconfigured, I'm not sure exactly how. I seriously doubt the video card's defective as I used it in another machine machine (also with a 250 watt power supply) for over an hour today in Unreal Tournament and it played fine. Is 250 watts adequite for such a machine? The only peripherals in it are the card and a modem. Now, I do realize the AMD processors are power hogs, but....

 

[Captain_Splat]

I had the same problem with my board, it cost me forty bucks to find out that I'd plugged the hard drive led onto the wrong pins. The a7a manual gives two different locations for the pins, the right pins are a bit off to the side of the row of pins on the board. Hope this helps, I've had no problems with stability since then, and I haven't even updated to the latest bios yet!

 

[Crashman]

But......but the main thing that comes up short with Athlons and Geforce 2's is the 3.3v power. So the number of drives would not be the most significant reason you need more power, the video card and processor are. Most people suggest a MINIMUM of 350 Watts for your motherboard, processor, and vidoe card combination. But I would suggest that as long as your only using a couple drives and a couple cards, you can get by with 300 Watts. Since you may want to add more stuff later, you really should consider a 350 watt power supply.
Most power supplies are overrated, so I would look for a quality unit. I have heard good things about Leadman power supplies, but have never used them myself. The only brand name power supply I have had problems with has been Antec, so I would suggest that most other brand name power supplies are fairly good.
You should know that power supplies can reduce output as they get hot, which could be part of your problem. And the resistance of the circuites increases as components get hot, meaning that at about the same point as the power supply begs for mercy, the components beg for more! So get a bigger power supply, and no lightweight garbage, get a good one!

Cast not thine pearls before the swine

 

[Crashman]

So to answer the original question, who's to blame, the answer is probably that you are. I don't know of anyone who recommends using a 250-watt power supply on ANY Athlon, especially when equiped with a GeForce2!

Cast not thine pearls before the swine

 

[Guest]

"I'm given some funky error message"

please copy the error messages here with a post so we can review the situation with this in mind. also, i agree with Crashman on the power supply issue. that is a huge athlon and really needs a larger supply. Also, in my experience IRQ conflicts can cause these sort of nasty lockups. check to see if any of your devices are sharing IRQ's. chances are they are......with any luck you'll discover that your video card and sound card are sharing an IRQ (or something like that). this kind of a conflict would only manifest problems (such as lockups or errors) when both items are used heavily.

ignore everything i say

 

[Prien]

Thanks, I read that they (being the Athlon processor and the GeForce) were power hogs but in the manual for my video card it said that 250 watts was fine. I suppose they didn't count on an Athlon 1.33 Ghz processor=) My other computer is a PIII 500 with 2 hard drives, a DVD drive, a burner, a TV tuner card, a modem, a network card, and an ethernet card. I'm using 250 watts on that and haven't had any problems so i figured this'd be the same.
As far as my "funky" error messages, it's either a 3 line long message about an incorrect pointer reference, which is usually preceded by the controls freezing (the game goes on, but I stand still and get shot). Same thing for Age of Empires. I'm testing right now by playing AoE at a low resolution which I would think would require less power, and should be fairly stable.
Is there any high end application (read: non-directX) that I could try to crash the machine (I try compiling the Linux kernel=)) with to isolate the problem from being between my motherboard and directx or does what's happening so far sound like enough to be positive it's the power supply?
Oh and BTW, I checked for IRQ conflicts before I did anything. Each device on this machine has its own IRQ.

<P ID="edit"><FONT SIZE=-1><EM>Edited by prien on 06/29/01 11:49 PM.</EM></FONT></P>

 

[Crashman]

I am about 60% sure it's your power supply. You need to remember that most of your drives run at 12v, so it is less of an issue with your PIII system. My advice if you want to verify that it's a power supply problem? Lower your system bus to 100MHz and try all your games. Try 3D-Mark 2000. Then raise your bus speed to 133 and try non-3D high stress applications.

Cast not thine pearls before the swine

 

[Prien]

It seems I found the solution to my problem. Like so many other problems in life, doing nothing seems to have been effective. That's right. After AoE2 crashes running at 1280x1024 in 32 bit color, i rebooted and left. I then played AoE2 again, but in lower resolution. For some odd reason, I decided to up the resolution to its previous state. I played for 1/2 hour and it didn't crash. Hrmm...that was wierd, i thought. I then ran 3DMark2000 and it didn't crash either (and posted some nice scores too). I then decided "What the heck, let's try Unreal Tournmanet again to make sure there's something wrong, this machine hasn't crashed in 6 hours". So I did. After running it in 1280x1024 32 bit color high detail (map: deck16II with 12 players) for about an hour, it didn't crash (I consider this a good test of performance). I'm dumbfounded. I'm keeping my fingers crossed, but no doubt when I wake up in 8 hours, it'll probably be crashing like its old self. If it continues to work, I'm sealing it shut and taking a disk image=) Oh well, off to bed.

 

[Guest]

As an aside to the graphics complaints. I have a Promise RAID FASTRAK100 card for my hard drives. The cmos is set in pnp mode but the machine will not see it. I have ripped every other card out & exchanged video cards and exchanged slots with the RAID; no go. I changed the cmos to nonpnp and the RAID shows up on IRQ 11 and NT2000 boots up. There is something weird to this Asus ATA266 at the H-ware level and I'm not even overclocking. I also have this machine on a switch box. When it does work (non-pnp)and switch over to an Apollo VIA P3V4 based machine, I lock up the mouse(IRQ12) and or the keyboard(IRQ1). Your note on IRQs is interesting, but I had only the video and the RAID troubleshooting at one point.My RAID card works like a dream, but only if the CMOS is set at non PNP OS (even tried an old PCI video card). I also noticed a article on the ALiMagik1 in Tom's in this chipset being lousy at overclocking incidently. Sorry if I haven't got the answer, but I think it's pointing towards the chipset. As the months go by, either the chipset manufacturer or the MB manufacturer may offer a trade in or credit through your vendor. I got suckered by the Rambus, and got a new MB.

 

[Crashman]

Tom did not like the fact that this motherboad did not have multiplier swithches, which means that it is still excellent for overclocking by FSB, but multiplier changes are dependant on the pencil trick. As for your IDE controller problem, you should have had your BIOS set to PNP-OS-NO anyway, because that forces BIOS to detect the card instead of Windows. Why would anyone want their RAID controller to only work AFTER windows has loaded? That would mean booting from the onboard controller! You need BIOS to configure the card, and that means PNP OS-NO!

Cast not thine pearls before the swine

 

[Prien]

As a general rule, you don't use PNP OS with Win2000 anyway, otherwise all PCI devices will be assigned to the same IRQ.

 

[Prien]

I test it, it works. My brother tests it, he managed to crash it. I probably didn't run it long enough on one map to do anything. The error message I get in UT is as follows:
Title "Critical Error"
Text "History: UD3DRenderingDevice :: DrawComplexSurface <- URender :: DrawFrame <- UObject :: StaticFindObject <- UObject :: StaticAllocateObject <- UObject:: execClassContent <- (TMale1 CTF-PolarOpoosites. TMale2 @ Function Engine.PlayerPawn.RecievedLocalizedMessage : 0024) <-AActor:rocess State <- Object Enforcer CTF-PolarOpoosites.Enforcer155, Old State State Botpack.Enforcer.NormalFire, New State State Botpack.Enforcer.NormalFire <- AActor :: Tick <- TickAllActors <- ULevel: Tick <- (NetMode = 0) <- TickLevel <- UGameEngine :: Tick <-UpdateWorld <- MainLoop"

This occured while playing a map entitle "Polar Opposites", my brother was using the Sniper Rifle at the time and there were 15 bots. Resolution 1280x1024 32 bit color, high detail. I'll probably get the power supply now, but I don't want to spend $60 on something I don't need.

Oh and lastly the temperature last night reached a peak of 150 degrees Fahrenheit. How how should the CPU be allowed to get?

 

[Crashman]

You are overheating. Your processor should NEVER go over 140F. Never. In fact, if you can keep it UNDER 140F at all times, you are doing good enough for cooling.

Another poster mentioned that he needed the latest ALi AGP driver. Perhaps the one Asus supplies is an old revision. Asus makes great boads, but their software package leaves much to be desired!

Cast not thine pearls before the swine

 

[Crashman]

BTW, on the pwer supply issue-I have never seen the combination of the 1.33 and GeForce2 work with a 250 watt power supply, so the fact that yours works at all is pretty amazing!

Cast not thine pearls before the swine

 

[agpport]

AMD says it (Athlon) can take about 90 degrees C, but you should try to keep it below 60.

To me it seems obvious that the PSU is the bad guy.

What cooling do you have?

Anyways, what you might need is: A powersupply at least 300W, a decent HSF and maybe a nice case fan.

Good luck!

 

[Guest]

an easy way to test the heat theory (i agree that it seems way to hot) is to open the case and place a house fan about three feet away pointing right at the case. crank the fan up and then play for a while.....bet it settles down. I think your graphincs card is overheating, no the cpu. first, your resolution is probably on the high side. resolutions over 1024 X 768 really strain 32mb chips. yeah, they can handle it, but the extra duty forces them to exchange data that much faster. second, its hot in there.

So, open the case and cool it down. notch down the resolution (temporarily for troubleshooting) and watch it get nice and stable (we hope)......

ignore everything i say

 

[Prien]

It's not the card overheating. I tested it in another machine, remember? And it's a 64 meg chip thank you=)