So yeah, I'm having trouble with my computer already. As stable as Win2K was on my old machine, for some reason it's a buggy mess on my new one.

Every so often doing the most random thing will cause the whole system to restart. I went in to redirect the "My Documents" folder, I click on the "change" button and all of a sudden the screen goes blank and I'm booting up again. Disconcerting, yes.

Then smaller things will happen. Like I inserted my motherboard's driver disk, and the driver install program wouldn't run. Something like "Application failed to initialize." This is annoying, because I needed to install drivers to see if I could fix the problem of my CF card reader not being recognized anymore.

When I first set everything up, I put a CF card in the reader (it came with the mobo), and tada! There was a new drive in "My Computer." I was able to see it just fine. But then overnight, something changed and now when I put the card in, it does nothing at all!

The last thing I've noticed was that after about 5 minutes of playing GTA3, it'll just lock up and I can do nothing but a hard restart. This is most likely a problem with the game, though, as it sucks.

Do you think there's a general problem here? Is there something in my system that could cause most of these glitches?

Also, for some reason Windows keeps bringing up some screen about a "digital signature" or something every time I try to install new soft/hardware. Can I turn this off?

Thanks for any help you can give me!

trouble ...mess ...restart... screen goes blank ...Application failed to initialize... it'll just lock up... Do you think there's a general problem here? Is there something in my system that could cause most of these glitches?

Sounds like you've got a hardware problem (CPU, motherboard, or, most likely, RAM). Try booting a memtest86 floppy and leaving it going for a while. http://www.memtest86.com

Also try making sure everything (especially RAM) is seated properly. If all else fails, try underclocking the FSB/RAM. If underclocking makes it all work fine, it's fairly certain there's a hardware problem.

Peter

Oh man, I don't want to hear that....argh. that's not good. I'll run that test tonight.

Oh man, I don't want to hear that....argh. that's not good. I'll run that test tonight.

Other lines of attack might include removing one or more DIMMs (if possible) or checking on the CPU heat situation (i.e. is the heatsink on properly) using something like Motherboard Monitor http://mbm.livewiredev.com.

Peter

I agree with Peter. Some hardware thingie in your PC is bad, indeed, most likely the RAM.

Windows 2000 automatically reboots on occasions like that. Very frustrating, I know.
You can actually disable that "booting", so you can see the actual error.

Go to system properties, (right click my computer, then properties) -> advanced -> reboot and repair eh stuff (option on the bottom) -> uncheck third option with system errors ("automatically reboot")

I'm trying to translate the exact steps here since I'm using a Dutch windows 2000, but you should be able to find this no problem.

Of course, then the painful process of debugging you hardware only begins...

Thanks for the help guys. Man, I hope it's just the memory if anything at all. God please say it isn't the motherboard. I don't think I could deal with that. The CPU wouldn't be much better.

If it helps, I have gotten a few blue screens, mainly on shutdowns I think. And when Win2K was first installed, I saw on that mentioned like a memory dump or something.

I was interested in disabling "auto reboot on error" so I went to look for it. You got the directions just about right. After clicking "Advanced" there's a button marked "Startup and Recovery..." - clicking that will bring you to the dialog box with the "Automatically Reboot" option so you can shut it off.

When I put Service Pack 2 in I hadn't checked for incompatibilities - my mouse software made the computer restart something like 7 times before I caught the name of the driver and realized it was the mouse. I have upgraded so it won't happen again... at least not from the mouse.

When things go wrong with my (or a) PC this is the order I use to debug it:
First I remove all the not-so-necessary cards (soundcard, NIC,...)

Then I replace :

1) memory
2) motherboard
3) video card
4) harddisk
5) CPU

In that order. In case the CPU is an AMD I'd switch 4 & 5

my mouse software made the computer restart something like 7 times before...

You have moved your mouse. Click OK to restart your computer for the changes to take effect.

you mean replace as in "get a new one"? Oh cripes. I reeeeally don't want to replace the motherboard. The CPU and memory have warranties on them, and they're not so hard to replace physically. the motherboard would be a giant pain.

Please, please don't let it be the motherboard...

I really hope I don't have to replace any parts, frankly.

By the way, my Win2K came with service pack 2, and I also had to restart in the middle of an installation to get it to recognize it. The Logitech mouseware didn't even get loaded after that. I had to reinstall it before it finally took over.

Please don't let it be a hardware problem.....:(

ROFL. No, it was much more dramatic than that. Windows would not even have finished loading the GUI and I'd get Win2K's blue screen of death. The reason it restarted 7 times is because I couldn't catch the name of the damn driver and didn't know how to stop it so I could read it. So it took me that many tries to catch the filename before it restarted. That was great fun.

you mean replace as in "get a new one"? Oh cripes. I reeeeally don't want to replace the motherboard. The CPU and memory have warranties on them, and they're not so hard to replace physically. the motherboard would be a giant pain.

Uhm. yeah. I guess I'm spoiled because my best friend owns a computer shop and every time I put a new pc together and something isn't right, I just take a new piece of hardware of the shelf to replace the possible culprit. This has been going on for so many years now, I sometimes forget other people don't have that option. Sorry.

I don't want to be pessimistic, but the last time I encountered such a problem it was a combination of memory and motherboard. Both were replaced.

Here's a tip : if you replace your memory, make sure you replace it with brand-on-brand memory. Much less chance of failiure on those.

I assume you know about the Soyo forum at amdmb.com?

http://www.amdmb.com/vb/forumdisplay.php?s=3db8221832bde9e4a752177366262f6f&forumid=25

How can I be sure it's not the motherboard?

You can't be sure, but should try to make sure it's nothing else first.

Doublecheck all jumper settings first..
Then check the following, as they can all affect stabillity:
CPU not overheating.
Northbridge not overheating.
CPU core voltage.
Memory voltage.
AGP card not overheating.
AGP card voltage.
Memory timings (Start with the least aggressive automatic timings in the BIOS, not optimised)

All of these measurements should be readily available from MBM (someone linked this earlier).
Something else you can try is Sisoft's Sandria utility. The freeware version is limited, but still useful in your case. You can run a burn-in test, and select which part of the system you are stressing, ie memory, CPU etc.

The amdmb foruma should be able to tell you whether people have gained stability by eg upping memory voltage. I know that on my Epox 8k7a, upping the memory voltage was necessary to gain stability when using optimised settings.

Oh, and what bios are you running. If the info I've gleaned from amdmb.com is correct, a recent bios fixed some compatibility issues with 512MB memory sticks that was causing stability problems.

Hmm. Okay. I'll find that software you mentioned too.

I hope nothing's overheating. I haven't been putting any stress on the system, and there's four fans on the case and plenty of room.

I haven't really made too many aggressive settings in the BIOS. The only thing I adjusted was the FSB to 133 so it would register as a 2100 processor.

So nervous...

Well, when I had my first problems with the system, I flashed it with the latest BIOS on the SOYO site. I dont' remember the version number, but I'll check that out when I get home too.

By the way, could bad cables attached to more simple things like the CD/DVD drives and floppy drives cause problems? Maybe all my cables were bad...

Also, the OS installed drivers for a lot of the hardware. If I installed the drivers that came with that hardware, could that make a difference?

For example, the items on the fancy cables are:
Toshiba DVD drive
HP SCSI CDRW (on 50 pin SCSI cable)
ZIP drive
Floppy

The items that I did not install drivers for myself:
DVD drive
CDRW
SCSI card for CDRW
ZIP drive

As suggested, I'll also try disconnecting all these drives and pray it's a problem with them and nothing else.

By the way, could bad cables attached to more simple things like the CD/DVD drives and floppy drives cause problems? Maybe all my cables were bad...

Yes, that is a possibility worth considering.

Also, I have seen stuff like this happen when running certain AGP graphics cards at 4x speed on a motherboard that doesn't properly support 4x (despite the fact that it says it does). I have sometimes had to throttle the card back to 2x to get the lockups to stop.

I have also seen bad USB-port drivers or IDE controller drivers do this sort of thing. Have you made sure to get the proper USB/IDE drivers for your motherboard from Soyo?

I would start by unplugging everything except the video card and one hard disk. See if it all works. If it does, start plugging other things in until it fails again.

Thanks! I'll try all of this.

I also just wanted to say that I agreed with the earlier posts which listed the RAM as a prime suspect.

Ill second that (RAM as culprti)...

Or, uh... well I guess its more like thirding that.
Fourthing?


Zachary

I was also told by the people on the SOYO board that heat would be a problem. They said I could check it in the BIOS. Where? And what are good temperatures?

And when we say "RAM", we can also mean things like having the motherboard running too fast for the RAM. For instance:

only thing I adjusted was the FSB to 133 so it would register as a 2100 processor.

You sure you got RAM that will really run at 133?

PC2700 had better run at 133FSB

The bios probably has a health check screen that will report temps and voltages.

CPU temps - 55 degrees C seemed to be borderline for my old T-Bird. It'd run forever at 54, but didn't like being taxed at 56.
Northbridge - 35 degrees C seems to be max.

Again, this is a different CPU on a different mobo, but it should give you an idea.

Oh, you will have much fun at this. I've been trying to figure my computers problems out for months now. Every time I think I have it solved, a few days later the crap begins. Replacing the video card got it to stop going black. Pulling out the old ram got it to quit rebooting itself. Now it just locks up, sometimes it can go days without doing it and some days it locks up a dozen times within mintues of coming back up. When I feel like it I'm going to replace the mobo next.

Hey guys...I ran memtest.

Is it bad that it runs for 7 hours, gets 7 passes and 4625 errors?

Is there a BIOS setting I might possibly change to make it work? I want to do anything I can before taking these sticks out.

By the way, I got MBM5 and it ran, but it didn't give me any temperature readings.

Is it bad that it runs for 7 hours, gets 7 passes and 4625 errors?

Unfortunatly, yes. That error number should be 0. Make sure to elliminate both individual memory dimms and the slots as the problem before calling it solved.

Is there a BIOS setting I might possibly change to make it work?

You might want to check settings relating to CAS latency. Should be a 3 digit string of numbers, like 3-2-2. Check the memory specs to see what it should be, and ensure it's set right.

So this was the RAM I got:
http://store.yahoo.com/pcmemory-stores/51pc33ddrsta.html

So that's the stuff. There's a sticker on the sticks that say tey have a warranty, so I suppose I can send them back. Who knows if they'll just replace it with the same or actually give me my money back.

Oh, I should point out that the RAM isn't even warm, so it's definitely not overheating.

You can get memtest to tell you where the errors occur. Run memtest once and write down those numbers (hopefully, it should only be a few different locations). Then shut the computer down and run it again. Make sure that the locations are the same. If they're not, I'd tend to suspect the motherboard.

I have a friend who's having this exact problem right now, and the locations of his errors are different every time, and the memory works fine in another computer.

Of course, I'm no expert here.

Um, it was a whole lot of locations. A WHOLE lot...

I still have to test the individual sticks, though. Unfortunately they say this:
* 1 year free replacement warranty

but they do say this:
* Special design for overclockers and 100% guaranty work

So if I can get the memory returned, I think I might go for reliability and speed over sheer size. I've been recommended this stuff, what do you think?

As I've said before, I strongly recommend buying brand on brand RAM, from known brands like infineon, samsung or apacer.

Make mine a fifth...Of Black label please.

Yea, it sounds like the RAM to me too, test each stick by themselves.

You didn't mention whether you'd read the forums at amdmb.com

One of the sticky threads there is "How to set up your Soyo Dragon", (paraphrased...). Amongst a lot of very good information that you definately should read is this gem...(again paraphrased)... "The Dragon is very picky about memory seating. Even if it appears as though the memory is seated correctly, it may not be. If you are having memory problems, then try reseating it."

Do you suppose that would go for the Pentium version as well? I'm planning to upgrade in a few months and I think I like that mobo very, very much.

Yes, I've been posting in that forum I saw that, and I really should try to reseat it. But I can't tell if it's seated properly. They sure don't look perfectly even as it is. I'm not sure why. I'll try it though.

Oh, and despite these problems, I'd definitely suggest the Dragon Ultra. It's a great motherboard even if the documentation is pretty bad.

By the way, I still have problems that I doubt are RAM related. For one, my CF reader.

The reader is in a add-on bay that came with the mobo. It has it's own connector that takes up one of the USB ports on the mobo. When I first got into windows and put a CF card in it, it saw the device and a new drive was detected. I was able to transfer files and all that. Then every time after that I couldn't get the thing to register. I put the card in and ejected it several times and still nothing was seen in my computer. Then I replaced my fancy cables with the ones that came with the mobo, and the first time I booted up, the reader was seen again, then the next time: same thing. This is puzzling to me. Any ideas?

Also, the cables did seem to be messing the system up. I replaced them and everything is running smoother, bootup was faster, and it seems like that alone let me play for much longer in GTA3! It lasted about 25 minutes before I got a program error, but at least it was an error, and didn't completely freeze the system.

One last thing. Microsoft Digital signature or whatever. How do I turn it off? I'm sick and tired of getting that annoying box every time I install new software.

I don't know. If you're talking about the Fire Dragon, and the connectors on it are from the same manufacturer, then it is a possiblity that you should be aware of. I'm sure that Tony could point you in the direction of an Intel-centric forum board that'd be able to answer this.

One last thing. Microsoft Digital signature or whatever. How do I turn it off?

Install Windows 2000 instead of XP?

Hmm. The first sentence of this thread is:

So yeah, I'm having trouble with my computer already. As stable as Win2K was on my old machine, for some reason it's a buggy mess on my new one.


I don't know whether you can turn this off in Win2k or not.. It doesn't bother me that much, but then again I'm paranoid

I thought he was referring to the XP license registration feature. Sorry.

No problem. Yeah, it's that digital signature thing. It's just a bit annoying that it pops up every time even a little piece of software wants to load. Oh well, I guess if it doesn't hurt the installations it's not too bad.

A quick update:

I ran memtest86 with each stick of RAM in the first slot individually. The first stick I tried got 3 errors. The second one got none! That one might be good, no?

The 4625 errors that I got before were because the test repeats, and I left it on for 7 hours.

I'm going to try with the sticks in the reversed slots of where they were before, make sure they're fully seated, and see what I get. But do you think that the one stick is okay then?? Is only 3 errors acceptable? The good news is that repeat testing brough up the same 3 errors, so as someone said earlier, it's not the motherboard, right?

Is only 3 errors acceptable?

Is only 3 crashes an hour acceptable?

It is not acceptable to have any errors.

If you're running Linux, however, there's a kernel patch available to mark certain address ranges as bad so that you can still use that memory. It's the badram patch, IIRC.

In fact, recent versions of memtest have a display option to show you the exact parameters you should supply to the kernel for that badram option.

I still believe that you should write down the addresses at which your memory has errors and then repeat the test to see if the addresses are the same. If they are, then it's almost definitely the memory's fault. BTW, I find that it's easier to record the badram-style report than the default report, so you might want to change the output to that mode.

I realize that you're making a point, but, since he asked the question in the first place, we can only assume that he is not as familiar with the concepts involved as we are. So, there is no direct correlation between number of memory errors and how often your computer will crash. Even one error could have dire results, if, for example, it were used when writing some registry data out and it ended up corrupting the registry. Your computer might not even crash, but the next time it did, for whatever reason (we are talking Windows, here, right ), it might be corrupted beyond bootability.

To sum up, memory errors are bad, very bad.

Errors are bad, mmmkay?

Okay, I'm glad I've got that straight. So in all likelyhood, the one that tested fine with no errors is a keeper, right? If so, that changes my plans. I'll send the bad one back with the intention of replacing it instead of returning it, just so that I can have matching sticks in there.

Thanks for all your help, you guys!

If you leave the bad one out, and the good one in (i.e., run with less overall RAM in the system), do all your crashes go away?

Well, with the good one in and the bad one out, everything seemed to run fine, but I did get one reboot. I then turned off that setting that was discussed earlier in the thread (auto reboot without error message), so if it happens again I may see if there's something else wrong. However, I played GTA3 last night for about 30 minutes at 1152x864 with absolutely no slowdown and a solid framerate. Again, the only thing that stopped it is a program error of some sort (couldn't tell), and not a lockup. I was able to exit it and get back to windows.

Oh, and I found the temperature readings in the BIOS. They were well below what was said here (I think the CPU was at 43).

One problem though. Last night I put in both sticks again, and when I woke up, there were ALOT more errors. By themselves, one stick had 3 errors and the other had none, but together there were about 30 that I could see. What does this mean?

It could be that the bad one is throwing the timing off on the good one, or it might mean that the good one is marginal. I'd try just getting a replacement for the bad one first and see what happens.

Also, assuming you installed Windows with that bad one in, I'd reinstall Windows, too. No telling what kind of odd errors it might have introduced.

crap. you're right though who knows, that might solve the problem. plus, if I just try to get one of them replace for now, it means my computer won't be out of commision at all.

Well, under the advice of someone on that SOYO board, I'm running Prime95's "torture test." I've also been surfing a lot and opening applications like Outlook Express and have been doing this all for about 4 hours. So far there hasn't been a single glitch, and all the tests in Prime95 have passed. Not one restart (except for the one that the thunderstorm caused, but we can't blame that on the RAM), and not one hiccup. Rock solid is how I'd describe it.

But I don't think I'm too stable now

Hi Dignan17.

Seems like noone else noticed this, but you made the 100,000th post to the BBS. I would really like to give you some sort of award, but I can't think of one that would be nice to have (for you) and affordable (for me). Maybe Rob will find something to give to you.

cu,
sven