hardware indicated random fatal error...

OK, this is a switch from a fatal error problem that started over on the software forum. Originally it was thought that the random crash was the result of a corrupted driver because the various Microsoft crash analysis' blamed 13 of the crashes on a software driver. The objective of the software forum was to access the Minidump files to determine what driver was the corrupted device.

Finally accessing the minidump files showed that each driver blamed crash involved a different driver. The overall indication is a probable hardware caused problem. Some other crash analysis documents indicated corruption of a RAM file reported to Windows by the CPU, Which to me indicates probable RAM defect, but, so far I have been unable to run the program DocMem2_2, a recommended Memory Diagnostic. It does not seem to want to run on my system. I ran Memtest86, and Windows Memory Diagnostic with no negative results, but do not feel this eliminates the RAM as the Suspect problem. I left a message with the Simmtest website to see if they have an idea why their program will not run on my system, while I look for a more comprehensive program than I have run on the Memory so far.

Any suggestions or input would be appreciated, for those who are interested in the final outcome of this long and tedius investigation. I have been chasing this problem for five months with little effect so far. Thanx for your input.

 

I have seen that happen before too, a friend of mine had the random blue screens then crashes, and Microsoft Error Reporting said it was a different driver each time. Somehow or another it actually ended up being ATi's driver software on his particular system. I'm not sure exactly what happened, but we formatted his pc and it worked perfectly after that. This is just my particular experience, but I've seen it happen before were it was also caused by a faulty motherboard too. So don't rule out hardware necessarily, but don't limit yourself to calling it a hardware problem either. Check your drivers too, sometimes drivers fail and the hardware is still fine.

 

When i get driver that i can't identify playing up, i disable all that i can, reboot, and bring them into play, one at a time. If you do identify a screwy driver, disable them again, but leave the problem driver enabled. If it still plays up, then, in one move, you have removed the possibilty of any of the disabled drivers being in conflict. One thing i did in an act of madness once, i deleted every driver i had, and i mean everything, even system. Windows auto detected them on reboot, and reloaded new ones. To do that you need all the driver disks, and you do run the risk of a format and reload of windows.

If it's memory, and you have more than one stick, take them out, one at a time, and see if that makes a diference, swap them around a little. If it doesn't then rule out the memory. To have one go happens, to have them all go at the same time is very rare indeed.

Is this a problem since a reloaded system? Have you checked for a scratched windows disk?

Have you tried Scandisk, set up for bad sectors?

If it isn't a reloaded system, then have you thought about registry or system file corruption?

Have you thought about viruses or the likes?

Have you run DR Watson, try it and post the results in the thread, it runs in the background and adds a synopsis to the error messeges.

If your using memory intensive programs, then chances are the swap file has been used, have you checked for a corrupt swap file?

Have you got the system set up for complete Memory Dump recording (under 2gb of memory only). If so, do you have the swap file more than a couple of mb larger than the physical ram. You get the full inventory of the memory recorded to log, at the point of the crash.

Stop believing what Windows reports is the problem, it reports them at a time when Windows itself is unstable, and often gets it wrong.

Only a few ideas, hope it can help.

 

Enough to hunt it down I bet but if you are very unlucky you might need even better memory/bios setup tester than Memtest86. www.mersenne.org have Prime95 one of the toughest cpu/mem tester there is. Almost industry standard, like Memtest86, for good reason. Not so hard passing Memtest86 and still fail P95, if it matters or not is other question. A stable computer can run it forever. P95 does not replace Memtest86, different tasks and both needed.

How you test btw? The longer the better, try 4-6 hours as min. Read "stress.txt" from P95 download, says all about stability. Takes forever but Ive found that is I set up a "custom" test I can trigger ram-timing/voltage/tweak problems faster than usual. Set both values to 4096 and tell it to use half of system memory so computer is still usuable. As with Memtest86 a good idea to let computer work on it while you sleep. If you find it fails within seconds or few minutes you have either found source of problem, or you have one more to fight.

You should also have good eye on temperatures of all sensors, same with voltage levels. Everest is good for this, http://www.majorgeeks.com/download.php?det=4181 btw, without knowing much about memory dumps Im pretty sure you get better help sending a complete Everest report. Give good detailed picture of computer. So detailed you must edit what to include or you will reveal personal info. Notice voltage levels before and after you have fired up Prime95. Unstable power supply or motherboard part dealing with power you can compare with bad memory, anything can happen. If it turns out motherboard is half-dead in some areas you can run in circles forever - only way to track it down is to exclude. Easier if you have computer no. 2 so you can check parts elsewhere.

Well as a crazy overclocker I know about Windows error messages and yes very true dont trust them for 2 seconds but often they can be translated, for example with help of http://aumha.org/win5/kberrmsg.htm Same info is at MS. Just a word or code can be enough. May be you have that piece of hardware, out of 100s, given a known problem with X driver - MS have all that info but you need min. 1 keyword to focus.

 

Ok, just to cover a few quick points that you guys should know, about my situation. The system is brand new as of February from SysteMax, and the crash came with the new computer. No downloads before the first crash on the first day.

This crash is from the original hardware/software combination. All original drivers were reinstalled, and the driver updates SysteMax has issued since the system was built were downloaded and installed incase they were the problem. Three complete reinstalls of the factory OS were done, with no correction for the problem. accessing the minidumps was finally the last point that pretty much eliminated the drivers as the culprit since the suspected cause was a corrupted driver, and all minidumps credited a different driver.

Another fact was the Microsoft crash analysis stated a hardware problem four out of 26 crashes. One of the hardware causes stated that the CPU notified Windows of a Memory file that was corrupted and the system needed to close.

Since the driver seemed to not be the problem, suspect 2 was the Memory. As of today, I started out at Six AM and was running Memtest on the OS in the background to see if anything surfaced. Till about 1045AM No errors were recorded. Then I switched back to Memtest 86 and ran it from 1045AM until 1125PM, for a 12 and a half hour run of 105 passes, again with no errors.

I am going to go out on a limb here and say the RAM is probably no longer suspect though these tests are rarely conclusive. With no Extra sticks to substitute, this is the best I can do for now.

So, two down, we think, so what is the next most likely candidate for a possible suspected cause, and how do we test?

I think that pretty much covers everything. Any questions or suggestions?

 

No offense man but what kinda sh*t is SysteMax trying to pull? A straight-out crash as soon as you open the box? Return it and buy from someone better.

 

Wow, lots to do here. Will check out Prime95, and read up on Everest, then get back to you. Thanks a bunch.

 

I do wish it were that simple Oh-man, but it's alot easier to buy a defective computer, than it is to return it. You know what Customer support is with this system?

"Gotta problem mac? Mail it to us and we will look at it and if we find something wrong we will fix it and send it back to you. In the meantime, sit on your hands and wait.

 

Chuck the thing back at the suppliers, it came with a fault, wether software or mechanical. Any messing around inside could invalidate the warranty. If you've been inside, then check the BIOS won't tell on you. Reload to factory specs, and just throw the thing back.

Is that an option?

 

Well good testing with Memtest86 but I think you need to get on top of tempereatures and voltage readouts. Dig into Everest, you should have that installed anyway - most love it and it wont do any harm.

Then may be some hours of Prime95 but allready now you have done more than most users would. Look at it that way and send it back within 1 week max. You could tear computer apart, put it back together again - hoping for miracle but then they might notice. You can/should check cables, wires, connection, cpu fan but that should be it.

If memtest86 and Prime95 can run for many hours those tests are conclusive - as much as is possible. If temperatures and voltage also look normal during Prime95 you wont get random crashes due to powersupply, cpu, memory. Motherboard can still be faulty though. Hard to test motherboard, a little almost invisible scratch can be enough!

Yep, if shop is any good they will take it back in a flash. But I see your problem, I would also like to know what is wrong if it was my computer. Even so, settle for some cable-fiddling/check inside case, memtest86/prime95, temperatures/voltage levels and dont forget a safe boot XP session. You still get lockups in safe mode there must be hardware error. If you can find it or not should not be of importance once you have done what can be expected, or more... What about XP event viewer?, anything in log-file of interest?. XP does not allways pick up crashes but can be helpful. Punch in "eventvwr.exe" in run box or find it in Control Panel, Administration somewhere.

Those people who will unpack your returned computer will not do all this - they just send you new computer. Time is money.

Look on the bright side, you know more about how to check computer than you did before. Should do that with new one even if it seems ok.

 

Yes, I do agree Brambo, I do know more now than I would have if I had gotten a good computer first off, but I do wish I could locate the problem. I have checked the cables first thing, both inside and out of the case. I could possibly send it back and get a refund, if I demand it and use the state lemon law as leverage, but then, that leaves me without a computer for as long as it takes to get a new one.

The bottom line is I paid for a computer that works, and I expect them to give me one. Their Idea is that I do without a computer until they get it fixed, they already told me they would fix it, not replace it. If they would replace it they could send it to me first and then I could box up original and never be without one. They say they don't do it that way.

What I would like to do is find the problem, and make them send me the replacement parts, and I will install them. However, the deeper the problem, the harder it will be for me to repair myself at my skill level. That is why I was hoping it would be the ram, I can do that With my eyes closed.

As far as the event viewer goes, I do get some errors, usually when my AVG starts running. Other than that I don't seem to be showing any problems recently. At first I seemed to have a number of things that triggered Viewer notations, but other than AVG, no other entries under Applications or System, that I have noted since my last FActory restore.

Just so you know though, I have documented every one of the last 26 crashes once I figured out how to do it, and kept both hard copies and digital on the system. If push comes to shove and they don't want to stand behind their product, I have enough to file a lawsuit if that becomes nesissary. I Hope it does not, because I hate dealing wiht courts. So much paperwork, Ya Know?

All I want is my machine fixed right. My whole objection to sending it back is losing internet access. Even for a month is too long when you paid for a new computer with the expectation that it would give you years of uninterupted sevice before you had to do anything major to it.

This is the first time I have turned my computer on since I left you guys the last message so I haven't had time to look at what you recommended to me so will do that tomarrow. I will talk to you then.

Thanks for the help.

 

Hi, guys, I am about out of time to dedicate to this today, so thought I would update you so far. To Mushroomcloud, I did create a couple of shortcuts, one to Dr. Watson and one to DR. Watson32. if I execute the fist I just get a message of no errors found, and the other brought up the setup menu on the error reporting so I changed it to the Full dump setting while I had it up. I actually did that before, but after I did one of the restores, I didn't reset it. So, I will need some more info on what I have to do to get it to do some of the things you describe. Not really familiar with the program, or even what it is suppose to do.

Scandisk, I couldn't find, but I think I had it on my system once, though I probably lost it in the restore process. That is one of the system tools on the windows disk, as I recall, is it not? On inside the case, I have been inside the case, and have installed some upgraded hardware. A new second slave harddrive, and a DVD ROM and CD Burner inplace of the CD ROM, compliments of my Sister and brother-in-law. All new items are using Windows drivers and software so compatibility should not be a problem. The CD buner is a Yamaha, and is no longer supported by Yamaha, so it is just rudimentary backup file use. Windows is good enough for my purposes for now.

Systemax has an area of the original disk that is not accessable to the user, and and is described as "Unallocated" This probably has something to do with recording system processes so they can see what the system has been doing. Call me paranoid if you want, but I would bet on it having records of system changes all of the way back to when the system was first activated. It is the only reason I can think of for Disk Management to show a space of "Unallocated" space I have no access to.

Now, Brambo, downloaded Prime95, and Everest. Had Everest before but took it off as it was like a program I had called Motherboard Monitor, except seemed at the time to have less things on it. Might be I just wasn't looking at everything.

Temps are as follows: Motherboard,13 degrees C to 66 degrees C, depending on the sensor it is reading. CPU is 32 degrees C, AUX is 40 degrees C, Harddrive #1 is 24 degrees, and harddrive #2 is 37 degrees.

volts are as follows: CPU core, 1.39V; 2.5V=3.36V; 3.3V=2.99V;5V=3.66V; 12V=11.9V; 5V Standby=3.06V; VBat=2.03V.

Haven't run Prime95 yet, but will get to it.

Basically I think you will agree as a overclocker the temps are pretty much enviable, since the ones you overclock probably don't see anything this cool ever again. The voltages I really can't comment on as I have no real knowledge there. I really don't see anything that I don't find respectable, do you?

Oh, also bookmarked the windows error page you posted. In case I need it. Once I run Prime95, will post the results. Anything I missed so far?

Oh, I forgot. On safe mode. I use a dialup modem so have no access to the internet when in safe. So there are some things I just can't do there. Let me know what I can do and I will.

Well, guess this is a lot of stuff here and I will be lucky if the server software doesn't dump it when I upload it so catch you later. Thanks for working with me on this. I expect it will be a long and tedious exercise in the process of elimination.

 

Just to keep you guys updated I have started running Prime95, It is runnig as I write. It is 53% through the tests, and so far nothing has shown any problem. However, My system fatal error count has climbed one more to 27 recorded events. Add another 10 or 12 for before I figured out how to keep records of them.

Last night just before I started Prime, I was scanning a document, and my scanner wouldn't scan. I rebooted to try and get it back online, and it finally got the job done, but was immediately followed by an error message that the scanner software had to close and I should report the problem to microsoft. I fired up my AOL software, and bingo, it crashed before I could get online.

When it came back online I had another error message to send. the alleged cause of the crash, AOL software. Strike one up for Mushroomcloud, it blames server software, when the system was already malfunctioning before the server software was involved.

It is really irritating to have paid for so meny problems all at once. I think for the rest of my life I will use Systemax as the term most descriptive of time and money wasted. I have Systemaxed five months away. Has a nice ring to it, don't you think?

By the way, I have told them they need to come over here and get involved in this troubleshooting if they want to save the reputation of their product. We will see if they really give a **** about the people they sell to I guess. Well, gotta run. let you know what the outcome is.

 

Well, I did run Prime95, and now I have questions about it. It appears to have run flawlessly for the duration, about 13 hours. Well, actually one flaw, it completed the test while I was away, and when I came back, the results were not there. The program appearantly just dumped the final information and continued to try to start a new run of the program, but could not access the internet so never started.

I decided to look at the program to see if there are temp files it uses to store information, but didn't find any. I did however find where the program was set to a memory limit of 8 MB of memory to perform the calculations. That seems to defeat the purpose of a test program to check your systems stability. I would think that I would want to bump the memory usage up to max in order to work the system as hard as possible. So, should I change the memory usage?

Secondly, as near as I can figure out, this program is a prime number algorythm that calculates various prime numbers, basicly putting high stress on you computers electronics to see what goes. Probably not originally designed for that purpose but serves well. What is it that I am missing here? You guys did say something about the results, but I don't even know whre mine went.

Wish I could give you more information, but I wasn't here when it finished. So, what can you tell me? Thanx.

 

Well, I just got an e-mail from Systemax, and they deny I have ever mentioned this problem before. Big surprise. They don't seem to want to take a proactive response and have their techs input here. Guess they don't want people seeing their systems have problems. So, what is the deal guys,
About Prime95? I could use some further input. Got anything that will give me a little more insight?

 

So, What is the deal here, guys? You aren't afraid of posting some negative publicity for a company like Systemax are you? A company that is basicly selling junk that they don't want to fix once you have it? You have a voice here, a very powerful voice that is capable of doing what is nessisary to protect the public from being predatorized by the kinds of people who think the customers aren't important once they have your money in their pocket. So, why don't you use it?

You don't want your friends being ripped off by OEM's that don't stand behind what they sell, do you? This is really important and you can make a difference. If you don't get involved what are you going to tell the next guy that say's he bought one of these machines, especially if it is one of your friends or family?

All you have to do is do what you normally do and that is talk about the troubleshooting of the system. If you don't talk, you are giving companies like Systemax license to sell defective merchandise and make money off of you then ignore you when you have a problem with their product?

I find it hard to believe you people are afraid of these guys because they have money. They do this because people don't complain, and that is all the permission they need. So, come on, work with me here!

 

If you make a big stink to their tech-support and return it then they'll have to spend money and effort on you, I highly doubt the guys at systemax are surfing majorgeeks.com

IMO .. it's probably either combination of hardware that's faulty or like one capacitor that's blown ... if it's the latter and you're handy with a soldering iron well then go for it ... but I think it would be best to just return it, call, don't email, systemax and talk to a living breathing person, and tell them you've been trying to fix this for more than 5 months and persist untill they take it back

You still have the reciept, right?

 

OK, thanks. It has come down to the point that I seem to have no choice but to either sue them or make them take it back even though I really hate having to give up my access to the net. This means that I have to go through the whole process of buying a new computer and everything. Scince AOL is the one responsible for marketing this system maybe I can pressure them to put some heat on Systemax to replace it first. I've just got this thing set up the way I like it and it really Irritates me to have to return it and start over. I appreciate all of the help you have tried to give me and will continue to read and participate when I can. To any who have thought about getting a system from these guys, I have one thing to say, "DON'T!!!"