New System Troubleshooting: Help :(

S

SelRahc

Limp Gawd
Joined
Aug 12, 2004
Messages
389
Hi all,

I just got done completing my 2 identical builds for project stackers. However, one of the computers seems to be having an issue. Naturally its mine >.<;

Here are the specs:
AND FX-53 939
MSI Neo2 Plat mb
512x2 Crucial Ballistix PC-4K
Antec Neopower 480
ATI X800XTPE (real)
Sound Blaster A2 ZS
XP All patches with the exception to sp2

Everything is running STOCK, no OCing.

Here my problem:

Reboots. This is the biggest problem now. The computer will just up and reboot on its own, as if I hit the reset button. It does it when in games, it does it when im just browsing file folders. Happens several times a day.

What I've done:

I ran MemTester, and it reports an error every 100 cycles or so. I contacted Crucial and they told me to get DocMem and test with that. I did get it.
DocMem ( from http:\\www.simmtester.com ) is free and is a bootable tester. It does many kinds of tests. It ran over 22 hours burnin looping with not 1 error. So, according to crucial, my memory is really ok.

I loaded MSI PCAlert 4 to monitor MB temps. As soon as it initialized, it was giving me undervoltage warning on my 12v+ rail to the MB. (the 4pin square one). It was giving me a reading of 1.98v, which didn't seem right at all. Would the system even boot if it were that low? When I reboot and check the volt in the bios itself, it is well within target @ 11.70v. I watched it for over 10 minutes and it never dipped below 11.69 and never higher than 11.72 so I don't think that is the problem either? I think PCAlert doesn't seem to be reading it right... I think?

Here are my specs when I check them in the MB Bios:

Temp CPU: 56c
Temp System: 22c
Vcore: 1.47v
3.3v: 3.07
+5v: 4.95
+12v: 11.70 * PCAlert4 says its @ 1.98, bios says its stable @ 11.70
Fan speeds seem within spec. CPU Fan is at 6K.

CPU Clock: 2412
DDR Clock: 200

MB Bios 1.1

Video Card drivers are the 4.9 cat drivers only, no control center. The GPU has been very stable since I did this without the CCC.

I am going to reformat again, just to make sure I didn't get a buggey, and I need to do it again for testing anyways.

What the heck could be going on? It's driving me nuts.
 
Someone mentioned memory conflicts...... Could the Ballistix be causing a memory conflict that is crashing my system? So the memory may not be bad, but it may not be playing with my MB? Would that cause reboots like that???

I need to add though, that the other stacker is identical and is not having these problems.
 
is auto restart disabled?

control panel -> system -> advanced -> Startup & Recovery -> Settings -> uncheck "Automatic Restart"

I think that's how you get there...
 
Disabling auto-restard is the first thing that I would do. With a problem like you're experiencing, the PSU or RAM is almost always to blame but you can make a good guess at which one it is if you disable auto-restart, write down the error code, and google it.
 
K, I'll do that.

I also downloaded Prime 95, and it haults the torture test after 1 minute with an error >.<
 
I took a look in the event log, and I am really leaning towards the memroy as being the problem.

Take a look at this. Each of these errors represent a point where the system crashed today:

Faulting application wmplayer.exe, version 10.0.0.3646, faulting module wmvcore.dll, version 10.0.0.3646, fault address 0x0004c56d.

Faulting application iexplore.exe, version 6.0.2800.1106, faulting module wmvcore.dll, version 10.0.0.3646, fault address 0x0004c56d.

Faulting application quicktimeplayer.exe, version 6.5.1.17, faulting module unknown, version 0.0.0.0, fault address 0x00d48870.

Notice that the top 2 are at the same memory address..... that is the physical memory address right?

Also, prime errored in blend mode. I'm going to try small fft's now and see if that errors as well.......
 
Yeah, it's a memory problem. Also, you were right about the hex. That is the physical memory address where the error occurred.
 
Prime's small FFT torcher test is working smoothly, so im fairly sure the cpu is okey dokey.

My problem now becomes RMAing the memory. Crucial says use DocMem to test it, and DocMem says its ok. When I ran MemTest, I get errors. What memory test program should i go by, and what does everyone else use to test with?
 
lie? you paid good money for it and it doesnt work the way they advertised your american, get your money back! or replacement either way is good :-p
 
Lie to them and tell them it was passing DocMem but now it won't. Docmem isn't much liked IMHO. Memtest 86 not only tests RAM but CPU and Chipset usage as well. It's a great tool for testing any kind of memory issues.

One thing I recommend, when you do call Crucial back INSIST on an Advanced RMA, where you give them a credit card number to hold. They ship you replacement RAM and you ship them your RAM back in the replacement RAM's package. This will decrease downtime and you're guaranteed to have a different stick (or sticks) in hand. If you send it to them first, then your'e without RAM and they may just send it back to you.

If they say they can't do that, ask to talk to a supervisor. There is no reason why they can't, I've done it several times. This trick also works with hard drives, Dell's Axims and DJs and occasionally a monitor. Then again I don't tolerate phone customer service reps very well. (evil grin)
 
Morphes said:
lie? you paid good money for it and it doesnt work the way they advertised your american, get your money back! or replacement either way is good :-p
Click to expand...

Haha. One way or another ill get new memory. ;)

The Memtest I was running wasn't memtest86, it was some other windows memtest. I have since gotten the correct version, Memtest 86, and ran the test from the floppy.

OMG the errors. It wasnt running 5min before the first errors... at the exact same address. So I have no doubts I have a bad simm. I'm going to figure out which one it is later today and get it replaced.

Hopefully that will fix my problem, I really need to get on with the testing heh.

By the way Morph, EastCoastLan rox. I just moved from VB alittle over a month ago :)
 
The_Mage18 said:
Lie to them and tell them it was passing DocMem but now it won't. Docmem isn't much liked IMHO. Memtest 86 not only tests RAM but CPU and Chipset usage as well. It's a great tool for testing any kind of memory issues.

One thing I recommend, when you do call Crucial back INSIST on an Advanced RMA, where you give them a credit card number to hold. They ship you replacement RAM and you ship them your RAM back in the replacement RAM's package. This will decrease downtime and you're guaranteed to have a different stick (or sticks) in hand. If you send it to them first, then your'e without RAM and they may just send it back to you.

If they say they can't do that, ask to talk to a supervisor. There is no reason why they can't, I've done it several times. This trick also works with hard drives, Dell's Axims and DJs and occasionally a monitor. Then again I don't tolerate phone customer service reps very well. (evil grin)
Click to expand...

the problem comes when the advanced RMA replacement is deffective :p

orelse you live in the EU, where you have 7 work days to return anything without having to give ANY motiv. VERY good for flat screens and dead pixels.
 
Ill stray from the norm...

Discount all wires going to your reset switch from the motherboard to the case. Those filmsy wires cut real easy and you could have a simple short.
 
Here's a thought. Why don't you try to swap RAM before going through the RMA process. You said you built anohter system, correct? Try swapping memory with that box and see if the problems stop on your machine and start with the other one. This is probablly the best way to test the quality of RAM. Also, have you ever tried re-installing windows? I know this is a bit of a tangent, but incredibly effective for ruling out software glitches. You also don't have to deal with useless support reps. Good luck.

P.S. Are you sure you have the right drivers for your mobo? If you go the windows reinstall route, be sure to just use the drivers from the CD's for both your mobo and video card. While this may not be your preferred configuration, it will at least show whether or not your barebones config is in proper working order.
 
I did that and it seemed to clear up some of my problems. I guess this is what we get for being bleeding-edge. Sometimes, I feel like we're the real guinea pigs. At least try and swap your sticks for a different set, like mentioned above. That way you can be for sure. Also, try to read up here: www.houseofhelp.com ( I think). Google it. It comes from Corsairs site but it would most likely apply. Some folks have fixed this reboot issue by relaxing timings, upping voltage, or manually setting timings. Good luck.
 
SelRahc said:
I loaded MSI PCAlert 4 to monitor MB temps. As soon as it initialized, it was giving me undervoltage warning on my 12v+ rail to the MB. (the 4pin square one). It was giving me a reading of 1.98v, which didn't seem right at all. Would the system even boot if it were that low
Click to expand...

you probably found this out already, but... this is a a know issue, also you could try updating the BIOS to 1.2

for a more accurate voltage try MBM5 (motherboard monitor 5) or check the included utility on the CD (corecenter) :)
 
besides being an obvious memory problem, there a million + 1 people saying they solved a similar issue by switching keyboards... to a usb keyboard... just throwing that in there
 
SelRahc said:
Prime's small FFT torcher test is working smoothly, so im fairly sure the cpu is okey dokey.

My problem now becomes RMAing the memory. Crucial says use DocMem to test it, and DocMem says its ok. When I ran MemTest, I get errors. What memory test program should i go by, and what does everyone else use to test with?
Click to expand...
Call Crucial, and tell them the truth. It will pass DocMem, but not Memtest86+ and tell them you're not comfortable with that. My guess is that they make it right; there's a reason they are one of the best in the business.

Before you do that, repeating the obvious is probably worthwhile;
- Latest MoBo bios installed? It makes a difference.
- Memory Vcc / timing parameters correct in the BIOS?

Good luck - B.B.S.
 
Questions for you:

1. Have you run MBM (or similar) and watched the voltage readings while having a load going?
2. Have you verified the mating between the cpu and cooler?
3. Examined the seating on all of the power contacts to make sure that no contact has come out of it's socket (especially the primary power connector for the MB)?

Let me know!
 
Random reboots are normally only caused by faulty ram or pulling to much power from the psu. If replacing the ram doesn't do the trick, swap out that power supply. Oh, and your readings in the bios are not true readings for voltage lines, as most of the time not everything in your system is on. Whereas when your in your OS they are, unless you have things set to turn off after a predetermined amount of time (right click desktop/properties/screen saver/power).
 
I learned how important a high quality PSU is this year. I was having all kinds of problems with some kind of generic japotech PSU. I finally got my act together and bought a thermaltake PSU. It was definitely higher quality. All of the cables were sleeved and the rails had much higher ratings and tolerances.

It definitely sounds like you've uncovered that its a memory problem, which hopefully you'll be able to resolve since Crucial is a great company.

Also, you're specs look bitchin' :p .
It'd be a shame if a crappy PSU ruined the whole machine. Don't forget how important a good psu is.
 
Just another thing to consider...

A lot of people with Socket 939 Abit boards have had major memory compatability issues with the 939 chips. Since the mem controller is on the chip as opposed to the board its possible that any socket 939 might echibit these issues. Additionally, the Antec PSU when used with the Abit AV8 causes random reboots and simply will not work with that board. It is possible that the MSI board is exhibiting similar issues.

Warden
 
Ive had to trouble shoot a memory prob also on a NF7-S board. It was a pain in the ass to do. I would do 2 things. 1st test ur memory in another board, and test other memory in ur board. Try different the other 2 slots for the memory in ur system also. Also I woudl try just running 1 stick of memory in each slot and get it to pass memtest86.. great program. If your board passes with the other sticks of memory, then you can test each of ur sticks to find which one is givin your the problems. With the nf7 i was trouble shooting, I found that it was acutally the mem slot on teh board that ws givin me the trouble. It was slot 3.. since I was running dual channel in it. and 1 and 3 had mem in it.

I might also set the memory timings in the bios urself incase your mobo isnt reading the memory correctly, and set the mem voltage to the correct voltage from crucial. If there is a prob with your ram, crucial is a great company. and they will RMA it. GL
 
I recommend run the memory tests with one or the other stick installed. I bet you have one stick with a problem. In fact, i'll bet ya a case of Guinness :)
 
I had a similar problem, random reboots, no heat problems, strange behavior etc. I replaced all hardware except memory. I found that the problem was that in BIOS if you set the memory configuration to AUTO, it can be a little over-aggressive in its timings. Try manually setting the Mem timings to the least aaggressive they can be per your mobo, and go from there. Also keep an eye on any settings that say memory speed: normal/fast/turbo. As before, start with the least aggresive and go from there.
 
This will sound crazy but if you go into MSCONFIG and on the Startup tab you should see a entry that has:

/L:ENG

uncheck that and reboot the PC. It was causing my machine to do that and after that it has been working fine.
 
An odd twist.

When the memory is tested together, it errors alot. However, when the memory is tested individually, 6 hours each and no error. When they are put together again, errors galore within minutes.

So it seems the memory is erroring when used together. But when used by themselves, there is no problem at all. Dule channel failing?

So now I have a few questions:

I have 2 channels on the mb. It says Chann A is green, and Chann B is purple. When I use them together, I have 1 in a green and 1 in a purple. Is this how it's suppose to be? Is that concidered dule channel?

What would happen is i ran them both on A? Or should I even do that?

I'm going to dig around alittle and see what I can come up with, cause i'm running out of ideas here.

As for the voltages, I downloaded Motherboard Monitor 5 and it is reporting the voltages as normal, so i think PCAlert was having an issue.
 
SelRahc said:
So now I have a few questions:

I have 2 channels on the mb. It says Chann A is green, and Chann B is purple. When I use them together, I have 1 in a green and 1 in a purple. Is this how it's suppose to be? Is that concidered dule channel?

What would happen is i ran them both on A? Or should I even do that?

I'm going to dig around alittle and see what I can come up with, cause i'm running out of ideas here.

As for the voltages, I downloaded Motherboard Monitor 5 and it is reporting the voltages as normal, so i think PCAlert was having an issue.
Click to expand...

You need to run your both your ram in the same color. That would be dual channel, not one in one color and the other in another.
 
I agree with Turbonium... throttling back on the memory timing settings may do the trick for you. When I put together my AMD 64 3400+ /ASUS K8V DELUXE/mushkin PC3200 system, I was experiencing all kinds of errors running MemTest...after researching the mushkin web page, I found this article they put out about setting the memory timings for their memory & AMD 64's:

www.mushkin.com/epages/Mushkin.storefront/415dceb00035c8e02740c0a801020643/UserTemplate/27

For my AMD64, once I set "tRCD" back one setting (in my case it was from 2 to 3 ) all my problems with memtest went away.
 
Random Digits said:
You need to run your both your ram in the same color. That would be dual channel, not one in one color and the other in another.
Click to expand...

This man is correct!!!
 
Pied said:
This man is correct!!!
Click to expand...

Actually, he is incorrect.

1 in each color is Dule Channel as I thought. This is confirmed by CPU-Z and MemTest86. When they are both put in the same color, it is and shows up as single channel mode.

The layout of the mb is:
Purple, Green, Purple,Green. With green being CHA and purple being CHB.

When I placed them both in the same channel (single channel) the tests do work with no errors. So it's Dule channel that is failing for some reason.
 
I forget the stats of the other system, but is it possible to put the bad sticks in the other box, and run memtest on that? This would help eliminate motherboard incompatibility. If neither set of ram works in the msi 939 board, you have an entirely different problem on your hands. It would also get weirder if the bad ram works perfectly in the other system. It would, however, help you track down the source of the problems.

Also, if you're really good, you could buy a good voltmeter and use THAT to check the output voltage of the psu, just to make sure you're not undervolting something.

That was my .02 for the situation...

J
 
SelRahc said:
Actually, he is incorrect.

1 in each color is Dule Channel as I thought. This is confirmed by CPU-Z and MemTest86. When they are both put in the same color, it is and shows up as single channel mode.

The layout of the mb is:
Purple, Green, Purple,Green. With green being CHA and purple being CHB.

When I placed them both in the same channel (single channel) the tests do work with no errors. So it's Dule channel that is failing for some reason.
Click to expand...
Look at what i posted, then look at what you just posted. I think you'll find your error.

And you guys are making this more problematic than it needs to be.
 
Ya, I'm going to flipflop the memory into the other system tonight, but I'm thinking its no the memory now. Don't know though. They test fine by themselves, and together when in single channel mode.

The other system is running identical parts, ram included, and hers runs and tests flawlessly in dule channel mode.

Checking the psu voltage is also on my last for later if its needed.
 
Someone suggested trying the memory at it's raited voltage of 2.8v. I did that but still got errors, so I reset it.

Well, I don't know what happened, but when I set my memory to 2.8v now I am getting graphical glitches on screen. Even reset to defaults, the glitches are still there. Its like a refresh line that appears at random on the screen... I just don't get it. Nothing was changed with the video... it even does it while im in the bios screen now.....

The memory is raited for 2.8v so that isn't my concern. Would the memory cause these flickers? I really think this board may be bad...

EDIT: Even after shorting out the bios reset, I got flickers now. It's official... something is seriously messed up with this stupid system >.<

I'm going to play with this alittle more, but If i cant get it working I am going to shut down and install everything on the Asus AV8 board. If everything works fine on the AV8, I'm going to RMA the MSI board. If it doesn't, then I'll just hafta go back to the drawing board......
 
Ok so,

I swapped the memory from the other system into this system.. and guess what....

Workes flawlessly. No problems whatsoever. No errors... no glitching... no crashing. Memtest86 reports perfectly working memory in dule channel mode.

When I put my memory back into the system, the problems return. Errors in Memtest86, glitching and crashes.

So...... friggin hell. Back to the memory being the problem? I'm going to RMA the memory on monday and ask for advance replacement. Hopefully they will let me do that and pay for overnight shipping.

This is so frustrating.
 
SelRahc said:
Ok so,

I swapped the memory from the other system into this system.. and guess what....

Workes flawlessly. No problems whatsoever. No errors... no glitching... no crashing. Memtest86 reports perfectly working memory in dule channel mode.

When I put my memory back into the system, the problems return. Errors in Memtest86, glitching and crashes.

So...... friggin hell. Back to the memory being the problem. I'm going to RMA the memory on monday and ask for advance replacement. Hopefully they will let me do that and pay for overnight shipping.

This is so frustrating.
Click to expand...
Your forgetting what happened earlier with your video problems. Once you take tha into account, it doesn't look like ram at all, it makes it look like your mobo. Test what you think is your "problem" ram into your "good" system and see if it checks ok. If it checks good then its your mobo not the ram. BTW its dual not dule.
 
Random Digits said:
Your forgetting what happened earlier with your video problems. Once you take tha into account, it doesn't look like ram at all, it makes it look like your mobo. Test what you think is your "problem" ram into your "good" system and see if it checks ok. If it checks good then its your mobo not the ram. BTW its dual not dule.
Click to expand...

Well, the memory does the same thing in the other system. It fails memtest. So there may be a ram issue.

Also, the video glitching came back again....

EDIT: While i was typing this it just flickered real fast and reset, as if i changed resolutions.... bad... bad.. bad.... Wonder if that may be my monitor having problems.....

I'm at my witts end. I may just RMA the MB and the memory and say screw it. Hell I should just RMA the entire computer. I'm so over this.... :(
 
You must log in or register to reply here.
Back
Top