Computer crashes completely, but no signs and no BSOD

Status
Not open for further replies.

JBauerisGod

Posts: 15   +0
I built my computer back in late '05 so it's getting on up there in age. Here are some specs:
OS: Windows XP Home
Motherboard: MSI K8N Neo4F
Processor: AMD Athlon 64+ 3000
RAM: Corsair ValueSelect (2x512)
GPU: nVidia GeForce 6600GT
PSU: Thermaltake 500W

I just bought the PSU the other day because the problem has been going on for a couple of weeks now. It seemed like a PSU problem to me because I can be doing anything from playing games to surfing the web and the computer will shut down completely. It's almost like the power went out.

Afterwards, the power light on the front flashes on and off until I turn off the switch on the PSU. I've tried disabling automatic restart on Windows failure, but I still don't see a blue screen.

My temps, according to SpeedFan, are fine. The GPU idles around 60C (a little high, but nothing crazy). The CPU idles around 39C. I've checked the BIOS temperatures and it confirms these numbers.

I haven't installed anything lately (that I can recall). I've done a virus scan to no avail. I had stupidly failed to clean out the inside of my computer for over a year and the dust bunnies had raided it, but I even cleaned it out to see if that would help anything. For over a year or so now, I've been getting this error at start up, something to the tune of "SQLDumper failed to initialize" or something like that, but it hasn't affected anything before.

Should I try repairing Windows? I'm not sure what else to do. I don't feel like building a new desktop because I'll be getting a Laptop for graduation.
 
This comp is so finnicky. The SQL solution removed the error and the comp was fine for 5 or 6 hours, but then it crashed again. It won't even go to the BIOS. Whereas I used to get a beep, now there is no beep.

I've tried resetting the CMOS but that doesn't work. I reseated the RAM, unplugged everything, and reseated the gfx card to no avail.
 
Either one of the Thermaltake 500W should be good enough for your system (there are different 500W models) still could be failing, how hot does it get? this could be hard to know if it has been sitting OFF for a while :)

First thing, have you checked the capacitors on the motherboard for any sign of bulging or leaking?

Do you have other DDR modules to test?

Can you disconnect everything but the RAM, Vid card, CPU heat sink /fan and the keyboard, does it POST?

If it does, start connecting components one at a time and re-boot.
 
Would the damaged capacitors be pretty easy to spot? From a quick glance, everything looks fine, but I'm not real sure what to look for.

I disconnected everything but what you mentioned, and still no post. The first attempt after unhooking everything led to continuous beeping about once a second, but now that doesn't even happen. I tried only 1 stick of RAM as well. Is it looking like a new mobo or CPU or what?
 
Capacitors that are bulging will be bulging at the top. They should be flat at the top.

Also, do you know the name of the BIOS your MSI mobo utilizes? For example Phoenix or Award. It should be in your manual.
 
The manual sucks (not even a troubleshooting section) so I'm not entitely sure of the BIOS, but I believe it is Award.

I may have some bulging capacitors, but once again I can't be too sure.

When I do hear the beeps, this little diagnostic tool shows an error at memory. Maybe it is the memory?
 
With that AWARD BIOS beep code it is either missing memory or the memory isn't seated right.

1. Turn off your PC and then unplug from your power source.

2. Take off your side panel and make sure you are grounded because just one small tiny static electricity can fry a motherboard. Touch your power supply and you should be grounded.

3. Now check your memory sticks. Are they set correctly? You may want to take them out and reseat them in the slots you found them in. Make sure they are aligned right but don't force them. There are two latches, one at each end that should snap in place.

* What are the results?
 
Capacitor plague

Next, you want to try another power supply, even if the caps are bulging a bit, just to know.

I'm off for a game of pool ... yay ... you're in good hands with Route44.


EDIT: Also try another vid card, even an old PCI one just to check after you`re done other checks.
A capacitor could leak from the bottom too which makes it harder to notice and detect.
 
After reading the Wiki, I'm pretty sure that I do have 4 bulging capacitors. Am I screwed?

I'll try reseating the memory again, though.

As a side note, it's really annoying typing everything with a PS3 controller! :) Thanks for all the help.

Edit: Reseated the memory but I still get the same results.
 
MSI does use cheaper capacitors but better than caps in mobos like PcChips, you might still get some life out of that one, but if you get it going, I would always have a back up of all important data.

With bulging capacitors, the health of all other components is questionable, where regulators are involved.

I would try another power supply, open the Thermaltake check it's state inside.

If you do get it going, do a few BurnIn Tests back to back, get the free version, does 15 minute tests only.

Do this before you begin using the system on daily basis if you do get it going, I have seen motherboards with caps bulging a bit that still run for quite a while, it all depends on few additional factors!

If you can afford it, time for a new rig perhaps.
 
You can get capacitors fixed but I think Sharam would tell you the cost might outweigh the benefits.

Getting a 939 board is difficult but not impossible. You can get new ones on eBay (I got a used EPoX 9NPA+Ultra from an IT person and it runs great), believe it or not, or there might be some good local tech stores that might have a few in storage.
 
I doubt the Thermaltake is damaged as I just installed it Thursday, unless of course the constant shutdowns have taken a toll. I will, however, plug in my old PSU to see if I can get a couple more boots out of it.

As my cash reserves are limited, would a new mobo and CPU do? What would you recommend? It would have to support DDR400 RAM. All I really use the comp for is playing some games (nothing intensive like FEAR or any game like it) and surfing the web.
 
If the PS is new, then we have to assume the mobo is the culprit, to get a better picture of how everything started, can you give us the make and model of the old PS and a reading from the sticker for the +3.3V, +5.0V and the +12.0V max output in Amps? What are the same outputs on the new PS? I've had bad units out of the box, it does happen!

How many hard drives, optical drives, case fans over the time period you've had the system, even if at one point you had another vid card, any other changes. This will help to understand how hard the components have been running over the years so we can decide what has to go and what can stay.

Trying to see if we can get some life out of this, if you want to bother with it.

Do you have a multimeter to check the output voltage (this does not prove load) Yellow is +12V, Red +5V and Orange +3.3V. Get a paper clip or a small solid wire, with the power supply disconnected from everything, turn the rocker switch OFF at the back of the PS, on the ATX 24pin, locate the green wire (the fourth from right looking at the clip side) and short it with a black wire (best to use the one above it to the left or right, 3rd or 5th in the upper row. Plug in the power supply and turn it on from the back. All + rails should be within + %5. Do you smell what we call a "burnt peanut smell" out of the PS?

It is perhaps time to take the motherboard out of the case, completely disassembled. Check both sides, any burnt mark at the bottom, any other things you notice? Remove the battery while you visually inspect the board, With the motherboard out of the case, re-install the CPU, heat sink and fan, RAM, vid card (can you get your hands on another PCI or PCI-E card?) and the keyboard, put the battery back, use a proper bed for the mobo, double check the CPU fan wires to the mobo, connect the power supply, plug it in, short the front panel power ON pins on the mobo, does it POST?

As Route44 suggests, it is harder to get a new decent socket 939 mobo, but many used around as in the provided link, EPoX is a very decent make if you asked me, the A8N series from ASUS are decent and they have the same chipset if you decided to do a repair install with the new mobo but I'm concerned with data integrity and suggest to delete the partition, re-create a partition, perform full format with a fresh install if possible.
 
Sharam, what is your experience concerning FoxConn 939 boards from the 939 heydays? By the way, did you win your pool game? :stickout:
 
After reading the Wiki, I'm pretty sure that I do have 4 bulging capacitors. Am I screwed?

I'll try reseating the memory again, though.

As a side note, it's really annoying typing everything with a PS3 controller! :) Thanks for all the help.

Edit: Reseated the memory but I still get the same results.
If your caps look the same as the ones in the Wiki article (the second picture to the right under Symptoms) with the white stuff on the top, don't bother with the motherboard tests and replace it without trying to power up again.

If there are just blown up a bit, you can proceed with the tests if you want.
 
Sharam, what is your experience concerning FoxConn 939 boards from the 939 heydays? By the way, did you win your pool game? :stickout:
I'm not a big fan of FoxConn and have had few in the shop for repair.

Nah, was just a reason to get out for quick brown pop with friends, they are still there. Nothing like a decent snooker table, I have to blame the tables in the bar cross the street, the ones you put coin in :)

EDIT: tried to merge my 2 posts, too late and no option to delete :(
 
Old PSU came with the case. It's a Raidmax. 420W
3.3V - 26A
5V - 32A
12V - 13A

Thermaltake Purepower 500W
3.3V - 22A
5V - 32A
12V - 15A

As for changes, the only thing I added was a TV Tuner in '06. Everything else has been the same. Here's all the other specs:
Sony CD/DVD Burner
Western Digital 80GB HDD (connected through SATA)
1 floppy drive

I haven't had any problems before this.

As for shorting wires, how exactly do I do that? I'm pretty sure my dad has a multimeter for me to use. I think the PSU is fine because I bought it in an effort to stop the crashes. It only now stopped POSTing.

I know this is kind of a silly question but how do I remove the CPU heatsink? I flipped the safety lever but I could not figure out how to undo the other bracket.

Thanks again.
 
Sure, blame the coin operated tables... :rolleyes:

Anyway, the reason I asked is because he is searching for a motherboard and a number of new ones can be had for around $70. I, too, have never been a fan, but I've never had to repair them either.
 
Route44, that is about all available now for Socket 478 and 939 from our suppliers brand new as well as Asrock last I checked. Better to get a good old proven working Asus or a like as much as I don't like their design.

13A and even 15A on the +12V rail is not good for me, I don't go bellow 18A these days but prefer 20A+.

Your processor takes about 6A from the +12V rail and the vid card around 4.5A, that leaves little for the optical and hard drives (3A together)

Even though the system might have ran OK, it has been pushing the power supply over the years, prove is in the cheaper caps not taking the load. The heat due to the dust you mentioned, helped speed things up.

The 500W should run your system OK, if you upgrade the processor or the video card and add more drives in the future, I would replace the power supply at the same time to avoid getting where you are at today.

As I mentioned before, if the caps are just blown up a bit, you might be able to get some life out of this, the output on the +12V and the +3.3V rails of the old power supply were just enough for the components in your system, the dust didn't help much but again, can't be sure how good it was able to operate under load and we don't know yet the output voltage of either of the two.

You can have fun and run some test if you are like me and want to get to the bottom of this even if you change the mobo, I'm still not %100.00 sure of the health of the video card and the RAM, BurnIn test would be able to fill in the blanks somewhat.

EDIT: let's not remove the CPU till the time comes when and if you get a new motherboard, would've been good to remove it but then we have to clean the old Thermal paste and apply fresh paste, we can do this if it worked or when you get that new motherboard.

I looked at the picture of the motherboard, it does not show a bracket so I don't know what type of retainer you have.

Make a small U from the paper clip (snip the rest) and push it in the openings looking at the ATX 24pin as I explained (green to black) this will power up the PS when the switch at the back is turned On allowing you to take a reading from the wires.

EDIT 2 I knew I was tired, the TT 500W has 2 X +12V rails, +12V1 14A and +12V2 15A. So you are more than OK for now and future upgrades.
 
Is there any story that I can pick up thermal paste for the CPU? I forgot to order some and don't feel like putting off the install of the mobo while I wait for the thermal paste. Would Radio Shack have it by any chance or is a local mom and pop tech shop my only choice?
 
You are looking for Arctic Silver 5 from your corner computer store, I don't know about RS! The small tube would last you a while.

When you do get it, please let us know so we can provide proper steps to take. You are going to need some sandpaper, 1000 or 600 Grit would do, if you can get your hands on some [ame="http://www.amazon.com/Electrical-Contact-Cleaner-oz-aerosol/dp/B000Q8C748"]Electrical Contact Cleaner[/ame] most probably not from your corner computer store, that would be great.

The links are just a quick search I did.
 
Got the new mobo in today. I'm going out to pick up some thermal compound and to see if I can find some electrical contact cleaner. Here's a pic of the CPU heatsink clamp that I couldn't figure out. I hope this helps a little bit. On the other side is the safety lever.

cpupic.jpg
 
Protect the area under and around the hold down clamp in the pic, put some heavy material directly on the mobo, even on the other side of the clamp. use a flat head screwdriver and insert it in the opening, make certain the screwdriver is in, push down on it then away from the heatsink, it should come out, release the other side, wiggle the heatsink as it is most probably stuck to the CPU, once there is movement, pull the heatsink up.

You should do this with the ZIF socket's lever in the locked position, if the heatsink "bond" with the CPU is not broken by moving it clockwise and counter clockwise first and you pull up hard on it, the CPU is going to be forced out of the socket, breaking or damaging the pins possibly.

Once you get the paste and stuff let us know.

EDIT: Where do you live, if in Canada I know where you can get some contact cleaner.
By heavy material I don't mean metal or anything that can scratch the mobo.

Since your RAM is so close, remove the RAM and other components before attempting to remove the heatsink
 
Status
Not open for further replies.
Back