GeneralThis forum is for non-technical general discussion which can include both Linux and non-Linux topics. Have fun!
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
I've been using the same tower pc for close to 6 years now and was able to get 3 more years out of it thanks to Linux. Today that pc fried itself while it was on...I smelled smoke and then the screen went. I thought it was the monitor, but that worked just fine on another pc. I pulled all the pci cards from it, swapped video cards, put another harddrive in and swapped CPUs. The fans are all running. The pc would start, but the red harddrive indicator wouldn't light and nothing appeared on the screen allowing me to go into the setup area. It won't even boot off floppy (I heard it check that drive). Everything in the pc has been upgraded except for the motherboard. Perhaps the motherboard went? How could I check the RAM since it won't boot off of floppy or CD?
Short of putting the RAM in another computer to test... if you've got more than one stick, then you can pull the memory individually to see if it makes a difference. I doubt that's the problem, though. If RAM goes in a way that can cause you to smell smoke, you'll see craters on one or more of the memory chips, and probably a fair amount of debris around the area.
Have you tried a bare boot? Pull absolutely everything that isn't needed to get into BIOS. All the PCI cards except video (if no OB video), all hard drives, cdroms, floppies, and all but one of your memory sticks (or two if it's 72-pin). Usually the BIOS will have some kind of hardware status monitor that can help you figure out what's wrong.
Check the motherboard for chips with craters in 'em. It doesn't happen much in newer hardware, but I've seen north/southbridges explode, as well as RAM and anything else that can get hot. Only time I was ever glad I forgot to take off my glasses . Such a crater could also be under a heatsink on the motherboard. I've also seen capacitors explode, though they usually aren't quite so dramatic when they go. Usually, they'll just overflow a yellowish sludge onto the motherboard. If you see it, don't touch it. Depending on the capacitor manufacture, that will be mildly corrosive, and could be very corrosive.
You might also have lucked out, and find that it's just a cold solder. If something got too hot, and you had a cold solder, the solder could come undone, and the "smoke" you smelled could just be the binding agent.
In any event, if you smelled/saw the magic blue smoke, and the computer decided to shuffle off its mortal coil shortly thereafter, it's probably not worth the headache to troubleshoot.
Originally posted by linuxhippy Good ideas-it wasn't enough smoke that I could see the remnants anywhere or smell the leftover. Hopefully it was the solder. If it's solder, could my pc come back to life tomorrow?
If you're good with a soldering iron, yeah... the hard part of that exercise is finding out which solder came undone. As long as you have the right fine-point soldering iron, and have the motor control and confidence to do it yourself, you could resurrect it.
It probably isn't a cold solder, though, given the age of the device... I did say that option would entail luck.
I suppose that it all depends on much time and effort you want to invest in trouble-shooting this problem, and also whether you have access to duplicate parts against which to isolate the cause.
On the assumption that this is going to be a project that you'll take on, in addition to the foregoing suggestions, I'd be interested in examining the following areas too:
1. When your PC boots up, does it make a series of beeps (i.e. the POST - Power On Self-Test series) just after the power up? If so, count those and then against the type of your BIOS run a google for POST codes to check out the decoding of those beeps. For example, on my Asrock K7VT2 mobo (using AMI BIOS), a series of 1 long and 8 short beeps meant that the video card wasn't seated properly. The beeps are meant for diagnostics and runs off of your BIOS - it's about as close to a machine hardware diagnostic as you are likely to get without relying on the OS which will pick up the ball once the BIOS is satisfied everything checks out and hands over control to the kernel. The type of BIOS your board is using should be apparent from a look at the board.
I found some good clues from these two sites when I had to run this level of diagnostic previously:
2. It might be possible that there was a power-surge which fried something, like a capacitor or some such. As a general rule - if you have the patience and the time, the best bet would be to look at testing each of the components. It is unlikely to be your PSU otherwise the machine wouldn't power up at all. Current is obviously flowing because you can hear it checking drives and running through its boot routine ... at least to some level. If you can't see anything on the monitor, it is difficult to say for sure just how far it gets in its sequence, isn't it. It could be located closer to your hard drive itself. Try a knoppix or other live CD and see if you can get into an OS that way, which will begin to reduce the options for the fault. However, you still need a video card function, which might mean that the video slot is cooked?
If you do figure this out, please post back here for reference on what it was and how it was fixed.
I hear no beeps. It doesn't turn over when it powers on....usually I hear the power turn over to the harddrives and see the monitor light up after all the fans start turning. Right now there is no signal going out to the monitor and the power indicator on the monitor remains orange and then turns green when there is a signal (usually). The fans turn...and that's it.
I took killerbob's advice and again pulled out all the pci cards except video and pulled 2 banks of RAM and left the first bank with 128 MB of RAM. I then unplugged power from the drives (floppy included) and unplugged the data cables from the motherboard. I still saw no visible evidence of a burn out and smelled nothing unusual. In the end, though, my pc fans would only turn and that's it....no boot.
I'm going to go out and buy a new power supply today. I replaced that about 3 years ago and it never got hot. I'm thinking that since I cannot see physical evidence of a burn out that it may be the power. Hopefully.
If I wind up buying a new pc, would a bare bones system be ok since I have all needed components?
If the new PSU doesn't fix the problem, at the least you have a spare PSU for a future use, so you can't really go wrong. Thinking past the PSU (assuming that is working), in an earlier post you mentioned that the system checked the floppy disk drive when you tried booting with a boot disk : "It won't even boot off floppy (I heard it check that drive)." But, the boot function itself didn't kick in and take over. That does sound suspiciously like a motherboard issue and could be related to the ROM on the board.
Installing a new motherboard is a valuable learning experience. The mobo I mentioned previously was a self-install upgrade from a smaller processor and board. With a new board and - depending on your budget - processor, cards, etc., you could build yourself a system, but if the PSU doesn't do the trick, I'd suggest get a new board (NB make sure that your current processor, etc will work with the new boards specs!!!), and you can probably use all your existing cards, etc. This way, there is no need to go out and buy a new PC unless you felt the need to specifically do that. Rebuilding your system is an enjoyable way of spending a Sunday afternoon and will in all likelihood be easier on the pocket. Also, because GNU/Linux software generally uses hardware more efficiently than some other types of OS software, performance-wise your current hardware should be still fine to use, I'd have thought.
Well the new power supply didn't fit the motherboard (slot on the PSU is too big). So I'll return that. I had an old motherboard from a Gateway (Pentium, 200 MHz, 64 MB RAM) that I wound up swapping out for the other. Once I got the original MOBO out I did see yellowish burn marks near the CPU-MOBO fried. I installed the other MOBO and now my tower lives again. Slow, but with Slack it's doable.
this new mobo already had the CPU and thermal sink dissipator installed. Should a heat sink fan also be attached? It's only a 200 MHz CPU and my power pack has a dual fan system-one exhaust fan is right above the cpu.
Last edited by linuxhippy; 08-14-2005 at 06:23 PM.
If you have a spare fan laying around, it wouldn't hurt. But my experience is that 200 MHz isn't really enough to burn out a CPU as long as it has a heatsync. I have a (shelved) 200 MHz box too -- an old AST. The ventilation is nowhere near the (weirdly-placed) cpu, but I never had any stability problems with it. All it has is a heatsync -- pretty typical of the old pentiums.
hmm...that's what I was hoping. I do have an old CPU fan for heatsinks, though. Unfortunately the 200 MHz heatsink doesn't have any place for fan screws. How would I attach it (glue)? Also, could I keep the case open with a big window fan on the pc?