« Welcome little Amy! |
Main
| Observation on friendships »
April 27, 2005
Tech advice needed
Hi folks. Been having technical problems everywhere today, from rats in the roof cavity to boot problems with my PC. The rats I can handle (they're not getting in to the house yet), but the PC I'm worried about. I've included in the extended entry, the symptoms and the diagnosis I have made. Any advice or assistance anyone else has on the matter would be much appreciated.
EDIT: I should add this is a motherboard that was purchased and installed last September and the chip was purchased then, too. And the bios beeps I'm getting would indicate a memory error.
I think my PC problems might be layered. I've been having troubles occasionally with the CDRW (not recognising disks, not opening door, etc) and it seems to be the CD spin start that has stalled startup on the occasions that the BIOS beeps have occurred. I'm also concerned about CPU and HD0 temperatures and the seating of the RAM.
But I guess I should just give you a low-down of what happened, complete with details. Any advice any of you may have on possible/likely causes would be much appreciated.
1. WHAT HAPPENED
It began a week or two ago when the CDRW basically wouldn't read anything or even open its door until I rebooted. I just thought of it as a glitch.
Then a few days later I had long beeps on cold boot. Beeps lasted about 3-4 seconds and then there was a 1-2 second break before it happened again. Nothing came up on the monitor. I allowed myself the obligatory curse, and turned the machine off, waited 10 seconds and tried again. Same thing happened. I turned it off, waited 10 minutes and tried again and it started up fine, but I got a "CMOS checksum error: Defaults loaded" message. I pressed F1 and all seemed okay. Must admit I didn't diagnose the problem though, beyond looking at my sheet of beep explanations, which seemed to indicate a memory error (at least that's what most BIOS beep error code reports are saying).
No problems for the next few days, so again I just hoped for the best and then this morning I got the long beeps on cold boot again. I cursed again, and then turned the machine off for a few minutes. Once again the system started up fine with the exception of the checksum error message again. I started trying to look up similar problems online and after 10 minutes of working the system simply turned itself off and tried to boot again with no warning. Only it couldn't boot and was emitting the long beeps. I left the machine off for a few minutes and it booted fine again with the checksum error. This time in the few minutes I had, I checked the system properties and found RAID controllers without drivers. But by the time I tried to look at a few sites to conitnue diagnosis, the system had switched off and was emitting the beeps again.
2. WHAT I DID NEXT
So I did the next logical thing and unplugged everything and took the side off the case. All seemed to *look* okay and I had a go at just jiggling and pushing the RAM dimm to ensure it was properly seated, reconnected keyboard and mouse and restarted - fine. All fans were working (on power supply as well as CPU, but the power supply fan was running extremely quietly). So I reconnected monitor and network cable and started again - no problem.
Next I reinstalled RAID controllers from the mobo CD (in the DVDRW drive just to be safe), and checked SystemFan temperatures. Here are the readings as I type this:
Temp1 (internal): 33 degrees. Temp2: 27 degrees. Temp3 (CPU): 49 degrees. HD0: 41 degrees.
I've seen the HD0 temp up at 43 degrees and since reconnection of all devices, the CPU has got up to 50 degrees, but only for a few seconds before reverting to 49.
So temperatures aren't *great* as I have the side off the pc at the moment and so air is getting to the CPU more easily.
3. DIAGNOSIS
Basic specs are as follows...
Mobo: Gigabyte K7 Triton
CPU: Athlon XP 3200+
VGA: on board (long story attached to that, but I have a very nice VGA card I'm not actually using at the moment because in spite of the fact that it is also a Gigabyte brand device, the mobo won't boot if it's connected... and yes, I've tried changing the BIOS; no diff).
After all the above symptoms/activities occurred, I left the PC running an hour to see if it would shut down again and it hasn't done so. Yet. But have checked BIOS upgrades, etc, and found I have the latest driver. Details are as follows (from eSupport BIOS agent):
Program: eSupport.com BIOS Agent Version 3.44
BIOS Date: 04/21/04
BIOS Type: Award Modular BIOS v6.00PG
BIOS ID: 04/21/2004-KM400A-8237-6A6LYG0WC
OEM Sign-On: GA-7VM400AM(F) F3
Chipset: VIA 82C3205 rev 0
Superio: ITE 8705/SiS 950 rev 2 found at port 2Eh
OS: WinXP SP2
CPU: AMD Athlon(tm) XP 2171 Mhz MAX: 2800 Mhz
BIOS ROM In Socket: Yes
BIOS ROM Size: 256K
Memory Installed: 256 MB
Memory Maximum: 512 MB
Memory Slot 01: 256 MB
Memory Slot 02: 0 MB
I've also checked the CDRW and it seems to be working. Or at least it's reading stuff.
(And yes, I know it's insane I only have 256M RAM.)
So for the time being I don't see what else I can do, other than keep an eye on things and maybe investigate new cooling for the system. And I'm pricing RAM too. I had forgotten I'm only running on 256M RAM and that appears to be Kingston PC3200 DDR RAM at that.
4. YOUR ADVICE?
So. Have I missed anything? Should I turn this thing off and not reboot until I have better cooling? Should I remove the CDRW? Any advice and any diagnosis suggestions would be much appreciated. Also, FWIW, I'm speaking of a box which is running XP SP2, fully patched, firewalled and virus-protected. Not that that means jack when it comes to hardware problems (which this clearly is) but it might help in terms of driver compatibility.
Anyway - I'd be grateful for any suggestions.
Posted by jj at April 27, 2005 12:56 PM
Comments. Please PREVIEW your comment before submitting. This helps the server to save your entry.
I don't think the temperatures are particularly atypical.
How MANY long beeps and in what pattern?
Why are you running RAID anyway? What sort of drives have you got?
If you're getting flakiness from the CD drive, the memory and the checksum and spontaneous reboots... and they're not happening consistently... eewwwr.
Possibly your BIOS/Mobo has reached its MTBF and is about to go to the GJITS.
How long have you had this machine?
Grab a Memtest CD, boot off it several times in a row and put the PC through the hardest tests possible. Also (not instead) grab a live Linux CD like Knoppix, boot off it and check the dmesg log (su; dmesg | less) for specific error messages.
Are you prepared to do a BIOS flash?
And be prepared for the eventuality you might be looking at a new motherboard.
Posted by: P at April 27, 2005 1:13 PM
jj, I am not a hardware tech, but the beeps as you hit the bios mean things. Have a look on the internet for what the beeps mean for the make of computer or the bios. It might be something as simple as bad RAM.
Posted by: Cameron Riley at April 27, 2005 1:38 PM
Could be something as simple as a bad CMOS battery. That would cause the CMOS to reset to defaults, which would cause various devices to not be recognised.
I'm guessing (only a guess) that it's not the battery itself, but a connection to it, which is fine when the machine is cold but fails when it expands as it heats up.
Posted by: Stephen Soymonoff at April 27, 2005 2:57 PM
Re: Edit: It may be that the Mean in MTBF is a little left-of-the-bell-curve skewed and may (or may not) still be headed for the GJITS. The nature of statistics.
Still strongly recommend the Memtest and Knoppix CDs. Run, say, five or ten Memtests, maybe boot Knoppix five or so times and dump the dmesg log to a text file on a USB drive, see if you can get different hardware failure messages.
Let me know when you're in next, I'll prep one of each for you along with the DVD I promised (ooh, and more goodies if you wanna beat, say, the ABC to it).
I honestly don't think the drivers are gonna make much of a difference; no drivers apply at BIOS time, and drivers for IDE/ATAPI devices should be so frighteningly standard that the version of Windows should make very little difference...
Failing that, if you have or can borrow more/other RAM, swap it in and see if it works. Again, really, do the Memtest.
The only reason I'm not putting it all down to memory is because of your descriptions of other problems.
Posted by: P at April 27, 2005 3:30 PM
hmm. i'd suspect a motherboard and/or bios problem. Is there a fan on the (not sure which one your motherboard would have) north or south bridge?
i'd try updating the bios (if you were willing to chance it). i'd say there is at least a motherboard/bios problem, as while memory problems may cause what you're seeing (cdrw problems) it shouldn't affect the bios. you could also try replacing the cmos as cameron suggested to see if that works.
Posted by: westyx at April 27, 2005 3:57 PM
With fluctuating temperatures you can sometimes get what is known as chip creep. The chips (the ones in slots) slowly move in the slots, and yes I have seen this happen.
Try reseating the RAM and any other cards and see if that makes a difference, sometimes it cures a lot of problems.
Just my two cents...
Posted by: Misty at April 28, 2005 11:55 AM
Post a comment
Please PREVIEW posts before submitting. This helps my host server to save your comments.