Random Rebooting

[tetrakarbon]

Let's do the specs first:
A7N8X-Deluxe (nforce2 board, if you don't know)
Athlon 2100+ (don't remember the core)
3x256MB Ram running at 133mhz (so there's a divider worked in there, I think my proc is a 166FSB), generic
Antec 380 (can't remember W so well, will need to take this out)
1 Maxtor 120GB, 7200RPM drive
1 Maxtor 4GB drive, probably 5200RPM
Generic DVD-RW and CD-RW drives
ATi Radeon 9800Pro (128MB)

Hi guys, here's the problem: random reboots. I'm using WindowsXP, with the setting that shows me a bluescreen if the OS hangs, rather than just rebooting. I am sure this is not related to Windows or any other program - it happens very randomly, and I don't get a bluescreen.

Should note that this computer is about 2 years old, and I haven't had this problem before. Reformated just a month ago for no particular reason other than a cleanslate, problem started about 3-4 days ago.

After letting the computer cool, I start up the BIOS. After a little while, idle temp gets up to 44 degrees centigrade. This is too hot, no? Motherboard hangs around 21 degrees C, which seems okay. Perhaps I ought to reset my heatsink (PAL8045).

I'm guessing it's a power issue. I have an Antec 380 (perhaps higher W, I will have to take it out to check the sticker), which ought to be fine for my system (gfx card only thing that could be pushing it, seems to me). Anyway, I've got these readings from the BIOS:

vcore - 1.79 (set to 1.75, perhaps this is at fault?)
3.3v - 3.16V
5v - 4.48 (this perhaps indicates the gfx card stressing things?)
12v - 11.9v (I think I might have written this down wrong).

Anyway, I've downloaded MBM and will start running that as soon as I reboot (which might be sooner than I expect!), but let me know if preliminarily you have any suggestions

Thanks guys!

EDIT: Make this one quick before another reboot. It rebooted, and temp was at 54-57 deg C (it moved up after a few seconds) when I went into BIOS immediatly afterwards. Then, it did a loop of rebooting (got to the part right before the WinXP loading screen, then rebooted from the POST again) did it about 5-6 times before I turned it off. This was immediatly after trying to run MBM, but that might have been a coincidence. Tell me what you think

-- Ned Ruggeri

 

[Jonsey]

Normally, when you hear rebooting, you think PSU. Rebooting before you get into windows, though? You can test your PSU by unplugging everything you don't need to boot, like any extra HDD's or optical drives.

It almost sounds like your RAM or CPU is overclocked. Try underclocking everything to see if you can get into windows.

 

[tetrakarbon]

Thanks for the advice Jonesy. I will try that this weekend, as soon as I have some extra time.

I will check on the situation with the RAM and the proc. I have underclocked it already, to see if that would reduce the heat, but I don't think it took (still shows ~1.737Ghz in MBM)

I was running Pifast for a while, and MBM gave the socket temperature (which must have been the one given in the BIOS) as 57C or so, whereas the proc only got up to about 43 under full load, and is siting at 33 now. So I don't worry about that now.

Should have been clearer - rebooting before windows issue was only after using MBM. I changed a setting and the next time I used MBM nothing was wrong, though I can't be sure if MBM screwed things up. I might try to change back that setting to see if it causes that weird behavior again. In any case, I can now get into Windows again, though I am not sure if the original problem is resolved.

The +5v is running at 4.6v (though -5 is at 4.96). I assume this is outside of a comfortable range? I might try taxing the video card to see if that will cause a problem...

Thanks for the help, btw.

 

[Jonsey]

4.6 on the 5v rail is not too good. Don't the XP systems rely heavily on the 5v rail? That could be the problem.

 

[Slartibartfast]

I would def. reseat the heatsink (make sure you use some new thermal goop too). I pulled apart an athlon XP system I had used pretty heavily for about 2 years, (the hardware started doing really weird things) and the stock thermal paste was melted clean through.

 

[tetrakarbon]

Thanks for the advice guys. I will definitely reseat the HSF and reapply some thermal goop. But diode temp is 33C with 53C as the socket temp. So I think that's not the problem.

The +5v line dropped to 4.4v once (saw it in the MBM log), and never gets above 4.6v. This hadn't ever been a problem before, and I have had the gfx card (which would put the drain on the +5v line the most, wouldn't it?) for a long time. Maybe it's just the PSU deciding that the warranty is over now The other lines are generally pretty close, though the +12v line is running @ 11.31v.

Here's something strange - possibly the result of several reboots and not shutting down properly. I use Zone Alarm, and all the settings for it were set to default when I turned the computer back on after that cycle of rebooting. I don't know to what extent this indicates ZA is a factor...

-- Ned Ruggeri

 

[Slartibartfast]

Yeah those voltages are pretty screwy, which is odd cuz antec makes a good psu. I would try another one in there too, also try disconnecting one HD and the optical drives, see if that affects it any. If they're more than .1 off it's something to worry about, 4.6 on the 5 is pretty bad. I have an an Antec True Power 430 in a similar setup and the voltages stayed within .04 or so.

 

[tetrakarbon]

I'm writing from a friend's computer. It was definitely power. I unplugged the power connect, and a couple of the contacts were all browned out - both on the mobo & PSU cord. I guess I might try another PSU, just to see if that will work with the mobo - the damage to the contacts on the mobo make me wonder if it will work properly. Also, I wonder if replacing the end of the cord (where it plugs in to the mobo) might fix things, I'm wondering if the current drop was just becacuse of the corrosion that helped when the thing overvolted.

That, btw, must have been what happened. Antec freaked out and overvolted things. I assume that the lower voltage is either a result of the PSU freaking out, or perhaps it was the damage to the contacts that prevented some voltage on some lines. Regardless, I guess I shouldn't play with a busted PSU, so I'll buy another I think.

I'll put a picture up sometime for comments.

Edit: I reset the heatsink and applied some new goop (though perhaps not expertly), and the diode temp is now 20 degrees C, but a 43 degrees C reading on the socket. Which is generally more acurate? Does this 23 deg difference make any sense? Should I be worried - 43C at idle seems high, doesn't it?

I am still toying with fire by using the old PSU to write this. I need to get a new one and hope the mobo won't have any problems recovering (that is, the entire problem is with the PSU). Anyway, just goes to show Antec is unfailable.

-- Ned Ruggeri

 

[TheMostWantedPolishTwin]

4.6 on the 5v rail is not too good. Don't the XP systems rely heavily on the 5v rail? That could be the problem.

mobos that don't have the extra 4pin connector do run the CPU off the 5V rail...

 

[davidhammock200]

Test like this: http://forums.extremeoverclocking.com/showthread.php?t=137886

Select new PSU here: http://www.hardforum.com/showpost.php?p=1027898523&postcount=7

Good Luck,
Dave

 

[tetrakarbon]

Thanks guys. I will definitely stary looking. Having replugged in my PSU, I now have voltages as so:

3.3v: 3.25v
5v: 4.84v
12v: 11.31v
-12v: 11.81v
-5v: -4.86

So that's better, but not worth ignoring the initial problem.

Second, here are my load temperatures (after pifasting a good time), with the case open:

Diode: 29C
Sockett: 49C
Case:25C

That diode can't be right. Anyway, you guys think it's anything to worry about? Thanks again!

-- Ned Ruggeri

 

[wee96]

11.31v on the 12v rail is horrible, if thats the actual true reading (software isnt a true test).