3dfx Archive
http://www.falconfly.de/cgi-bin/yabb2/YaBB.pl
This & That >> This & That >> Wierd pc issue
http://www.falconfly.de/cgi-bin/yabb2/YaBB.pl?num=1273553402

Message started by bucksavage on 11.05.10 at 06:50:02

Title: Wierd pc issue
Post by bucksavage on 11.05.10 at 06:50:02
I've been having an odd issue with my pc the last while.  It only happens if there is a power outage, or if I have to turn the powerbar that my pc is hooked to off but when I get power restored, when i bott up, only the power pack fan anf the cpu and it's fan turn on.  If I press the reset button 5-7 times then my DVD/CD RW combo gets power, and after a few more resets the vid card and hd's fire up.  After everything finally gets power again, it runs like clockwork until a loss of power.  What could cause this? 

Title: Re: Wierd pc issue
Post by Tim on 11.05.10 at 09:18:15
I'm positive that's a power supply unit problem.

Title: Re: Wierd pc issue
Post by bucksavage on 12.05.10 at 02:29:20

Tim wrote on 11.05.10 at 09:18:15:
I'm positive that's a power supply unit problem.

Hmm, that's what makes me wonder because the PSU must get power and transfer it if the fan attached to it runs, wouldn't it?  The odd thing is that this never happens at all under normal situations. 

Title: Re: Wierd pc issue
Post by Tim on 12.05.10 at 14:27:02

bucksavage wrote on 12.05.10 at 02:29:20:

Tim wrote on 11.05.10 at 09:18:15:
I'm positive that's a power supply unit problem.

Hmm, that's what makes me wonder because the PSU must get power and transfer it if the fan attached to it runs, wouldn't it?  The odd thing is that this never happens at all under normal situations. 


I think it's either your AC current supply from the mains socket or powerstrip that's dodgy after the power outage, or your PSU must have gotten a bump after a blackout/brownout. What I would do in your situation is get a cheap multimeter and test the powersupply 3.3, 5 and 12v+ rails to see exactly if the power is stable.

Do this when you get the weird power issue.

To easily do this unplug every wire from mainbord, VGA, HD, drives, and only leave a few fans attached.

Switch off PSU by rocker switch on back of PSU.

Jumper the green pin and any black pin on the ATX power connector with a piece of wire. I use a paperclip.

Then power up the PSU. The fans should turn on. Now insert your red probe into a yellow or red wire on any molex (12 or 5v), and your black probe in a black wire, do this in this order, do the reverse of this when you take em out. Check to see if they deliver stable voltages, they shouldn't fluctuate and should be around the 12 or 5v mark.

Do the same for the brown wire on the ATX plug (3.3v)

If all this checks out allright, plug everything back in. And now when PC is fired up do the same thing. Also check under heavy load scenarious such as Furmark/3DMark to see if the rails are still stable.

If they are stable, it could well be your motherboard. Sometimes if your mains supply is unstable due to situations beyond your control ie. powerboard/grid defects, they can do damage to your electronics.

If your powergrid is dodgy, it's best to get a powerstrip that filters your powersupply, or even better a USP, uninterrupted power supply, basically a big battery that which lets you save your work in the even of a powerout and lets you safely power down.

Hope that helps.

Title: Re: Wierd pc issue
Post by bucksavage on 13.05.10 at 07:13:23
I'll have to check those ideas out somtime.  For now though it's running well, no power issues, at least until I a) have a power outage, or b) cut the power by turning off the power bar.  I only usually turn off the bar in the event of a thunderstor, cause I don't need to have a power surge or a lightning strike fry it. 

only other question I thought of was, would this also happen if the cmos battery was dead or very near dead?  Does it not initiate the start up of every thing other than the basic power to the fans and such? 

Title: Re: Wierd pc issue
Post by Tim on 13.05.10 at 11:40:44

bucksavage wrote on 13.05.10 at 07:13:23:
I'll have to check those ideas out somtime.  For now though it's running well, no power issues, at least until I a) have a power outage, or b) cut the power by turning off the power bar.  I only usually turn off the bar in the event of a thunderstor, cause I don't need to have a power surge or a lightning strike fry it. 

only other question I thought of was, would this also happen if the cmos battery was dead or very near dead?  Does it not initiate the start up of every thing other than the basic power to the fans and such? 


Stange that it's just running fine normally. I would suspect it to be the mains?

No a dead CMOS battery just forgets the settings, the machine will boot up and says CMOS defaults loaded.

Title: Re: Wierd pc issue
Post by bucksavage on 14.05.10 at 05:28:06

Tim wrote on 13.05.10 at 11:40:44:

bucksavage wrote on 13.05.10 at 07:13:23:
I'll have to check those ideas out somtime.  For now though it's running well, no power issues, at least until I a) have a power outage, or b) cut the power by turning off the power bar.  I only usually turn off the bar in the event of a thunderstor, cause I don't need to have a power surge or a lightning strike fry it. 

only other question I thought of was, would this also happen if the cmos battery was dead or very near dead?  Does it not initiate the start up of every thing other than the basic power to the fans and such? 


Stange that it's just running fine normally. I would suspect it to be the mains?

No a dead CMOS battery just forgets the settings, the machine will boot up and says CMOS
defaults loaded.


Hmmm....it's a curious porblem indeed.  I wondered for a secif there could be a loose connection in the wiring harness to the drives, but my previuos pc had that, and when the wire loosened, all the drives just didn't work, but the pc still booted up.  Almost sounds like the power isn't getting into the badic I/O areas to juice everything other than the cpu and the psu up.  I guess in the long run I can 't complain too much, I got this tower and motherboard for free.

Title: Re: Wierd pc issue
Post by bucksavage on 03.09.10 at 02:42:47
Just thought I'd post a little update on this as I recently installed a program that monitors temps, cooling and power.  For my PSU, these are the power reating as of right now: 
+3.3V standby - between 3.317 and 3.352
+5V - betwwen 5.026 and 5.052
+3.3V - varying between 3.180 and 3.214
CPU Core (+1.85V) - between 1.8 and 1.812
CPU I/O - (1.5V) - between 1.471 and 1.484

Would these be within tolerance? 

Title: Re: Wierd pc issue
Post by Drakan Player on 03.09.10 at 03:36:18

bucksavage wrote on 03.09.10 at 02:42:47:
Just thought I'd post a little update on this as I recently installed a program that monitors temps, cooling and power.  For my PSU, these are the power reating as of right now: 
+3.3V standby - between 3.317 and 3.352
+5V - betwwen 5.026 and 5.052
+3.3V - varying between 3.180 and 3.214
CPU Core (+1.85V) - between 1.8 and 1.812
CPU I/O - (1.5V) - between 1.471 and 1.484
Would these be within tolerance? 


The 3.3 is a little low on load but not to the point of crash happy but you need to monitor your 12v source. Any thing lower than 11.5v then you got a problem and need to upgrade/replace your psu.

Title: Re: Wierd pc issue
Post by Drakan Player on 03.09.10 at 03:42:05

Tim wrote on 12.05.10 at 14:27:02:

bucksavage wrote on 12.05.10 at 02:29:20:

Tim wrote on 11.05.10 at 09:18:15:
I'm positive that's a power supply unit problem.

Hmm, that's what makes me wonder because the PSU must get power and transfer it if the fan attached to it runs, wouldn't it?  The odd thing is that this never happens at all under normal situations. 


I think it's either your AC current supply from the mains socket or powerstrip that's dodgy after the power outage, or your PSU must have gotten a bump after a blackout/brownout. What I would do in your situation is get a cheap multimeter and test the powersupply 3.3, 5 and 12v+ rails to see exactly if the power is stable.

Do this when you get the weird power issue.

To easily do this unplug every wire from mainbord, VGA, HD, drives, and only leave a few fans attached.

Switch off PSU by rocker switch on back of PSU.

Jumper the green pin and any black pin on the ATX power connector with a piece of wire. I use a paperclip.

Then power up the PSU. The fans should turn on. Now insert your red probe into a yellow or red wire on any molex (12 or 5v), and your black probe in a black wire, do this in this order, do the reverse of this when you take em out. Check to see if they deliver stable voltages, they shouldn't fluctuate and should be around the 12 or 5v mark.

Do the same for the brown wire on the ATX plug (3.3v)

If all this checks out allright, plug everything back in. And now when PC is fired up do the same thing. Also check under heavy load scenarious such as Furmark/3DMark to see if the rails are still stable.

If they are stable, it could well be your motherboard. Sometimes if your mains supply is unstable due to situations beyond your control ie. powerboard/grid defects, they can do damage to your electronics.

If your powergrid is dodgy, it's best to get a powerstrip that filters your powersupply, or even better a USP, uninterrupted power supply, basically a big battery that which lets you save your work in the even of a powerout and lets you safely power down.

Hope that helps.


Very good post and I suggest that OP look into how his/her local grid functions during load ans spikes knowing that it has a big impact on every thing drawing power from it. My grandmother spent 12 years building high voltage stepping transformers so i got to see some of what takes place there. They are vulnerable to emp and solar flares that can cause spikes in the lines that can overheat the coils with in the transformers and storms often cause dips and spikes. The worst case is when there is a momentary spike that causes permanent failure of the grid till the damaged infrastructure is repaired.

Title: Re: Wierd pc issue
Post by bucksavage on 03.09.10 at 03:55:21
Ok, the only thing is that this program I have now doesn't show the 12V rating.  Is there one that will display that?  I also saw in an earliler post in this thread about something called, Furmark/3DMark.  What are they? 

As for my PSU, the one I have in the pc now is a Dainty DR-A300ATX; the PSU I intend to switch to, assuming it fits in my tower is an Orion HP400D, that also has one plug for a SATA drive.  Maybe I should switch sooner rather than later? 

Title: Re: Wierd pc issue
Post by Tim on 03.09.10 at 11:31:33
Forgot about using software to measure your rails. They are so off the mark it's not funny anymore. The only accurate way to measure them is by multimeter.

Title: Re: Wierd pc issue
Post by elfuego on 06.09.10 at 19:46:30

Tim wrote on 12.05.10 at 14:27:02:

bucksavage wrote on 12.05.10 at 02:29:20:

Tim wrote on 11.05.10 at 09:18:15:
I'm positive that's a power supply unit problem.

Hmm, that's what makes me wonder because the PSU must get power and transfer it if the fan attached to it runs, wouldn't it?  The odd thing is that this never happens at all under normal situations. 


I think it's either your AC current supply from the mains socket or powerstrip that's dodgy after the power outage, or your PSU must have gotten a bump after a blackout/brownout. What I would do in your situation is get a cheap multimeter and test the powersupply 3.3, 5 and 12v+ rails to see exactly if the power is stable.

Do this when you get the weird power issue.

To easily do this unplug every wire from mainbord, VGA, HD, drives, and only leave a few fans attached.

Switch off PSU by rocker switch on back of PSU.

Jumper the green pin and any black pin on the ATX power connector with a piece of wire. I use a paperclip.

Then power up the PSU. The fans should turn on. Now insert your red probe into a yellow or red wire on any molex (12 or 5v), and your black probe in a black wire, do this in this order, do the reverse of this when you take em out. Check to see if they deliver stable voltages, they shouldn't fluctuate and should be around the 12 or 5v mark.

Do the same for the brown wire on the ATX plug (3.3v)

If all this checks out allright, plug everything back in. And now when PC is fired up do the same thing. Also check under heavy load scenarious such as Furmark/3DMark to see if the rails are still stable.

If they are stable, it could well be your motherboard. Sometimes if your mains supply is unstable due to situations beyond your control ie. powerboard/grid defects, they can do damage to your electronics.

If your powergrid is dodgy, it's best to get a powerstrip that filters your powersupply, or even better a USP, uninterrupted power supply, basically a big battery that which lets you save your work in the even of a powerout and lets you safely power down.

Hope that helps.


After this post I can only say that Tim knows what he's doing. Only thing left to say is be careful when playing with current and personally, I would 'paperclip' the wires before plugging the power cord in the PSU, but it's just me ;-)

Title: Re: Wierd pc issue
Post by bucksavage on 07.09.10 at 18:26:45

elfuego wrote on 06.09.10 at 19:46:30:

Tim wrote on 12.05.10 at 14:27:02:

bucksavage wrote on 12.05.10 at 02:29:20:

Tim wrote on 11.05.10 at 09:18:15:
I'm positive that's a power supply unit problem.

Hmm, that's what makes me wonder because the PSU must get power and transfer it if the fan attached to it runs, wouldn't it?  The odd thing is that this never happens at all under normal situations. 


I think it's either your AC current supply from the mains socket or powerstrip that's dodgy after the power outage, or your PSU must have gotten a bump after a blackout/brownout. What I would do in your situation is get a cheap multimeter and test the powersupply 3.3, 5 and 12v+ rails to see exactly if the power is stable.

Do this when you get the weird power issue.

To easily do this unplug every wire from mainbord, VGA, HD, drives, and only leave a few fans attached.

Switch off PSU by rocker switch on back of PSU.

Jumper the green pin and any black pin on the ATX power connector with a piece of wire. I use a paperclip.

Then power up the PSU. The fans should turn on. Now insert your red probe into a yellow or red wire on any molex (12 or 5v), and your black probe in a black wire, do this in this order, do the reverse of this when you take em out. Check to see if they deliver stable voltages, they shouldn't fluctuate and should be around the 12 or 5v mark.

Do the same for the brown wire on the ATX plug (3.3v)

If all this checks out allright, plug everything back in. And now when PC is fired up do the same thing. Also check under heavy load scenarious such as Furmark/3DMark to see if the rails are still stable.

If they are stable, it could well be your motherboard. Sometimes if your mains supply is unstable due to situations beyond your control ie. powerboard/grid defects, they can do damage to your electronics.

If your powergrid is dodgy, it's best to get a powerstrip that filters your powersupply, or even better a USP, uninterrupted power supply, basically a big battery that which lets you save your work in the even of a powerout and lets you safely power down.

Hope that helps.


After this post I can only say that Tim knows what he's doing. Only thing left to say is be careful when playing with current and personally, I would 'paperclip' the wires before plugging the power cord in the PSU, but it's just me ;-)

I'm not that well versed in doing this kind of stuff, but, if I do check it, I'll make use of all the advice given here.  Thx everybody. 

3dfx Archive » Powered by YaBB 2.4!
YaBB © 2000-2009. All Rights Reserved.