LinuxQuestions.org
Help answer threads with 0 replies.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware
User Name
Password
Linux - Hardware This forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?

Notices


Reply
  Search this Thread
Old 01-28-2022, 07:16 AM   #1
Livestradamus
Member
 
Registered: Feb 2006
Location: Planet Earth (currently)
Distribution: Slackware & OpenBSD
Posts: 132

Rep: Reputation: 65
Unhappy Power Problem after outage


Hello All
I have a ~10 yr old desktop that’s had a problem staying powered on after an outage.
Here are some clues:
- Desktop can be in use or not but anywhere between 20 to 40 minutes it will shutoff hard.
- OS reports 2 out of 4 RAM sticks. Slots 0 and 2 are reported empty in the OS. In the BIOS it sees all 4. I have swapped them around to test. memtest does not even finish before it hard resets. I’ve tested couple of different Linux installs as well as systemrescue live CD
- BIOS settings completely resets every time this happens.
Is it the RAM modules or motherboard or something else you think that’s causing a problem?

Edit: Additional info

Have not noticed bad processes nor cpu spikes

Last edited by Livestradamus; 01-28-2022 at 07:24 AM.
 
Old 01-28-2022, 07:56 AM   #2
business_kid
LQ Guru
 
Registered: Jan 2006
Location: Ireland
Distribution: Slackware, Slarm64 & Android
Posts: 17,094

Rep: Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483
Sounds like the M/B or ram is blown.

Take the ram out of slots 0 & 2. Put the remaining ram into slots 0 & 1, and see will it boot and stay going with half the ram.
 
1 members found this post helpful.
Old 01-28-2022, 09:26 AM   #3
kilgoretrout
Senior Member
 
Registered: Oct 2003
Posts: 3,007

Rep: Reputation: 395Reputation: 395Reputation: 395Reputation: 395
Quote:
I have a ~10 yr old desktop that’s had a problem staying powered on after an outage.
Was there a power outage and the problem started afterwards?

I'd suspect either a ram problem of a dodgy power supply. If the problem started after a general power outage, a power spike when power resumed may have damaged the psu.

Quote:
BIOS settings completely resets every time this happens.
That would be indicative of a bad cmos battery, especially on a box that's 10 years old. You can try replacing it. They are usually 2032 3 volt batteries and are fairly inexpensive. If the cmos battery is bad, the BIOS will reset every time you unplug the box. The battery provides power to the motherboard to hold the BIOS settings when line power is removed. The fact that the BIOS is resetting after every hard shutdown would indicate that power from the psu to the motherboard is lost on shutdown and your cmos battery is bad causing a BIOS reset. That strongly suggests a problem with the psu.
 
1 members found this post helpful.
Old 01-28-2022, 01:59 PM   #4
onebuck
Moderator
 
Registered: Jan 2005
Location: Central Florida 20 minutes from Disney World
Distribution: Slackware®
Posts: 13,961
Blog Entries: 46

Rep: Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188Reputation: 3188
Member Response

Hi,

I also agree that the PSU could be the culprit. Along with the CMOS battery being bad and should be replaced. 10 years for cmos battery is far to long for dependable use.

Do you have a replacement PSU that meets the present PSU requirements that could be swapped in to see if that solves your issue(s)?

If you have DVM to test the levels while still connect to the system may show poor levels but if the switching/regulator circuits are failing then a DVM would not show this unless a major swing of levels.
You can look at this article to help with testing the PSU; https://www.techdim.com/test-computer-power-supply/

If you happen to have had a major signal/spikes through the PSU then damage to the system could be the issue.

If you happen to find the PSU is OK then be sure to test the memory.

Hope this helps.
 
1 members found this post helpful.
Old 01-28-2022, 03:35 PM   #5
Livestradamus
Member
 
Registered: Feb 2006
Location: Planet Earth (currently)
Distribution: Slackware & OpenBSD
Posts: 132

Original Poster
Rep: Reputation: 65
Quote:
Originally Posted by business_kid View Post
Sounds like the M/B or ram is blown.

Take the ram out of slots 0 & 2. Put the remaining ram into slots 0 & 1, and see will it boot and stay going with half the ram.
I tested all of the RAM on another desktop, they're perfectly fine
The problem remains using removing from the 0 & 2 slots and when using in 0 & 1. I guess its mobo or psu.


Quote:
Originally Posted by kilgoretrout View Post
Was there a power outage and the problem started afterwards?

I'd suspect either a ram problem of a dodgy power supply. If the problem started after a general power outage, a power spike when power resumed may have damaged the psu.
Thanks!


That would be indicative of a bad cmos battery, especially on a box that's 10 years old. You can try replacing it. They are usually 2032 3 volt batteries and are fairly inexpensive. If the cmos battery is bad, the BIOS will reset every time you unplug the box. The battery provides power to the motherboard to hold the BIOS settings when line power is removed. The fact that the BIOS is resetting after every hard shutdown would indicate that power from the psu to the motherboard is lost on shutdown and your cmos battery is bad causing a BIOS reset. That strongly suggests a problem with the psu.
RAM are OK. Thank you it must be the psu. I ran 'watch sensors' and saw some weird voltage warning. Must be psu unless something additional on the mobo.
And yes the battery needs replacing.
Thanks!



Quote:
Originally Posted by onebuck View Post
Hi,

I also agree that the PSU could be the culprit. Along with the CMOS battery being bad and should be replaced. 10 years for cmos battery is far to long for dependable use.

Do you have a replacement PSU that meets the present PSU requirements that could be swapped in to see if that solves your issue(s)?

If you have DVM to test the levels while still connect to the system may show poor levels but if the switching/regulator circuits are failing then a DVM would not show this unless a major swing of levels.
You can look at this article to help with testing the PSU; https://www.techdim.com/test-computer-power-supply/

If you happen to have had a major signal/spikes through the PSU then damage to the system could be the issue.

If you happen to find the PSU is OK then be sure to test the memory.

Hope this helps.
I have nothing to test the psu, ran watch sensors saw weird voltage warnings so yeah most likely this. I had replaced the psu 2 yrs ago that was old and extremely low quality.
The psu I could find at the time was a 'foxytech' a friend gave me they had laying around (used it in a miner).
Its all starting to make sense now.

Thanks!
 
Old 01-28-2022, 03:36 PM   #6
business_kid
LQ Guru
 
Registered: Jan 2006
Location: Ireland
Distribution: Slackware, Slarm64 & Android
Posts: 17,094

Rep: Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483
If the desktop lasts 20 minutes it's hardly the PSU. There's a 'Power Good' line which means all power lines are more or less there, and without it, the PC won't boot. Yes they can always go, but my guess this time is the ram. Unless you're saddled with a crazy power factor on your mains, the main effect of a power outage is power loss, not spikes.
 
Old 01-28-2022, 04:04 PM   #7
Livestradamus
Member
 
Registered: Feb 2006
Location: Planet Earth (currently)
Distribution: Slackware & OpenBSD
Posts: 132

Original Poster
Rep: Reputation: 65
Quote:
Originally Posted by business_kid View Post
If the desktop lasts 20 minutes it's hardly the PSU. There's a 'Power Good' line which means all power lines are more or less there, and without it, the PC won't boot. Yes they can always go, but my guess this time is the ram. Unless you're saddled with a crazy power factor on your mains, the main effect of a power outage is power loss, not spikes.
Interesting. I’ll test the RAM more. I wonder if it is the motherboard maybe where l insert the RAM sticks?
 
Old 01-28-2022, 05:14 PM   #8
obobskivich
Member
 
Registered: Jun 2020
Posts: 608

Rep: Reputation: Disabled
If there's absolutely no physical interaction that's happened here* I would look at the PSU or the motherboard. Just because it runs for 20 minutes does not excuse the PSU - it could be overheating or getting overloaded in that time, and when you're talking about power surges/outages I'm inclined to start looking at primary power components first (which is the PSU and the motherboard).

* If there was physical interaction it may be the CPU or CPU socket as well - losing memory channels is often a symptom of damaged contact between the CPU (and its IMC) and the rest of the system. I'm making an assumption this machine is <12 years old and thus has an IMC.
 
Old 01-29-2022, 01:38 AM   #9
mrmazda
LQ Guru
 
Registered: Aug 2016
Location: SE USA
Distribution: openSUSE 24/7; Debian, Knoppix, Mageia, Fedora, others
Posts: 6,245
Blog Entries: 1

Rep: Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173
Open up the PSU and look for electrolytic caps whose tops are no longer flat, or that are obviously oozing. If you find any, you can be sure the PSU isn't up to spec any more. With a name like "foxytech", I wouldn't expect top quality or long life.
 
Old 01-29-2022, 11:50 AM   #10
obobskivich
Member
 
Registered: Jun 2020
Posts: 608

Rep: Reputation: Disabled
Quote:
Originally Posted by mrmazda View Post
Open up the PSU and look for electrolytic caps whose tops are no longer flat, or that are obviously oozing. If you find any, you can be sure the PSU isn't up to spec any more. With a name like "foxytech", I wouldn't expect top quality or long life.
Do note that opening up an SMPS can be very dangerous, as those bulk caps can hold quite a charge - don't do this if you're not at least somewhat experienced with electronics (I do mean electronics, not computers). You can probably see the bulk caps inside most PSUs by peering through the fan opening with a flashlight, at least well enough to look for bulging caps, but that wouldn't be the only potential failure mode if you're dealing with a power event.
 
Old 01-29-2022, 12:51 PM   #11
mrmazda
LQ Guru
 
Registered: Aug 2016
Location: SE USA
Distribution: openSUSE 24/7; Debian, Knoppix, Mageia, Fedora, others
Posts: 6,245
Blog Entries: 1

Rep: Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173
Quote:
Originally Posted by obobskivich View Post
Do note that opening up an SMPS can be very dangerous, as those bulk caps can hold quite a charge - don't do this if you're not at least somewhat experienced with electronics (I do mean electronics, not computers).
Opening a PSU up to inspect needn't involve touching its components. I've been routinely opening PSUs and replacing bad caps for over decade without detecting such a charge. Apparently, the time it takes to extract and open an unplugged PSU from a PC is enough for retained charge to bleed down. Motherboards and graphic cards have electrolytic caps too. Nobody ever warns not to touch or replace a motherboard or GPU.
 
Old 01-29-2022, 01:14 PM   #12
Arnulf
Member
 
Registered: Jan 2022
Location: Hanover, Germany
Distribution: Slackware
Posts: 294

Rep: Reputation: 100Reputation: 100
You can touch 24 VDC (+12 V against −12 V) at maximum inside a computer. Therefore a computer can be an "open system".
You can touch 325 VDC (230 V · √2 against ground – in Europe; value depends on line voltage in your region, general: line_voltage · √2 against ground) at maximum inside a switching PSU like AT- and ATX-PSUs. Therefore a switching PSU is a "closed system".
 
1 members found this post helpful.
Old 01-29-2022, 02:20 PM   #13
business_kid
LQ Guru
 
Registered: Jan 2006
Location: Ireland
Distribution: Slackware, Slarm64 & Android
Posts: 17,094

Rep: Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483Reputation: 2483
On 220VAC, you can often find charged capacitors with ≤325V across the pins. Usually it's a bit lower, but not much.

All CE approved soldering irons have earthed tips nowadays, so you find that sort of thing out early on. Even Neutral --> earth faults trip the RCDs fitted in most places.
 
1 members found this post helpful.
Old 01-29-2022, 03:33 PM   #14
obobskivich
Member
 
Registered: Jun 2020
Posts: 608

Rep: Reputation: Disabled
Quote:
Originally Posted by mrmazda View Post
Opening a PSU up to inspect needn't involve touching its components. I've been routinely opening PSUs and replacing bad caps for over decade without detecting such a charge. Apparently, the time it takes to extract and open an unplugged PSU from a PC is enough for retained charge to bleed down. Motherboards and graphic cards have electrolytic caps too. Nobody ever warns not to touch or replace a motherboard or GPU.
Just because you have been safe/not had problems does not mean 'throw out all discussion of caution for everyone.' There's also a difference, as arnwulf and business_kid, point out, between the voltages and context of a PCI add-in card, and a power supply that connects to AC mains. Nowhere did I say 'this can never be done' but the "refer servicing to qualified professionals" tag is there for a reason - maybe you are "qualified professionals" and that's great, but not everyone is.
 
Old 01-29-2022, 05:38 PM   #15
mrmazda
LQ Guru
 
Registered: Aug 2016
Location: SE USA
Distribution: openSUSE 24/7; Debian, Knoppix, Mageia, Fedora, others
Posts: 6,245
Blog Entries: 1

Rep: Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173Reputation: 2173
One doesn't need to be an experienced tech to manage to not touch component innards. The reason for that kind of language is because of lawyers and corporate lawsuit paranoia, not because of serious risk.
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
[SOLVED] Centos 6 boot problem after power outage timl Linux - Software 4 01-27-2016 05:53 PM
Network card problem after power outage robrecc Slackware 13 06-11-2006 01:12 AM
Some machines won't ping www after power outage itsjustme Linux - Networking 1 05-13-2004 05:21 PM
Repair Questions after Power Outage Cynthia Blue Linux - Newbie 1 12-28-2003 12:31 PM
XServer not starting after power outage jcksrobbins Linux - General 2 08-24-2003 03:51 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware

All times are GMT -5. The time now is 03:59 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration