LinuxQuestions.org
Welcome to the most active Linux Forum on the web.
Home Forums Tutorials Articles Register
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 09-26-2022, 05:23 PM   #1
haertig
Senior Member
 
Registered: Nov 2004
Distribution: Debian, Ubuntu, LinuxMint, Slackware, SysrescueCD, Raspbian, Arch
Posts: 2,331

Rep: Reputation: 357Reputation: 357Reputation: 357Reputation: 357
Opinions? Apparent computer thermal shutdown at 48 Celsius???


I am apparently having thermal problems in my computer. It appears to be shutting down due to heat, but way below where that should normally occur. I am seeing a shutdown at a CPU temp of 48 Celsius.

This is a new thread, with new info, that was originally:

https://www.linuxquestions.org/quest...se-4175716525/

Since that other thread I have identified the problem as apparently being heat related.

This system has always run cool. Normal, routine use of the computer has the CPU around 32 Celsius, and when idling usually around 30 degrees. Ambient room temperature is about 22 degrees. All fans are operational (CPU fan, plus four case fans). System is dust free.

On to the testing...

I fired up the psensor application (it uses lmsensors data to draw graphs). 33 degrees CPU. I then started to transcode a video. I watched psensors graph the CPU temp increase to 48 where it leveled off. But the system shut down after being at 48 degrees for about 30 seconds. This is repeatable.

CPU is an AMD Phenom II x6 1045T 2.7GHz, which should have no problem going over 70 Celsius for reasonably short periods. Motherboard is a Gigabyte GA-MA78G-DS3H ATX, it is not overclocked. Cooling is OEM heatsink and fan. Thermal paste was applied correctly at installation, but has not been replaced for years - because there has never been any overheating, including while transcoding videos - and I always watch the psensors graph while doing that.

BIOS is set to "alarm" at a CPU temp of 70 Celsius. psensors program is set to warn at 55 degrees, but never gets to do that before the system shuts down.

The only thing I had done to this system prior to these shutdowns starting was to replace the memory. 4Gb Kingston ram was upgraded to 8Gb Kingston ram. Total memory replacement, no mix-and-match between old-and-new.

As part of my initial testing - before I determined that this appears to be heat related - I replaced the power supply (made no difference) and cleaned and reseated the memory modules (that seemed to make the shutdowns less frequent, but I think that was simply coincidence). More recent testing video-trancsode->shutdown, transcode->shutdown, transcode->shutdown seem to confirm a thermal issue as far as I can determine.

So why would my system be shutting down at the low temp of 48 Celsius? That is an expected temperature during video transcoding in my experience, and is not terribly high at all. It shouldn't be in the "thermal shutdown" range. The highest temp I can remember ever seeing for my CPU was 55 or 56 Celsius, which was during a video transcode. When that happened a few years ago, I used some Dust-Off to clean everything out (heatsinks, etc.) and the video transcoding temp dropped to a slightly lower 52 Celsius after that, if I am recalling the specifics correctly.

p.s. - When I start the computer and go into BIOS, that is reporting my CPU temp to be the same as psensors is reporting the temp after a full boot, which tells me that psensors is not out of calibration and is reporting temps accurately.

Last edited by haertig; 09-26-2022 at 05:28 PM.
 
Old 09-26-2022, 05:35 PM   #2
haertig
Senior Member
 
Registered: Nov 2004
Distribution: Debian, Ubuntu, LinuxMint, Slackware, SysrescueCD, Raspbian, Arch
Posts: 2,331

Original Poster
Rep: Reputation: 357Reputation: 357Reputation: 357Reputation: 357
From a post in the old thread:

Quote:
Originally Posted by Emerson View Post
How do you measure this 48° C? You sure it is not a bogus reading and it is the actual temperature of CPU?
I'm using the psensors app, which gets data from lmsensors, and the lmsensors value for CPU temp matches the BIOSes value for CPU temp.
 
Old 09-26-2022, 05:40 PM   #3
haertig
Senior Member
 
Registered: Nov 2004
Distribution: Debian, Ubuntu, LinuxMint, Slackware, SysrescueCD, Raspbian, Arch
Posts: 2,331

Original Poster
Rep: Reputation: 357Reputation: 357Reputation: 357Reputation: 357
I have out of town guests coming that I need to get ready for, but once I have more free time in a few days, my plan is to pop off the CPU heatsink, clean all old thermal paste off of the CPU and the heatsink, and reapply new thermal paste.

BTW, how long does thermal paste last once you've opened and re-capped the tube? The newest stuff I've got is about five years old. Is that still good?
 
Old 09-26-2022, 06:29 PM   #4
Emerson
LQ Sage
 
Registered: Nov 2004
Location: Saint Amant, Acadiana
Distribution: Gentoo ~amd64
Posts: 7,661

Rep: Reputation: Disabled
What I thought was perhaps this 48° C is coming from socket sensor, but actual CPU temperature is higher. My Intel has sensors on the die, actually four sensors, one for each core.

Edit for clarity: I think your stored thermal paste is still good, it is not supposed to cure over time.

Last edited by Emerson; 09-26-2022 at 07:09 PM.
 
Old 09-27-2022, 08:59 AM   #5
beachboy2
Senior Member
 
Registered: Jan 2007
Location: Wild West Wales, UK
Distribution: Linux Mint 21 MATE, EndeavourOS, antiX, MX Linux
Posts: 3,972
Blog Entries: 32

Rep: Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465
haertig,

As long as the thermal paste has been sealed and not exposed to the open air, it should be okay.

To be absolutely sure, buy the cheapest tube of Arctic MX-4 you can find. It is often supplied free with new CPUs from some retailers.

Get some lint free cloth (such as for use with spectacles), some isopropyl alcohol and thoroughly clean the mating surfaces on the CPU and cooler.

You may need to let the isopropyl alcohol soak in for a short while before the old paste can be completely removed. Go gently to avoid scratching the top of the CPU!

NB You only need a tiny amount (≡ 1.5 rice grains) of thermal paste in the centre of the CPU.

Having said all that, from what you have described, I don’t think this is a heat-related problem.
Your PC has been running perfectly for a long time.
You have already ruled out the PSU as a possible culprit.

Quote:
Originally Posted by haertig View Post
The only thing I had done to this system prior to these shutdowns starting was to replace the memory. 4Gb Kingston ram was upgraded to 8Gb Kingston ram. Total memory replacement, no mix-and-match between old-and-new.
My suspicion would be the new Kingston memory. Personally I would not touch it because of its poor reputation. Others may disagree. I only use Crucial or Corsair memory and over many years I have only had a single stick of Crucial RAM go bad (replaced under lifetime warranty).

If you happen to know a good techie at a local computer shop, you could possibly borrow some suitable functioning used RAM and see whether that makes a difference.

Good luck.

Last edited by beachboy2; 09-27-2022 at 09:23 AM.
 
Old 09-27-2022, 09:22 AM   #6
boughtonp
Senior Member
 
Registered: Feb 2007
Location: UK
Distribution: Debian
Posts: 3,599

Rep: Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546Reputation: 2546
Quote:
Originally Posted by beachboy2 View Post
NB You only need a tiny amount (≡ 1.5 rice grains) of thermal paste in the centre of the CPU.
Rice gains come in different shapes and sizes.

The thermal paste I have recommends a single 3-4mm dot in the centre for small CPUs, whilst for for mid-sized CPUs it recommends that plus four 2mm dots in each corner.


Quote:
As long as the thermal paste has been sealed and not exposed to the open air, it should be okay.
Mine states "Recommended storage time (before use) up to 3 years" ?

Seems particularly odd when it has "Recommended usage time (on the CPU) up to 5 years" - why is CPU time longer?

 
Old 09-27-2022, 02:53 PM   #7
Arnulf
Member
 
Registered: Jan 2022
Location: Hanover, Germany
Distribution: Slackware
Posts: 268

Rep: Reputation: 89
Quote:
Originally Posted by haertig View Post
CPU is an AMD Phenom II x6 1045T

[…]
Motherboard is a Gigabyte GA-MA78G-DS3H ATX, it is not overclocked. Cooling is OEM heatsink and fan. Thermal paste was applied correctly at installation, but has not been replaced for years
Your system is quite old. Aging reduces allowed maximum operating temperature of electronic components. Aging may dry out thermal grease and thermal pads whereby they become brickle and their thermal conductities decrease. Therefore cool down your system:
  • Replace any thermal grease/thermal pads with fresh thermal grease with a thermal conductivity of about 9 W·m⁻¹·K⁻¹ or more at northbridge and southbridge on mainboard and GPU on graphics card. Clean all parts with isopropyl alcohol before applying new thermal grease.
  • Replace OEM CPU cooler with a sufficient top-down-cooler, e.g. "be quiet! Shadow Rock TF 2" or "Scythe Big Shuriken 3" to cool down VRMs and RAM modules too.
Quote:
Originally Posted by haertig View Post
The only thing I had done to this system prior to these shutdowns starting was to replace the memory. 4Gb Kingston ram was upgraded to 8Gb Kingston ram.
I've found many incompatibilities between DDR2 RAM modules and mainboards over the years. Especially dealing with 4 GiB DDR2 modules is quite tricky. Are the 8 GiB RAM realised with two 4 GiB modules or four 2 GiB modules? Have you run memtest86 or memtest86+ with the 8 GiB RAM?
 
Old 09-28-2022, 01:13 AM   #8
haertig
Senior Member
 
Registered: Nov 2004
Distribution: Debian, Ubuntu, LinuxMint, Slackware, SysrescueCD, Raspbian, Arch
Posts: 2,331

Original Poster
Rep: Reputation: 357Reputation: 357Reputation: 357Reputation: 357
The new memory is 4x2Gb = 8Gb total. The old memory was 4x1Gb = 4Gb total.

The original memory was Kingston KVR800D2N5. The new memory has a very similar model number: KVR800D2N6. I assume the difference between the N5 and the N6 in the model designations is the size of the memory stick (2Gb vs. 1Gb).

The reason I went for Kingston memory as a replacement was because basically, that's what you find available these days for DDR2. DDR2 is pretty old stuff. And you mostly have to get it direct from China. Off of eBay. In newer systems I have built, I have gone with either Corsair memory or G-Skill memory. But this old system I am dealing with now, it got what I could find. There were a few other choices besides Kingston, but I had never heard of them. Kingston may not be high end, but at least I had heard of it before. And indeed, this old system had been running for years on Kingston memory. It was only $16 for 8Gb of DDR2 on eBay. That's what I was willing to risk financially, ordering from China.

I ran MEMTEST on the new memory when I got it. For a few hours. Long enough to complete one full pass. If this was more expensive memory, or in one of my higher end systems, I would have run MEMTEST longer, but I called it good after one pass for this system. I did notice my system getting warm during that testing, just by feeling the temp that the fans were exhausting. Oddly enough, the system did not shutdown from heat during that MEMTEST. It wasn't burning up or anything. Normally, if I feel the temp of my exhaust fans it always feels cooler than ambient. Obviously it's not, but it feels that way due to the convection. When I was running MEMTEST, the exhaust felt slightly warm. Not hot. Just not cool like normal. But a rise in temps is expected during MEMTEST.
 
Old 09-28-2022, 02:57 AM   #9
beachboy2
Senior Member
 
Registered: Jan 2007
Location: Wild West Wales, UK
Distribution: Linux Mint 21 MATE, EndeavourOS, antiX, MX Linux
Posts: 3,972
Blog Entries: 32

Rep: Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465Reputation: 1465
haertig,

Hynix, Micron and Samsung are all good makers of PC 2 6400 ddr2 RAM.

Examples:

https://www.ebay.com/itm/37407437009...Bk9SR5Lo2svwYA

https://www.ebay.com/itm/37399335647...Bk9SR5Lo2svwYA

https://www.ebay.com/itm/29502232957...Bk9SR5Lo2svwYA

Last edited by beachboy2; 09-28-2022 at 10:38 AM.
 
Old 11-12-2022, 11:51 PM   #10
haertig
Senior Member
 
Registered: Nov 2004
Distribution: Debian, Ubuntu, LinuxMint, Slackware, SysrescueCD, Raspbian, Arch
Posts: 2,331

Original Poster
Rep: Reputation: 357Reputation: 357Reputation: 357Reputation: 357
I thought I would close out this thread. My computer is up and running again. But you really can't call it the "same computer" now - the only things left from the original are the case and a DVD burner. It's got a new motherboard, new CPU, new RAM, new SSD and a new power supply.

It didn't cost me much for this upgrade. The mobo (Gigabyte GA-Z87X-UD4H) was $100. The CPU (i7-4770k) was free. The 32Gb RAM (G-Skill RipJaws DDR3-1600) was free. The CPU cooler (OEM, new in box) was free. The 500Gb SSD (Crucial MX-500) was $42. The 650 watt power supply (EVGA Supernova P2 Platinum) was $69. So about $210 total for everything. I was lucky to have a friend who offered up an unused CPU and unused RAM he had laying around. I found a compatible motherboard on eBay. The initial goal was to fix a dead computer. But I ended up with a solid upgrade in the process, for not too much money.

As far as the original (thermal?) problem, the old computer kept getting worse and worse. More frequent shutdowns. Immediately after boot when the computer was cool as a cucumber. So the problem didn't appear to be thermal (leave the computer turned off for 24 hours, boot it, and it shut down within a few seconds of boot). I did replace the thermal paste with fresh stuff, but that didn't help. When I used to do video transcodes with the old computer (back when it wasn't shutting down), it would slowly climb to around 62 celcius max during the transcode. I tested this "new" computer doing a video transcode, and it only got up to 48 celcius (and I let it transcode for over half an hour). So it's much cooler than the old.

Last edited by haertig; 11-13-2022 at 12:01 AM.
 
  


Reply



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
What are the labels for thermal sensors listed in /proc/acpi/ibm/thermal mglinux Linux - Hardware 2 09-28-2020 06:41 PM
CPU fans: thermal pads and thermal paste; newbiesforever General 17 11-23-2009 10:22 PM
Installation problem FC4 on Fujitsu Celsius hofman22 Fedora - Installation 0 11-01-2005 06:48 AM
KWeather in Celsius? Seiken Linux - Software 2 03-10-2005 07:11 AM
Workaround: Suse Linux on Fujitsu-Siemens Celsius H mhollstein Linux - Laptop and Netbook 0 08-03-2004 08:21 AM

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

All times are GMT -5. The time now is 09:42 PM.

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