Personal computing discussed

Moderators: Flying Fox, Thresher

 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

weird issue would like your input

Fri Apr 20, 2018 8:57 am

my machine:
crucial 650watt
gigabyte z68xp-ud4 (v1.3) (I upgraded to the latest uefi firmware from 2013 when I got the machine 2 years ago)
crucial memory 16gig (4x4gb dimms)
2600k
Radeon rx580 8gb.

I noticed over the weekend that I was no longer using the 1600mhz memory setting and had dropped to 1333 and my overclock had been reset to default. I had fired up cpu-z instead of gpu-z by mistake since I wanted to monitor video card temps while playing games.
So I shut the machine off to get to the bios and would not boot back up.
I did the usual trouble shooting reseated the front panel switch, mainboard power, and power cycled the power supply and it finally came back up, I shurgged and played skyrim for 4 hours at the default settings and then shut it down.

Wed it would not boot, I held in the power button for 20 seconds and hit the button and it fired up for 5 seconds and shut off.
I waited 10 more seconds and tried again and it stated the bios was corrupt and it was restoring from backup using dual bios.
It completed and rebooted and then shut off again. This repeated 3 more times and then ran out of time to mess with it.

Last night I fought with it some more and cleared the cmos and each time it would repeat the same process and restore the corrupted bios.
I replaced the mainboard battery and cleared the cmos again same thing. on the third go round it suddenly said the cmos had been cleared and did I want to load the optimized defaults.
I went with that and it booted.

I cold rebooted a few more times setup by bios options, set the xmp profile for memory (1600mhz and recommended timings), light over clock using boost (I mean turbo) back to 4.5ghz for a single core, 4.3ghz for 2 cores and 4ghz for all 4 leaving the default speeds and settings intact. I disabled the onboard video as well.

Warm and Cold rebooted a couple more times it worked fine every time.
I also fired up occt and some other stability tools and everything checks out, voltages are right, temps get no higher than 85c, machine seems 100% stable.

Anyone have any ideas what caused it to freak out? power supply? Mainboard? cmos battery?
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
just brew it!
Gold subscriber
Administrator
Posts: 51659
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: weird issue would like your input

Fri Apr 20, 2018 9:08 am

Until I got towards the end of your post I was going to say "PSU", but I guess that has been (mostly) ruled out.

Could've been a dead CMOS battery was the original trigger (if you still have the old battery and have a multimeter, check its voltage). Also sounds like whatever you were doing to clear the CMOS simply didn't work the first 2 times.

If it starts happening again, I would start suspecting the PSU.
Nostalgia isn't what it used to be.
 
MileageMayVary
Gerbil Team Leader
Posts: 290
Joined: Thu Dec 10, 2015 9:18 am
Location: Baltimore

Re: weird issue would like your input

Fri Apr 20, 2018 9:23 am

Can you verify that the PSU fan is spinning?
Main rig: Ryzen 1600, R9 290@1100MHz, 16GB@2933MHz, 1080-1440-1080 Ultrasharps.
 
Chrispy_
Maximum Gerbil
Posts: 4426
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: weird issue would like your input

Fri Apr 20, 2018 9:57 am

just brew it! wrote:
Until I got towards the end of your post I was going to say "PSU", but I guess that has been (mostly) ruled out.

Could've been a dead CMOS battery was the original trigger (if you still have the old battery and have a multimeter, check its voltage). Also sounds like whatever you were doing to clear the CMOS simply didn't work the first 2 times.

If it starts happening again, I would start suspecting the PSU.


Seen this a dozen time on old Sandy machines. The batteries are always the problem, I stopped even trying to diagnose them until I'd changed the battery because of the completely erratic behaviour.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Fri Apr 20, 2018 11:35 am

MileageMayVary wrote:
Can you verify that the PSU fan is spinning?


Actually... no.

I think I should really double check. I'm not sure if it blows down through the bottom of the case or out the back. And I don't remember there being dust on the dust filter. There is 140mm fan right next to it using the same filter and that has dust on its location.
I'll take a look tonight or over the weekend. Last thing I need is the PS suddenly blowing.
Last edited by Aranarth on Fri Apr 20, 2018 11:40 am, edited 1 time in total.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Fri Apr 20, 2018 11:39 am

Chrispy_ wrote:
just brew it! wrote:
Until I got towards the end of your post I was going to say "PSU", but I guess that has been (mostly) ruled out.

Could've been a dead CMOS battery was the original trigger (if you still have the old battery and have a multimeter, check its voltage). Also sounds like whatever you were doing to clear the CMOS simply didn't work the first 2 times.

If it starts happening again, I would start suspecting the PSU.


Seen this a dozen time on old Sandy machines. The batteries are always the problem, I stopped even trying to diagnose them until I'd changed the battery because of the completely erratic behaviour.


Thanks JBI, That was really weird and I was starting to suspect the mainboard and then it suddenly worked and everything was fine.

Thanks Chrispy! I suspected the battery but it was such a weird thing I wanted to double check if anyone else has seen it.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
MileageMayVary
Gerbil Team Leader
Posts: 290
Joined: Thu Dec 10, 2015 9:18 am
Location: Baltimore

Re: weird issue would like your input

Fri Apr 20, 2018 2:09 pm

Aranarth wrote:
MileageMayVary wrote:
Can you verify that the PSU fan is spinning?


Actually... no.

I think I should really double check. I'm not sure if it blows down through the bottom of the case or out the back. And I don't remember there being dust on the dust filter. There is 140mm fan right next to it using the same filter and that has dust on its location.
I'll take a look tonight or over the weekend. Last thing I need is the PS suddenly blowing.


An former machine of my brother's would crash extremely random . Couldn't figure it out. Then I burned my finger on the PSU. Unscrewed the thing and flipped it over, the fan wouldn't spin during POST or during any amount of load. Replaced that and he was good to go.

When I say burned, only minorly, but it's shell was about as hot as the hottest water from the tap so... 120-130ºf?
Main rig: Ryzen 1600, R9 290@1100MHz, 16GB@2933MHz, 1080-1440-1080 Ultrasharps.
 
just brew it!
Gold subscriber
Administrator
Posts: 51659
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: weird issue would like your input

Fri Apr 20, 2018 2:14 pm

Back in the day, the fan in the PSU of my dual Athlon (MPX) rig failed. I started to smell burning electronics, then noticed that the outside of the PC case (over the PSU, this was an old-style case with top PSU mount) was quite hot. Immediately shut the system down and swapped in a spare PSU...
Nostalgia isn't what it used to be.
 
The Egg
Gold subscriber
Minister of Gerbil Affairs
Posts: 2367
Joined: Sun Apr 06, 2008 4:46 pm

Re: weird issue would like your input

Fri Apr 20, 2018 3:13 pm

I'm actually glad I saw this thread. I'm getting ready to hand down my 2500k Sandy system to my parents, and it still has the original CMOS battery from 2011. Gotta say, this is the first time I've used something as my main rig for so long that the battery was due for replacement.
 
JustAnEngineer
Gold subscriber
Gerbil God
Posts: 18259
Joined: Sat Jan 26, 2002 7:00 pm
Location: The Heart of Dixie

Re: weird issue would like your input

Fri Apr 20, 2018 3:18 pm

just brew it! wrote:
Back in the day, the fan in the PSU of my dual Athlon (MPX) rig failed. I started to smell burning electronics...
Officially, there are no user serviceable parts and the capacitors inside the PSU might kill you. Unofficially... :wink:
i7-8700K, H100i v2, RoG Strix Z370-G Gaming, 16 GiB, RX Vega64, Define Mini-C, SSR-850PX, C32HG70, RK-9000BR, MX518
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Mon Apr 23, 2018 6:54 am

The machine is still doing it.

My usual over clock of 4.5ghz is no longer stable and causing the machine the freakout.
The weird thing is the cmos clear is not working either.
I have to pull the battery and power cord for a minimum of an hour before it comes back and lets me use the "optimized defaults" which are definitely NOT optimized for anything later than XP.

Machine is completely stable at default frequencies and sometimes I can get it to boot 4.5 ghz on one core. (eist)

I guess I'll just use the xmp memory profile and default processor speed.
The machine is plenty fast enough for now and I don't have the money for a new machine anyway.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Mon Apr 23, 2018 6:54 am

oh and before I forget, yes I can confirm the PSU fan is turning! :D
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Chrispy_
Maximum Gerbil
Posts: 4426
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: weird issue would like your input

Mon Apr 23, 2018 7:59 am

Has it been overclocked at 4.5GHz its entire life?

If the PSU is known to be good, even under load, the next most probably scenario is that your CPU is dying. That's one of the risks that comes with overclocking - increased voltage and clockspeeds will have exponentially higher electromigration effects on the silicon. If you were "auto" overclocking, this would make it even worse, because they typically apply more voltage than necessary. Disabling speedstep because of your overclock also means that the CPU never gets to downclock and cut back on the voltage that does the most damage.

  • At stock, I'd expect Sandy to last for 12+ years, based on some very rough extrapolation of overclocked endurance for older chips that were running 24/7
  • Light overclock (1.25V or lower) - maybe 8 years? I don't have any experience/knowledge of these failing through electromigration yet.
  • Heavy overclock (1.4V or higher) - 3-5 years. Failures of 5GHz Sandys are common knowledge among the 5GHz club and have been for a few years, unfortunately.

These all assume an 8h working day duty cycle. If you leave your PC on overnight that will obviously add to the duty cycle, and it depends whether your overclock leaves Speedstep enabled or whether it's a full-time overclock - with peak voltage running all the time too.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
The Egg
Gold subscriber
Minister of Gerbil Affairs
Posts: 2367
Joined: Sun Apr 06, 2008 4:46 pm

Re: weird issue would like your input

Mon Apr 23, 2018 8:48 am

I'm not a fan of Gigabyte boards lately. My parents FM2 setup randomly reboots since installing W10, and follows no pattern nor gives any indication as to why (it was solid under W7, so I can only guess drivers, though I've tried different versions of most everything).

More recently, my nephew bought a Z77 Socket 1155 board off eBay (without asking my opinion first), and the thing is flaky as all getout (go figure). It needed a BIOS update before it would function with his discrete 1050, and unbeknownst to me, his 2550k is one of the few chips without onboard graphics. That was fun to diagnose. Then after fighting with multiple different brands/speeds of RAM, I came to find that the failsafe defaults for the Gigabyte BIOS is to enable automatic overclocking by default. Very quietly in a buried setting. :evil:

In short, it might be a bit amateurish to blame the board manufacturer, but I'd like to take a moment to blame your board manufacturer.
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Mon Apr 23, 2018 12:49 pm

Chrispy_ wrote:
Has it been overclocked at 4.5GHz its entire life?

If the PSU is known to be good, even under load, the next most probably scenario is that your CPU is dying. That's one of the risks that comes with overclocking - increased voltage and clockspeeds will have exponentially higher electromigration effects on the silicon. If you were "auto" overclocking, this would make it even worse, because they typically apply more voltage than necessary. Disabling speedstep because of your overclock also means that the CPU never gets to downclock and cut back on the voltage that does the most damage.

  • At stock, I'd expect Sandy to last for 12+ years, based on some very rough extrapolation of overclocked endurance for older chips that were running 24/7
  • Light overclock (1.25V or lower) - maybe 8 years? I don't have any experience/knowledge of these failing through electromigration yet.
  • Heavy overclock (1.4V or higher) - 3-5 years. Failures of 5GHz Sandys are common knowledge among the 5GHz club and have been for a few years, unfortunately.

These all assume an 8h working day duty cycle. If you leave your PC on overnight that will obviously add to the duty cycle, and it depends whether your overclock leaves Speedstep enabled or whether it's a full-time overclock - with peak voltage running all the time too.


Before I bought it off of ebay I asked the previous owner who built it if it was overclocked and was told "no it was not" it was used for graphics and video editing only.
Who knows if that is the truth or not.
I overclocked myself I think I was using 1.20v. I can't find my notes. It ran fine like that for 3 years.
Overclock was all cores and fixed cpu speed. Temps hovered around 60c when playing games.

Changes made to the machine:
ocz triton 500gb ssd
good xigmatek after market cpu fan instead of intel.
new video card: came with radeon 5950 and upgraded to 7870 to rx 580 8gb


Sure I'd like my nice overclock back, Shrug, I'll just go back to a stable speed and go with that.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Mon Apr 23, 2018 12:52 pm

The Egg wrote:
In short, it might be a bit amateurish to blame the board manufacturer, but I'd like to take a moment to blame your board manufacturer.


I'll join you with leaving a flaming bag of AHEM on their front doorstep. :D

on the other hand my overclocked q6600 at 3.0ghz is running just fine and that is also in a gigabyte board. :D
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Chrispy_
Maximum Gerbil
Posts: 4426
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: weird issue would like your input

Mon Apr 23, 2018 1:25 pm

I think it's fair to say that all motherboard vendors make good models and bad models. The difference is in how much you have to pay to get a good one with each vendor ;)
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Wed Apr 25, 2018 7:23 am

I'm still wondering if it is the P/S...

It just did it again.
I can also reproduce the fault which is nice: Boot the machine, Do nothing much for 2 mins, shut the machine off, give it 10 seconds and boot it back up and boom it won't boot or gives a bios corruption error. Just like you would be doing if you were attempting overclocks and wanted to see if the machine would boot. Pull the battery and wait a couple hours and then load bios defaults and you are back in business.

If the machine is running it is fine.
If you shut the machine down and don't boot it backup for a couple of hours it is fine.

I don't think it is the M/B or it would be unstable while it is on.
I believe JBI is correct and the P/S is suspect. It seems to be a temperature issue cracked solder joint or something. Luckily a new Seasonic P/S is only $80.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Chrispy_
Maximum Gerbil
Posts: 4426
Joined: Fri Apr 09, 2004 3:49 pm
Location: Europe, most frequently London.

Re: weird issue would like your input

Wed Apr 25, 2018 7:34 am

It's starting to sound like a C7 sleep state issue that is effectively a compatibility problem between your OS, PSU and BIOS.

If your PSU was flaky, it would be fine at idle (boot/POST) and the fault would manifest at high power draw - which you've confirmed is not the case because it'll game for 4+ hours.

Have your run a simultaneous Furmark and Prime95 on all threads, yet? (or the built-in CPU burner in Furmark?). Stability for more than 1h at full CPU+GPU load usually certifies the PSU as rock-solid.
Congratulations, you've noticed that this year's signature is based on outdated internet memes; CLICK HERE NOW to experience this unforgettable phenomenon. This sentence is just filler and as irrelevant as my signature.
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Wed Apr 25, 2018 11:48 am

Chrispy_ wrote:
It's starting to sound like a C7 sleep state issue that is effectively a compatibility problem between your OS, PSU and BIOS.

If your PSU was flaky, it would be fine at idle (boot/POST) and the fault would manifest at high power draw - which you've confirmed is not the case because it'll game for 4+ hours.

Have your run a simultaneous Furmark and Prime95 on all threads, yet? (or the built-in CPU burner in Furmark?). Stability for more than 1h at full CPU+GPU load usually certifies the PSU as rock-solid.


I'll give it a try.
I can also try turning off all power saving features and doing a mild over clock and see if that sorts it out.
Drives me nuts that the machine was running fine for years and then suddenly it has quirks.

I want a machine I can turn on an just game on for a few hours a week.
I rarely have time for anything else between work, the farm, and kids.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Fri Apr 27, 2018 8:20 am

Aha!! found it!
It is the xmp memory profile.
I went with the default 1333mhz instead of 1600mhz and it works just fine.
Over clocks to 4.5ghz the whole bit.
Turn it back on after a cmos clear and the machine boots to windows.
Cold reboot and the machine freaks out and you get the corrupt cmos error.

Something is causing the memory to fail during post when the xmp profile is turned on.
I'm positive I had it turned on before it started freaking out.

Whelp at least the memory is not running at 533mhz any more.... :D
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD
 
just brew it!
Gold subscriber
Administrator
Posts: 51659
Joined: Tue Aug 20, 2002 10:51 pm
Location: Somewhere, having a beer

Re: weird issue would like your input

Fri Apr 27, 2018 9:21 am

It is stories like this which reinforce my decision (made approximately a decade ago) to run stock CPU and RAM clocks on all systems which aren't explicitly "screw around" or "smoke test" boxes. I don't want to be dealing with weird behaviors like this when I'm trying to use the system for something else!
Nostalgia isn't what it used to be.
 
Aranarth
Graphmaster Gerbil
Topic Author
Posts: 1130
Joined: Tue Jan 17, 2006 6:56 am
Location: Big Rapids, Mich. (Est Time Zone)
Contact:

Re: weird issue would like your input

Fri Apr 27, 2018 10:03 am

just brew it! wrote:
It is stories like this which reinforce my decision (made approximately a decade ago) to run stock CPU and RAM clocks on all systems which aren't explicitly "screw around" or "smoke test" boxes. I don't want to be dealing with weird behaviors like this when I'm trying to use the system for something else!


If you ask the wife, this IS my screw around machine...
I only use it for playing games! :D

I had it running with eist on and multipliers set to 45, 43, 40, 40. Then I ran occt for an hour at and it chose to run at 3.7ghz (I did see all four multipiers being used during spool up etc.).
I ran out of time for testing but it also booted with dividers set to 45, 43, 43, 41 and occt ran for a few minutes at 4.1ghz.

Max cpu temp was 55c during the hour long test.
Main machine: Core I7 -2600K @ 4.0Ghz / 16 gig ram / Radeon RX 580 8gb / 500gb toshiba ssd / 5tb hd
Old machine: Core 2 quad Q6600 @ 3ghz / 8 gig ram / Radeon 7870 / 240 gb PNY ssd / 1tb HD

Who is online

Users browsing this forum: No registered users and 1 guest