Temperature limits and fan control

Petirdo

Commendable
May 28, 2016
24
0
1,510
So I brought MSI GTX 1060 Gaming X 6GB, and i have some questions about temperature limits and how fan operates.
I will start with the problem i faced to make me come to the conclusion that my PC is overheating.
I have had this Graphics card for a week now, and during this week i faced one problem that my screen would freeze on the last picture it stayed on, no sounds, no power to keyboard or mouse, fans are working but it feels like the computer has shut down or something. All i can do is to hard shutdown my computer from my case power buttons, and when i open it again it boots normally. This happened randomly ingame and not ingame.
The first time it happened was when i left my computer open during night downloading something and my screen after 10 minutes shuts off to conserve power until someone clicks on the keyboard or moves the mouse. So, when i woke up i found the screen stuck on the last thing which was black screen and nothing was working.
At first i didn't understand what was happening or why, and then it happened again ingame.
Another time, after I hard reset my PC, it just wouldn't boot up again. I looked at my motherboard and i saw this blinking LED on DRAM. I looked it up and found that it means there is something wrong with my ram. So, i thought maybe that was the problem all along. However, when i changed the ram slot i was using before it booted up normally. So, i put it back in the old slot thinking that the slot was damaged, and to my surprise it booted up normally too. This made me think that it wasn't a ram problem, or a VGA problem, and it was a temperature problem. The motherboard didn't boot up until it cooled down.
Also, i tested my ram using memtest and made like 1000 loops with no errors.
So, i moved my PC to a room with air conditioning. Also, I configured the fan control curve in MSI Afterburner to suit my comfort zone. Before that, on idle my GPU's temperature would be at 60°C , now it is 45°C when the air conditioning is off, and 40°C when the air conditioning is on.
When i am ingame it jumps to 50-55 at most according to the game.
Also my CPU temperature is 41 on idle, and goes up to like 50 on load.
Taking into consideration, i don't have an intake fan, only one exhaust. I left the case door open when i moved it. The weather conditions are normally hot here like 35°C to 45°C in summer. Also, I don't always have a good airflow in the apartment.
The problem is not occuring anymore too.
So my questions are:
- Was it really an overheating problem? Is there some kind of mechanism that locks everything until it cools down for safety measures?
- Without the fan control curve i set on Afterburner, the GPU temperature would rise to 60°C even in the room without air conditioning on idle. Is this normal because it looks to me that it is hot?
- Also, how does the automatic fan works without any curves being set? Meaning that my GPU is at 60°C why doesn't the fans operate faster to cool down to 50°C for example? Could this be a faulty GPU or just settings in auto fan and that i should just stick to my fan control curve?
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
Was it really an overheating problem?
Probably not. Also, it's customary to post full specs when asking for advice.
Is there some kind of mechanism that locks everything until it cools down for safety measures?
Yes. It's called thermal throttling. It makes components slower to output less heat.
Is this normal because it looks to me that it is hot?
It's normal. Below 80c for most components is considered safe.
Meaning that my GPU is at 60°C why doesn't the fans operate faster to cool down to 50°C for example?
Lower noise and longer fan life. 20% increase in fan speed might lower temps by only 5c. So manufacturers try to keep it as low as possible until it becomes hot enough. The default curve should look like a half U. It starts slow but rises sharply toward the end.
-------------------
Your GPU BIOS states 92c as the max temp (it throttles at this temp) and 83c as the tested temp that it works at.
https://www.techpowerup.com/vgabios/185951/msi-gtx1060-6144-160630
-------------------
If you suspect GPU, test it at another system.
Also, open Control Panel > Event Viewer > Windows Logs. There should be errors if your PC keeps freezing.
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
Yes. It's called thermal throttling. It makes components slower to output less heat.
Can it lead to a freeze in some cases?
Lower noise and longer fan life. 20% increase in fan speed might lower temps by only 5c. So manufacturers try to keep it as low as possible until it becomes hot enough. The default curve should look like a half U. It starts slow but rises sharply toward the end.
I searched online and i understood that the manufacturer of the MSI Graphics card sets the fans to not work unless under load or above 60°C.

If you suspect GPU, test it at another system.
Also, open Control Panel > Event Viewer > Windows Logs. There should be errors if your PC keeps freezing.
Well till now the PC has been open for more than 48 hours and the problem is not occurring.
Also, i closed the side panel in my case, the temperature increased about 5°C but it is still working nicely. No freezing or anything.


 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
Can it lead to a freeze in some cases?
It can but mostly it lags and stutters. It should happen after 85c+.
I searched online and i understood that the manufacturer of the MSI Graphics card sets the fans to not work unless under load or above 60°C.
It's called fan stop. But it should be 50c+. 60c+ is a little high in my opinion.
Well till now the PC has been open for more than 48 hours and the problem is not occurring.
That's good.
Also, i closed the side panel in my case, the temperature increased about 5°C but it is still working nicely. No freezing or anything.
High CFM/Airflow (55+) fans can help with that.

BTW, most cases have fan position on the side for GPU. You could install a fan there.
 

Petirdo

Commendable
May 28, 2016
24
0
1,510

So, I used prime95 with core temp to monitor the behaviour of my core temps.
I ran small FFT test, and the temperature just boosted to 95°C and kept increasing till it reached 105°C. I stopped the test because i know that my T junction is 105°C
So what does that mean??
My CPU has been used for over 5 years now.
Should I apply thermal paste?
I am also considering buying a new case such as corsair carbide series 330R or obsidian series 550D, do you think it would help with the temperatures.


 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
My CPU has been used for over 5 years now.
Should I apply thermal paste?
If you have not changed the paste before, it will help a lot. Recommended to do it every 1-2 years.
I am also considering buying a new case such as corsair carbide series 330R or obsidian series 550D, do you think it would help with the temperatures.
Good fans + aftermarket CPU cooler should help more. Examples:
Don't know if it these fit. Check your case support. I can recommend smaller ones too.
https://es.pcpartpicker.com/product/cwPzK8/be-quiet-cpu-cooler-bk009
https://es.pcpartpicker.com/product/hJFPxr/deepcool-cpu-cooler-gammaxx400

https://es.pcpartpicker.com/product/kFBv6h/enermax-case-fan-ucmaa12a (If the numbers are correct, it's a great deal.)
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
Is the temperature rising to 105°C in a stress test normal using Prime95?
I was gonna buy the noctua 120mm or corsair for my current case, or just buy a whole new case all together.
I will give the thermal paste a shot.
Thank you for your help and patience mate ^_^
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
Is the temperature rising to 105°C in a stress test normal using Prime95?
On stock Intel heatsink with dried/old thermal paste + AVX, it's normal. Your game temps look much better.

Also, test with Prime95 v26.6 (Small FFT). Newer versions have AVX which outputs much more heat.
I was gonna buy the noctua 120mm or corsair for my current case
If you meant fans, you don't need to go that high-end. I bought 4 fans similar to this while they were on sale.
https://www.newegg.com/Product/Product.aspx?Item=N82E16835103060

CFM means airflow. You want 60+ for good circulation.
dB means noise. You want less than 25- for a quiet setup.
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
On stock Intel heatsink with dried/old thermal paste + AVX, it's normal. Your game temps look much better.
Also, test with Prime95 v26.6 (Small FFT). Newer versions have AVX which outputs much more heat.
Yeah i am afraid that it maybe the problem behind these recent lockups.
Ok i will try the v26.6 now and see where it goes.
I will get the thermal paste tomorrow and check the temps after that.

If you meant fans, you don't need to go that high-end. I bought 4 fans similar to this while they were on sale.
https://www.newegg.com/Product/Product.aspx?Item=N82E16835103060

CFM means airflow. You want 60+ for good circulation.
dB means noise. You want less than 25- for a quiet setup.
I don't mind buying it, it seems to be really good from reviews.
I will see my budget and which one is available in my country and buy it.
Thanks for the extra info
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
So i tested using Prime95 v26.6 (Small FFT).
The temperature increased slowly but in the end they reached 100-105°C and i kept it up for like 6 minutes. It came out with no errors.
But i am a bit confused now, the cpu can go high enough to 100°C and the PC doesn't freeze.
I checked the event viewer. It shows these errors:
The forced shutdown happened at 10:30 AM ( I woke up and found the PC frozen)
Error: The previous system shutdown at 6:25:36 AM on ‎9/‎6/‎2018 was unexpected. Event Source: EventLog, Event ID: 6008
Critical error: The system has rebooted without cleanly shutting down first. This error could be caused if the system stopped responding, crashed, or lost power unexpectedly. Event Source: Kernel-Power, Event ID: 41
Warning: The driver \Driver\WudfRd failed to load for the device ROOT\WPD\0000. Event Source: Kernel-PnP, Event ID: 219
I don't know if this is related to the problem or because i force shutdown my PC, because the same events are there when i have an electrical outage.
The only difference is the time in error with the event ID: 6008
when i have electrical outage, the time seems to be at the moment the computer shuts down. However when the PC freezes, as you can see in the times provided, i shut down the PC at 10:30 AM when i woke up, and the event log said "system shutdown at 6:25:36 AM"
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
You said you did memtest86 = no problem with RAM
Prime95 = CPU load/throttle doesn't cause the crash
-------------
That leaves Windows, Motherboard and the PSU + GPU

Test GPU with FurMark. PSU can be tested by using another one. I borrow from a friend
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
okay so i stress tested my GPU with FurMark for 30 minutes.
I used:
1600*900 Full Screen which is my screen resolution
Anti-aliasing 8x MSAA
3D test options: Dynamic Background
Burn-in
Xtreme burn-in
Post-Fx
My fps was stable at 48, and my GPU temperature increased gradually to 63°C.
Everything was fine, there were no crashes.
So that leaves windows, motherboard and PSU.
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
PSU can be tested with softwares but without hardware, it's not 100% accurate.. Open HWMonitor, run Prime95 + Furmark to maximize power draw from CPU and GPU. After using it for a few minutes, stop everything and look at the min/max +5/+12 voltages at the motherboard section. Less 5% variation is good.

Full thread:
http://www.tomshardware.com/forum/290251-28-testing-stability
So that leaves windows, motherboard and PSU.
BTW, you said you have power outages? PSU/MB have built-in safety mechanisms. Power fluctuations might cause that too.
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
PSU can be tested with softwares but without hardware, it's not 100% accurate.. Open HWMonitor, run Prime95 + Furmark to maximize power draw from CPU and GPU. After using it for a few minutes, stop everything and look at the min/max +5/+12 voltages at the motherboard section. Less 5% variation is good.
Okay tried that and it is working at the moment as well

CPU VCORE: Min 0.808V Max 1.096V
+12V: Min 12.096V Max 12.192V
AVCC: Min 3.360V Max 3.402
3VCC same as AVCC
+5V: Min 4.960 Max 5.04
According to these values they are well inside the tolerance boundaries.

BTW, you said you have power outages? PSU/MB have built-in safety mechanisms. Power fluctuations might cause that too.
So, what should i do regarding this. (taking into consideration this happened before i got power outages which only happened in one day this week)
I can reinstall windows on my new hard drive.
Also, i can remove my VGA card and use my integrated intel and see what will happen for a couple of days :/
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
I can reinstall windows on my new hard drive.
Also, i can remove my VGA card and use my integrated intel and see what will happen for a couple of days
I usually test by swapping parts to rule them out fast (a friend has pc repair shop). Since you can't do this, you have to do it the long way.

If it's driver issue like the event error said, reinstalling windows should fix it. You could try the iGPU as well.

Also, by power outages, I didn't mean power going out every time, I meant power fluctuating in your area causing the parts to stop responding. If it was a one-time affair, don't mind it
---------------
BTW, when exactly did these freezes/crashes start happening?
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
If it's driver issue like the event error said, reinstalling windows should fix it. You could try the iGPU as well.
Yeah i will go for it.
If it was a one-time affair, don't mind it
yeah happened in one day so no worries there
BTW, when exactly did these freezes/crashes start happening?
After i upgraded my PC :(
It just crashed again twice an hour ago.
I was playing yu gi oh duel links and then it crashed.
The second time happened while i am browsing chrome, after i finished the stress test by like 30 minutes.
Both crashes happened within an hour and a half.
I searched online again, and i was surprised when i found that a lot of people are facing the same problem in the same timeline. Some fingers are pointed at the latest windows update responsible for these freezes.
I just reached a wall where i don't know how else can i try to troubleshoot this except for reinstalling windows and working with iGPU
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
If you mean the GPU upgrade, try installing DDU.
https://www.guru3d.com/files-details/display-driver-uni...
I added a new hard drive, PSU, and the GPU.
I am now using my iGPU, and it's doing fine for 18 hours
I don't know if it is a GPU problem, and if it is indeed then why doesn't it crash in stress test, and why does it work normally for like 4 days straight and then starts crashing in one day all of a sudden.
I think what i can do is to install my new GPU in another PC and see how it behaves...
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
I added a new hard drive, PSU, and the GPU.
I am now using my iGPU, and it's doing fine for 18 hours
I don't know if it is a GPU problem, and if it is indeed then why doesn't it crash in stress test, and why does it work normally for like 4 days straight and then starts crashing in one day all of a sudden.
I think what i can do is to install my new GPU in another PC and see how it behaves...
I have seen people getting crashes/freezes caused by leftover driver settings after GPU upgrade. DDU should fix that. Using another PSU was a good idea.
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
Using another PSU was a good idea.
I got that with the new GPU upgrade so i don't know if it is faulty or not.
I have seen people getting crashes/freezes caused by leftover driver settings after GPU upgrade. DDU should fix that.
i used DDU when i installed the GPU 2 weeks ago, i will give it another try in a couple of days if i didn't return my GPU to the shop :/
Also, yesterday the DRAM LED started to blink again after the second crash, and the PC didn't boot up. After a while it just stopped blinking and booted up normally.
 

zebarjadi.raouf

Commendable
Jul 10, 2018
862
2
1,310
You said it worked fine on iGPU without any problem. If I'm not wrong that means PSU, GPU or driver. Since you changed PSU, That leaves GPU itself or driver. Try different driver version as well.
 

Petirdo

Commendable
May 28, 2016
24
0
1,510
Since you changed PSU
I changed my old PSU which was 400 W to a the new PSU seasonic 620W the moment i got my new GPU. So, i didn't change it. Sorry if i misinformed it.

My problem is this freeze happens totally randomly keeps working fine for 4 days or 3 and then it crashes again. So, i don't know whether my iGPU is working well or not. Also, i could install an old driver and then it work for couple of days and then it crashes. I have 4 days left to be able to fully return this GPU without losing any money in the process.