[SOLVED] PC restarts only during Gaming, tried lots of troubleshooting, finally here for help!

[ankur121192]

CPU - i5 7600k - Running without any Overclocking.
Motherboard - MSI Z170 mpower gaming titanium.
GPU - Gigabyte Nvidia 1080OC 8GB Windforce edition - Running without overclocking.
RAM - Corsair Vengeance 16GB @ 3000 MHz X 1 Module.
PSU - Corsair RM750
Storage - 1 Sata HDD @ 7200 rpm, 1 Sata SSD on which the game is installed ( Samsung 870 EVO), 1 M2 NVMe SSD on which Windows 11 is installed (WD Blue M2 NVMe SSD).
Display - LG GK650f 32inches Ultragear connected via Display Port.
Cooling - 2 X 120mm intakes, 1 X 120mm exhaust, all set to a smart Temp vs Speed setting through BIOS. Temperatures are going to 60-65 deg C max during OC.;70 degrees max during stress testing.

Main Issue : PC restarts; as if there was a momentary power cut, the power would go off and suddenly restart with a gap of hardly 1-2 seconds in that power off state. This happens only when playing a game, tested with 2 games - GTA V and Cricket 22 by Big Ant.

Some setup and settings : Apart from above mentioned PC components, there is a Mechanical keyboard (Redgear Invader MK 881) and a gaming mouse (Razer Deathadder) and Xbox X controller connected via USB. Nvidia GEForce is installed as well.

Some mix and matches that I tried and some troubleshooting that I tried but no luck :

1. Tried Memtest, 6 Passes and 0 errors, had no patience to test beyond that.
2. Tried to Overclock the CPU at my previously tried and tested settings, all 4 cores clocked at 4.5 GHz, Fan setup custom, CPU temp was reaching 55 Deg C and System Temperature was close to 65 Deg C.
3. Tried to Overclock the GPU with Gigabyte's own utility as I have not much idea of GPU overclocking, so simply tried both the gaming and the turbo boost modes.
4. Tried to have both the CPU and the GPU Overclocked and running at that combination.
5. Tried BIOS reset to default, NO OC at all, and then running the games at Nvidia GEForce recommended Optimal settings.
6. Tried to Unplug the Xbox Controller and play with Mouse and Keyboard only.
7. Furmark for GPU, Prime for CPU, OCCT for PSU. I tried to stress test the components, no restart during stress testing, even if the component usage was reaching as high as 95-100% and temperatures were going pretty high but within the recommended limits, as I live in India and it gets pretty hot these days.
8. Tried to run the game at below optimal settings, through Nvidia GEForce Experience app, there is that slider between performance and graphics quality. There is one optimal point, I moved slider all the way down from the optimal to midway between performance and quality. There was no OC during this.
9. I was earlier running Windows 10 on Sata SSD and the Game on NVMe SSD. I clean installed Windows 11 and formatted all my storage as well to bring things to the current mentioned configuration.
10. I was doubtful about the PSU, hence the OCCT testing, but it didn't fail in that.
11. I already cleaned the whole PC, replaced the thermal paste on the CPU and the GPU both. Cleaning was thorough enough to clean the dust between all the fins and everything.
12. I tried to alter the Virtual Memory thing, manually, set it to let windows decide option as well, no difference.

My doubts here :

1. The house I live in, is not subjected to any visible voltage fluctuations. Could there be invisible, minute fluctuations that can cause this. If so, why only during gaming!
2. There is no UPS, basically the power cord from the PSU is directly connected to the Wall outlet, no extension/multiplug in between. Do I need UPS for stability purposes? Or a Voltage Stabilizer?
3. If it was a failing GPU, how would I stand Furmark.
4. As it is said that the realistic way is to run the CPU, GPU stress tests in parallel, what tests and durations do you recommend, so that I won't end up burning my system to ashes.
5. My PSU is modular, on this particular model, do I have to worry about the Voltage rail thing? Basically 24 pin ATX could be plugged into one particular slot, it is sitting there only. Same goes for CPU 4+4 pin. The GPU 8-pin i.e. PCIe is something that I have tried switching PSU slots for, no difference.
6. It shall go away if it was Software-related or OS related, right? Can GEForce experience be the culprit here? I shall try restoring my system to a restore point before GeForce experience installation, and update accordingly. As I have read in a few forums how it causes issues at times.
7. Like GeForce experience, I have read similar stories about Razer Synapse being a culprit at times, but it is really difficult to manage without that utility.
8. I don't have mine or any friend's PSU to test things. If it indeed is PSU, my only option would be to get a new PSU, which could be a useless investment if PSU is not the actual culprit here.

This is really getting annoying people, and hence such a detailed post, I am trying to really sort this out, and I seek genuine help on this. The only issue would be the ability to invest in new PSU, new GPU one by one without having a clearer picture of the problem.
If anyone has faced these things in past, have something to contribute, please help me out!

 

[Cloudflare]

Could It Be the PSU? (But You said you already tested it with OCCT) SO what OCCT basically does it that it Pushes Every (almost every) component In the PC To Simulate VERY VERY High Loads and tests if the PSU can Handle that much power draw. SO.... IF it didn't restart (a.k.a Crash if I'm right) During the OCCT Test Then The PSU is FINE.

But We have a Problem. The PSU of yours Is 750W if im right. So Your GPU needs 540W (recommended by manufacturer) And your CPU needs 400W. In total you need 940W (More than That If you include the other stuff like the Motherboard) BUT Mostly it wont be a problem. Im running my i7 3770k In a 350W PSU With a Gigabyte Motherboard + AMD Radeon rx 590 in my 350W PSU, But as I don't STRESS my CPU or GPU much, I haven't experienced any issue with my PSU.

Also, How Old Is your PSU?

To test your PSU you could head to some PC repairing shop and give it to them. They Will Connect it to a PSU tester And check if It is fine by checking if the Voltage output is as it should be.

Happy To Help

 

[sizzling]

Could It Be the PSU? (But You said you already tested it with OCCT) SO what OCCT basically does it that it Pushes Every (almost every) component In the PC To Simulate VERY VERY High Loads and tests if the PSU can Handle that much power draw. SO.... IF it didn't restart (a.k.a Crash if I'm right) During the OCCT Test Then The PSU is FINE.

But We have a Problem. The PSU of yours Is 750W if im right. So Your GPU needs 540W (recommended by manufacturer) And your CPU needs 400W. In total you need 940W (More than That If you include the other stuff like the Motherboard) BUT Mostly it wont be a problem. Im running my i7 3770k In a 350W PSU With a Gigabyte Motherboard + AMD Radeon rx 590 in my 350W PSU, But as I don't STRESS my CPU or GPU much, I haven't experienced any issue with my PSU.

Also, How Old Is your PSU?

To test your PSU you could head to some PC repairing shop and give it to them. They Will Connect it to a PSU tester And check if It is fine by checking if the Voltage output is as it should be.

Happy To Help

Where are you getting those wattages from? Gigabyte recommend a 500w psu for a system running that model 1080, this will include an allowance for high wattage CPU’s of the time. That cpu needs nowhere near 400w, no desktop cpu takes 400w

 

[Diguelo]

Asking a question here regarding the Graphics cards age.

I know a lot of people go and buy graphics cards, turn the settings up way past their recommended and then sit back to enjoy awesome detail enhancements. Settings are recommended for a reason. A lot of it is down to cooling, or lack of it in most cases.

If your G Card is over two years old, you may have a bad heat pad or as I found in a repair recently no pad, just the surrond outside of the die on the card. I removed all the junk from the die area in the card (Involves stripping down the card, if your not sure how ask a pro to do it or check youtube for relevant vids). Instead of a pad the GPU got a layer of paste covering every part of the die (GPU shape) . Reassembly took a short time and when tested its Temps (avg and hotspot) before 100 plus on both and would shut down once the cutoff temp was reached. After runs all day happily peaking at 72 deg C and hotspot 80 deg C. Few pence converted a shutdown into a gaming pc again.

I also installed MSI Afterburner on his machine and set it for the 100% fan to kick in if the Avg temp hit 75 .

Granted the recommended game settings may not show mindblowingly brilliant flares and every leaf on a tree,. but the machine runs all day at 60FPS and doesnt even get the MSI warning fans at 100%

Sometimes just installing MSI Afterburner or EVGA precision is enough to alter your cooling profile. Air is better then water, it cools down a lot faster and easier and in a lot of cases a lot cheaper

HWMonitor a hardware monitor program will help point out hotspots in the PC Hardware. The program is a goldmine of info and can save a lot of time tracing a heat shutdown or indeed a PSU fault. https://www.cpuid.com/softwares.html (Makers site, not a clone) You can set this to record a profile of a machine, once it crashes in the problem, analysis of the log will help you find the fault.

This information is given on Personal experience. Its up to you if you wish to use or ignore it.

 

[Lutfij]

OP, Please disregard what @Cloudflare has said about wattage demands of components.

As for your query, an UPS has battery backup. If you choose a good quality unit you will have voltage stabilization built into it. Out of curiosity, what BIOS version are you working on for your motherboard? As opposed to buying components, it's best if you can source a donor PSU that has at least 550W(from a reliably built unit) at the entire system's disposal to rule out your unit being the root cause of the issue.

You might want to try relocating to another room to rule out a discrepancy with the outlet you're working off of.

You forgot to mention the make and model of the case...and the placements of the fans(although their airflow orientation is stated) as well as the cooler you're using to cool to that processor.

 

[ankur121192]

Could It Be the PSU? (But You said you already tested it with OCCT) SO what OCCT basically does it that it Pushes Every (almost every) component In the PC To Simulate VERY VERY High Loads and tests if the PSU can Handle that much power draw. SO.... IF it didn't restart (a.k.a Crash if I'm right) During the OCCT Test Then The PSU is FINE.

But We have a Problem. The PSU of yours Is 750W if im right. So Your GPU needs 540W (recommended by manufacturer) And your CPU needs 400W. In total you need 940W (More than That If you include the other stuff like the Motherboard) BUT Mostly it wont be a problem. Im running my i7 3770k In a 350W PSU With a Gigabyte Motherboard + AMD Radeon rx 590 in my 350W PSU, But as I don't STRESS my CPU or GPU much, I haven't experienced any issue with my PSU.

Also, How Old Is your PSU?

To test your PSU you could head to some PC repairing shop and give it to them. They Will Connect it to a PSU tester And check if It is fine by checking if the Voltage output is as it should be.

Happy To Help

Do you suggest a particular duration of OCCT, I ran it for close to 15 minutes with no crashing.

While it could be the PSU, I have been suspecting some sort of house wiring/voltage issue, as in this morning, PC was going into a restart loop right after POST beep every time. This sometimes happens in the morning, maybe load variation in the building or something like that. I am no Electrical Engineer to pin point that, but an observation.

I haven't checked the Wattage recommendations by the manufacturer, rather I used that power requirement calculator that is usually available online, which, despite entering all the details of peripherals as well, ended up suggesting 650W PSU.

Thanks for the suggestion on PC Repair Shop PSU testing, I shall do that. That way at least one piece of the puzzle can be sorted.

This PSU was bought from a fellow member of a forum in my country back in 2018, This PSU is 8 years old.

 

[ankur121192]

Asking a question here regarding the Graphics cards age.

I know a lot of people go and buy graphics cards, turn the settings up way past their recommended and then sit back to enjoy awesome detail enhancements. Settings are recommended for a reason. A lot of it is down to cooling, or lack of it in most cases.

If your G Card is over two years old, you may have a bad heat pad or as I found in a repair recently no pad, just the surrond outside of the die on the card. I removed all the junk from the die area in the card (Involves stripping down the card, if your not sure how ask a pro to do it or check youtube for relevant vids). Instead of a pad the GPU got a layer of paste covering every part of the die (GPU shape) . Reassembly took a short time and when tested its Temps (avg and hotspot) before 100 plus on both and would shut down once the cutoff temp was reached. After runs all day happily peaking at 72 deg C and hotspot 80 deg C. Few pence converted a shutdown into a gaming pc again.

I also installed MSI Afterburner on his machine and set it for the 100% fan to kick in if the Avg temp hit 75 .

Granted the recommended game settings may not show mindblowingly brilliant flares and every leaf on a tree,. but the machine runs all day at 60FPS and doesnt even get the MSI warning fans at 100%

Sometimes just installing MSI Afterburner or EVGA precision is enough to alter your cooling profile. Air is better then water, it cools down a lot faster and easier and in a lot of cases a lot cheaper

HWMonitor a hardware monitor program will help point out hotspots in the PC Hardware. The program is a goldmine of info and can save a lot of time tracing a heat shutdown or indeed a PSU fault. https://www.cpuid.com/softwares.html (Makers site, not a clone) You can set this to record a profile of a machine, once it crashes in the problem, analysis of the log will help you find the fault.

This information is given on Personal experience. Its up to you if you wish to use or ignore it.

The GPU is 4 years old roughly speaking. Also, I can assure you it is not the GPU heat up issues, I have run a stress test and monitored CPU temp, GPU temp, and System Temperature and this GPU model has 3 fans with a default setting of the third fan kicking in only when GPU usage is above a certain percentage. So whenever stress testing or gaming, the Third GPU fan kicks in and keeps it fairly below what I would say HOT temperatures.

Recently while replacing the thermal paste on the CPU, I had stripped GPU down, and cleaned it, and replaced thermal paste there as well. Moreover, I assume 2-3 minutes into the game cannot really heat the GPU up to that extent.

As for the applications, since I am using a Gigabyte card, I am using Gigabyte extreme to monitor or anything.

And when it comes to cranking up the settings from the game, from Nvidia settings or from GPU tuning utility, I always keep it a little under the recommended mark, just to be safe on the thermal side of things. I don't know why, but hot components freak me out, so, always below recommended settings.

Also, the Motherboard utility I have, got HWmonitor in-built. Now here is the fun part, I have tried to record a log of everything, so that I can trace what went wrong when system restarted. A lot of times the log is not written/lost maybe due to that abrupt power failure (By the way, Kernel - Power Failure - 41 as per windows logs). Even for the GPU, when I recorded the log, I checked it out later, and on the GPU part, everything was fine, fan speeds,, GPU temperature, clock speed, Memory usage, everything was pretty normal, nothing was going crazy.

 

[ankur121192]

OP, Please disregard what @Cloudflare has said about wattage demands of components.

As for your query, an UPS has battery backup. If you choose a good quality unit you will have voltage stabilization built into it. Out of curiosity, what BIOS version are you working on for your motherboard? As opposed to buying components, it's best if you can source a donor PSU that has at least 550W(from a reliably built unit) at the entire system's disposal to rule out your unit being the root cause of the issue.

You might want to try relocating to another room to rule out a discrepancy with the outlet you're working off of.

You forgot to mention the make and model of the case...and the placements of the fans(although their airflow orientation is stated) as well as the cooler you're using to cool to that processor.

Thanks for the response, yes I had several times calculated Wattage requirements from online calculators, so I guess 750W is definitely not a problem there.

I would probably go for that APC thing, which is international standard, and I guess to run 750W PSU, I would need at least 1200VA, I mean I had checked that thing, and if I can't rule the issue out, this is my next step in terms of investing money into the solution.

BIOS version exactly I can't remember as I am away from my system right now, but what do I remember 100% from around a couple of weeks back, is checking with MSI website if I am using the latest BIOS. Also, I have checked all the drivers and everything from Driver Easy - No issues there as well.

I will look for a Donor, otherwise, I have a 550W Non-Modular PSU, but from the maker Antec, I guess it is not really up to the standard as you would expect. So looking for a PSU Donor for testing purposes shall be my next step, or to at least take my PSU to a PC repair shop for testing purposes, as someone else suggested.

About the room and the power outlet possibility, I would get a reliable (Maybe 3M) extension cord and plug just the PSU into it, I know losses would be there due to the length of the cable of the extension, but that is the best bet for now.

About the query,
Case - Antec GX 200.
CPU Cooler - Cooler Master H410
2 X 120 mm fans at the front bottom and mid-height of the case for Intake.
1 X 120 mm fan at the rear top height for exhaust.

If possible, in a couple of hours I will be going home, and then I will click a few pictures to give a better idea.


Also, if you have any idea/experience on this, I can vaguely remember reading somewhere sometime that how if there is a little piece of metal or something, between the legs of the motherboard and the case metal body, it can cause some grounding/earthing issues which can result in this kind of behavior.

Edit - APC PSU I meant, instead of AEC.

 

[ankur121192]

Thank you everyone for trying.

You know what, with a hope that someone might stumble across this thread in future, and might have exact same issue where system restarts and windows logs show "Kernel Power 41" event, whenever gaming.

Maybe that someone would be scratching his/her head after successfully running a variety of stress tests and not witnessing any Crash/Failure/Restart.

Maybe they would suspect their GPU, PSU, Cooling and try a variety of things, oh hell even clean install the OS, fiddle with drivers and almost everything until they reach the point of replacing PSU, but then decide to search over the forums if something similar happened to someone else.


Here is what has been the root cause of the problem in my case, and in conditions that I live in :

Problem: " System turns off abruptly while gaming and restarts with a gap of 2-3 seconds. Happens with Resource-intensive games (Sorry but I haven't tried this on low requirement games, my test games were GTA V & GTA V Online, Cricket 22). "

Conditions I live in and probable cause: "I live in a metropolitan city in India, despite being metropolitan and everything else, the constructions here are often done by a third party/private builders. Now, these guys do House wiring in some XYZ way I believe, because this kind of thing I faced somewhere else in some other apartment too, while using a Cooler Master PSU with different PC specs. What happens, while gaming, and I say this based on the FPS shown by GTA V built-in Benchmarks, I guess, due to changing graphic demands, as well as changing processor load, due to dynamic conditions in gaming, The current/voltage (I am no electrical engineer, basically power related) demand fluctuates at a smaller level.

These kinds of fluctuations are often handled by good PSUs in a decent manner up to a certain point of time in their life cycle. Once there, and I say this based on the same experience twice, the PSU starts giving up against these fluctuations of power demand. And what I witnessed, this never was a problem while the PSU was brand new or during the Initial days. But using a PSU directly plugged into Wall socket, without any power backup or a UPS, and after having sudden power cuts for, I would say close to 30-40 times in a duration of ~3 years, something goes wrong with the capabilities of PSU to handle these fluctuations. Or maybe just the usage for that duration, dealing with power demand fluctuations over very short time durations as gaming would cause.

The conclusion is, using a PSU without a good UPS in conditions like in India, over a few years, the chances of PSU not being able to handle fluctuations in real-time while gaming, would simply cause the PSU to fail and end up restarting your PC. This would not happen during stress testing because it is not about the load it puts on PSU, it has more to do with the fluctuations in real time.

Solution: " Yesterday, after a lot of suggestions from lots of forums, PC enthusiast friends, and googling and playing permutation-combination of stress testing and everything, I was kinda suspect about this issue, so I ordered APC UPS from amazon with Same Day Delivery. This was more of a hunch after all of that brainstorming, and I was borderline ready to look for the PSU invoice and send it back for RMA. So, once the UPS came in, I plugged just the PSU cable into it, Monitor, Router, whatever extras are, still connected to wall outlets. After that, I went into the game, played on existing settings(much below Nvidia GeForce recommended Optimal settings) for a while, say 30-40 minutes. There was no restart or crash.

Then, I came to Nvidia GeForce experience and slid that slider to optimal settings, basically, reopened GTA V and went into its own Graphics settings, checked the options, and saw that running for 1440p, at 144Hz refresh rate, with the majority of settings at High, the Video memory usage was somewhere close to 3GB out of available 8GB. I ran Benchmark test from there, within the game. Benchmark tests ran pretty smooth, FPS ranged from 90-140 during different kind of scenes and simulations, I mean in one particular scene, the fluctuation wasn't too high, some 10-15 fps drop or rise within a scene only, but GTA V benchmarking has 4-5 different scenes, so that FPS range is for all the benchmarks test combined. The game shut itself afterwards, as displayed as a warning before starting Benchmarking. I fired the game up again, and this time I played for another 30 minutes or so. I could feel my system a bit hotter than usual when I touched the cabinet. Earlier I was avoiding to tweak System fans and GPU fans to run a bit faster as I had doubts about Power Drawl.

Now, I quit the game, went into Bios to link fan speed to CPU temperatures, linked the CPU fan, both the intakes, and the exhaust in a manner where CPU fan was running a little faster than the exhaust, and the two intakes were just a bit slower than the exhaust.

I also tweaked the GPU fan settings and linked it to GPU temperatures as well.

Now I again started playing GTA V online, I played for close to 2-3 hours in a go. Yes I could hear the fans spinning. But hey, the cabinet wasn't as hot as in case 2. I could put my hand a little away from the exhaust and feel the warm air coming out.

But you know what, I had dinner, and I again played for 2-3 hours.


So, from not being able to play even for 15 minutes to 2 sessions of 2-3 hours of gameplay, that too on High graphics settings and 144Hz refresh rate at 1440p resolution on a 32inch Monitor, the whole experience made me sleep like a baby later at night. All it took is a reputed UPS. By the way, my UPS is 1100 VA/660 Watt whereas the PSU is 750W. But I guess the PSU would never be drawing 750W even during gaming. So, I guess I am safe so far.




Now, Any PSU experts in the house? Shall I submit my PSU for RMA now? As I suspect something went wrong with all that operation without any UPS, and that something wrong might come back later to bite me in the a**. I have just 3 more months of warranty, so, please suggest whether I should submit for RMA.