Question Asus Surge Protection Triggered when gaming under heavy load

[Fanatic010]

I was gaming yesterday and my pc suddenly shutdown and I was greeted with a screen saying a surge was detected. Every time I go do anything intensive now the same thing happens. I'm guessing its the psu which is a EVGA Supernova 650w G2 that is only 3 years old. I was able to play less intensive games and have gotten the gpu to 50% utilization but anything more than that it just shuts down. Browsing and stuff seems to function normally. Not sure if a full load on just the cpu does anything at all but I would rather not risk continuing to trigger whatever is going on. I've also tried disabling Asus Surge Protection in my bios because some people said it was finicky sometimes. However, instead of shutting down the screen went gray and fans went to full speed so I quickly held the power button to shut it off.

Again I'm assuming its the power supply but I don't have a way to check and have never experienced anything like this before. I've checked the wiring and have tried plugging the gpu into another rail on the psu but nothing changed. It's a shame because it is a very well rated psu and I haven't seen anyone with the same one who has had this happen which worries me it could be something more severe like my gpu or something....

Any help or insight would be appreciated.

HWMonitor Screenshots:

https://imgur.com/WbV3fTO

View: https://imgur.com/WbV3fTO

https://imgur.com/6wIfAX9

View: https://imgur.com/6wIfAX9

https://imgur.com/g1iLtsX

View: https://imgur.com/g1iLtsX

 

[jimmysmitty]

The only test to verify this is the PSU is to use a known good PSU in your system or try your PSU in a similar system to see if the same results happen.

Considering it takes over 50% GPU load and the GPU would be pulling the most power most likely your PSU does have a fault somewhere. I doubt it would be the motherboard as the PCIe slot can only pull 75W but the PCIe cables can pull much more depending on if they are 8pin or 6 pin.

 

[Fanatic010]

I ended up using HWinfo and saved a log file from the point of idle until shutdown while running Unigine Heaven benchmark. It took less than a minute for the system to shutoff. Nothing appears to be off and the temperatures plus voltages all seem correct?

https://docs.google.com/spreadsheets/d/13SQt5WQ9ha_mc2IETOxlJNK8z0P8cqN0g8O80k7DpuU/edit?usp=sharing

The numbers don't appear to be off at any point but it also gave me a warning saying it might not be able to track certain sensors. I'm guessing its the ones with the "X's) Not sure if they are important or not.

https://imgur.com/oYva5gM

View: https://imgur.com/oYva5gM


My roomate has an alienware a51 that is similar to spec as my pc but it is very proprietary and the psu only has a bunch of 6 pins in a weird configuration so I can't do much with that. Would it be wise to plug in my gpu to his pc to see if the gpu is at fault or is that even worth doing? I'm a little worried about jumping the gun on parts if I don't know what the exact issue is. It's not easy to send parts back somewhere if I order something where I live.

Is there any other tests or things I can do that don't just involve trying new hardware that I don't have access too?

 

[LinuxDevice]

It's rather unlikely you'd get any useful information plugging your GPU into his system unless it is a test of complete failure. I've had similar effects from power supplies which were good, but not high enough wattage (or high enough quality at the given wattage) once games or other devices started drawing a heavy load. In my case temperatures were fine, and replacing the PSU did the job. You never said which GPU you use, but I'd never use only 650w for something "significant".

 

[Fanatic010]

I have a single EVGA ACX 2.0 980ti gpu and my psu is an EVGA Supernova 650w G2. From my understanding anything over 600w was enough for the card when i built my machine back in 2015. It has worked flawlessly up uptil now but I guess there isn't much head room if for whatever reason the psu wasn't outputting power like it did when it was new. The readings on the HWinfo log should show issues though shouldn't it?

 

[jimmysmitty]

650W is more than enough for the 980Ti, especially such a good unit.

Again if the GPU is hitting 50% load and then after that it shuts down my biggest guess would the the PSU as thats when you would be pulling more load from the PSUs 12V rail.

You can try putting your GPU into his system but Dell is very good at building systems with PSUs that just barely met what is needed to run the system so if he has anything that requires less power than a 980Ti it probably wont have enough power. If it has a 650W PSU it would be a good test to see how it acts in your friends system.

Otherwise you might have to fork over the money for a new PSU as a test. I would suggest Amazon Prime or anything with a no hassle return.

 

[Fanatic010]

Reiterating on my previous post, shouldnt hwinfo show voltage fluctuations before the PC shuts down or is it too abrupt to see?

Also when I turned off the asus mobo surge protection as a test the fans went crazy in my system at the time of heavy load and the psu didnt power the system down at least not before I held the power button to force the shutdown. Would that still be the psu at fault. It should shut itself down if it detects power issues?

 

[jimmysmitty]

The PSU does have the following:

Heavy-duty protections, including OVP (Over Voltage Protection), UVP (Under Voltage Protection), OCP (Over Current Protection), OPP (Over Power Protection), and SCP (Short Circuit Protection)

So yes it should by all means stop it from allowing an issue. But as I said the best test it to try the PSU in another machine or try a new/good PSU in your machine. I know it is not the easiest but when I was diagnosing systems that was my methodology and it worked 99% of the time.

There is the odd chance that the motherboards sensor is wrong but normally its designed to notice when something on the board is getting a surge. It could be possible as well its not the PSU but another device is pulling more than normal causing it to happen like a USB device.

HWInfo monitors a lot but it doesn't monitor every part of the system so there is possibly some point where power draw is higher than normal.

 

[Fanatic010]

Okay thanks for all the replies. I guess I will try to order a new psu and see what happens.

 

[jimmysmitty]

Again I suggest using something with a no hassle return policy.

 

[Fanatic010]

Update:

I just hooked up a brand new Seasonic Focus Gold 650w that i ordered and the same thing is happening....😡 It just shuts off under any kind of load. I've tried turning off the bios surge protection and it still shuts off. Now what? It must be either the motherboard or the gpu I assume . This is so frustrating.

 

[jimmysmitty]

Do you have a motherboard with display out that you can use to test with? Pull the GPU out and poiwer on and put the system under stress. WIll eliminate the GPU. At least you now know the PSU is good. It sucks but thats the best way to troubleshoot.

 

[Fanatic010]

My motherboard is a Asus Z-170A and it does have both HDMI and Displayport out. What sort of test should run with the gpu out? Would it pull enough power to even trigger it?

My roomate has a old GTX 770 I could plug in but I don't want to potentially ruin any hardware by plugging other stuff in or my stuff into his computer potentially to troubleshoot. Is there a risk of anything going wrong in doing either of these things?

Edit:

Well I ended up plugging in the GTX 770 and the benchmark worked fine and it ran for 45 minutes with no issues. Tried BF5 and Fortnite and they seemed to work just fine as well. I'm guessing the problem lies with my 980ti which is really disheartening. I never pushed the card through overclocking or anything and can't imagine the most expensive piece of my pc would be at fault. Unless there is still something I'm missing but it doesn't look like that's the case. Guess i will reach out to EVGA again and see if anything can be done...

 

[jimmysmitty]

Glad you figured it out. EVGA is one of the better companies. I am sure they will make it right.