[SOLVED] Screen goes black "No Signal" during PC games

[Spitfire7]

I seem to be having a new issue that I have seen very sporadically over the last few months.

In a heavy game CPU temps are 45c and GPU gets up to 75c tops. These temps never go over that. Randomly the screen will just cut out and then I see no signal. I can hear things still going on through the speakers. This is most often the case. However, the last time the screen went out to no signal, the computer also froze. I can only get the screen back by restarting the computer. So far this is only happening at heavy usage, but not so much just normal internet or work desktop use.

I did some research and most are saying its power related and the issue is power to the GPU speculating that the PSU might be going out not giving enough power to the GPU and as a fail safe it cuts the GPU off. Others have been curious if its a GPU related issue, but no 100% proof of this yet in the forums.

Here are my specs.

PC Specs:
ASUS Z790E ROG STRIX Gaming WiFi
13th Gen Intel Core i7-13700K 3.4GHz
32GB Corsair Dominator 6400MHz
Gigabyte RTX 3070 GeForce Nvidia 8GB
Corsair RM850x 850 watt PSU
Samsung 980 PRO SSD 2 TB
1 TB HHD backup
Windows 11 64-Bit
65" TCL TV Monitor 4k resolution

Here is what tried to troubleshoot:

• I got a new hdmi cable and plugged into all different inputs and still nothing.
• I even used a different tv monitor and still happening.
• I have updated my graphics drivers to the latest NVidia 576.88.
• I don't have anything overclocked.
• I have fans set to maximum while playing and temps are CPU temps 45c and GPU gets up to 75c tops.
• I re-seated that graphics card, blew out dust which there wasn't any and same with the PSU.
• I made sure all cable from PSU to GPU were tight and even used a different connector on my last try.
• I have also done several GPU max stress tests and they all passed perfectly.

The main game I am playing while this cuts out is Car Mechanic Simulator 2021 and MSFS2020. It takes a lot of memory and GPU usage. Maybe as soon as 4 months ago I was able to play this game for 3 hours straight with no issues. Now within 30 min, screen cuts to black and then "No Signal" shows on screen.

Have you heard of this issue and any ideas what could be happening? I definitely have enough power with the Corsair 850w to run my system. Please let me know other trouble shooting ideas I can try and if you know of a possible fix. Thank you.

p.s. I have been surfing the internet and typing this post over an hour without any issues. Normal Windows desktop work seems to not cause it. Seems to be related to heavy usage.

 

[Third-Eye]

This stuff right here will give the absolute best performance of all?

I can out it on the GPUs processor and then all the places that has normal thermal pads?

Give me the best link and I'll buy it.

https://a.co/d/8nY5Usu

I would test with the thermal paste and make sure the graphics card is funcitoning in games properly before looking to buy PTM 7950 or a Kryosheet.

It's specifically for the GPU itself, it's not for replacing thermal pads. You can get PTM 7950 from amazon or the LTT store and some other stores. The LTT store may guarantee genuine PTM 7950. You will get a larger sheet of PTM 7950 than you actually require for the GPU surface, and you will need to cut it to 25mm x 25mm to fit the GPU.

You can get Kryosheet from amazon or directly from the Thermal Grizzly store webpage. Kryosheet comes in precut pads, so you don't need to cut it. Pick the 25mm x 25mm option for the RTX 3070. https://www.thermal-grizzly.com/en/kryosheet/s-tg-ks-25-25

The kryosheet is more expensive, but easier to handle and install. It can only be reused a couple (about 2) times before it needs to be replaced. You shouldn't need to open the GPU except to maybe replace thermal pads in 5-10 years depending on the condition and thermals of the entire graphics card. PTM 7950 should be reusable, but you will still possibly have the same thermal pad life span issue.

 

[Spitfire7]

I would test with the thermal paste and make sure the graphics card is funcitoning in games properly before looking to buy PTM 7950 or a Kryosheet.

It's specifically for the GPU itself, it's not for replacing thermal pads. You can get PTM 7950 from amazon or the LTT store and some other stores. The LTT store may guarantee genuine PTM 7950. You will get a larger sheet of PTM 7950 than you actually require for the GPU surface, and you will need to cut it to 25mm x 25mm to fit the GPU.

You can get Kryosheet from amazon or directly from the Thermal Grizzly store webpage. Kryosheet comes in precut pads, so you don't need to cut it. Pick the 25mm x 25mm option for the RTX 3070. https://www.thermal-grizzly.com/en/kryosheet/s-tg-ks-25-25

The kryosheet is more expensive, but easier to handle and install. It can only be reused a couple (about 2) times before it needs to be replaced. You shouldn't need to open the GPU except to maybe replace thermal pads in 5-10 years depending on the condition and thermals of the entire graphics card. PTM 7950 should be reusable, but you will still possibly have the same thermal pad life span issue.

How much of a difference on cooling are we talking? Currently I have been pushing MSFS on ultra for the last hour and we are up and running great for the first time in a month. I am pushing it to its limits. Yes GPU temp is around 74c average, but I don't seem to have the spikes anymore which is amazing.

So PTM7950 is only to replace thermal paste? I still need the thermal pads?

I canceled the order on the thermal pads for right now. So please set me up with the best solution that will make a big difference. Lets pass on the kryosheets for now though. What size thermal pads do I need for the GPU? Do they use the .5mm or the 1mm thickness?

Did you have a good Amazon link option of the PTM7950 that you would recommend?

 

[Spitfire7]

Huge win today and I think we finally found the issue and the fix.

I took my card apart and reapplied new Arctic Silver 5 paste and put it to the test. I ran MSFS at Ultra in the rain and this would kill it every time. I just played for over an hour and GPU temps stayed at 74 tops and in testing my hot spots, it never got over 74c using FurMark GPU stress test.

So that was the fix. My card was spiking in the hot spots going up to 108c and the card would shut itself off and turn off the display giving me no signal or lock up the computer and no signal.

That is exciting. I have new thermal pads coming anyways and I will apply those as well. I am still interested in your PTM7950 idea, but I need to learn more about it. Like how much of a difference will I see and will those replace both the thermal paste on the GPUs processor and where the original thermal pads are currently.

What a crazy last couple of months. My TV broke unrelated, then got a new tv and then my Corsair AIO cool pump broke making my computer inoperative. Then the GPUs thermal paste was all dried up causing it to crash. Each thing seemed to not be related, but gosh all these thing happening within the same month really sucked. Oh and not to mention my 1970 Classic car Nova's wiring melted leaving me stranded on the highway. That is now fixed as well, but my goodness. All in one month.

 

[Spitfire7]

105C is the thermal limit in most GPUs. If the hot spot reaches that, even if the average of the card is at 71C, it will throttle. You may need to take out the cooler and put new paste on the card.
Its also possible that some core on the card is gone bad and causes that hotspot temperature and when the card tries to use it it crashes.
Also its not impossible for you to be a victim of the 13th and 14th gen intel degradation issue but the symptoms don't match, that usually causes BSOD.

This was the winner and the fix. Thank you so much for the idea.

 

[Spitfire7]

Hot Spot of 105c on most nvidia cards means the card is thermal throttling and likely shutting the system down to prevent damage to the GPU. I can't say that it's a very wide spread issue, but for the people I've been helping, this is becoming a common issue with 30 series cards and usually just means you need to reapply thermal paste to the GPU.

I've helped probably close to 20 people now, maybe even more, with very similar issues, all with 30 series cards of different brands, usually only the 2020 to mid 2021 manufactured models. Almost all of them had the same overheating issue at varying stages of defect, and I've had it on all of my own 30 series cards manufactured in 2021. It makes me wonder if they all used the same thermal paste supplier and if it was defective.

One of the easiest ways to check for this issue other than the Hot Spot temperature is if the fans suddenly ramp up to high rpm when starting a game and get very loud because of that. Under normal operation, the fans usually won't go past 70% speed at max GPU usage. You only notice the problem when you actually try to use the GPU for some workload like gaming, since the GPU doesn't need all that extra power for basic desktop stuff and web browsing.

Close second winner and I really appreciate all the advice and help as well.