Hello Friends,
I'm new to this world of high-end video cards.
My machine (description at the end of the message) was assembled in DEC/24 - and I waited until now to install the video card - an Astral 5090.
The card is operating 100% without major problems, without overheating, and with all drivers updated and basically still under factory settings.
But, for some reason, without any apparent reason, it starts receiving 0w and turns on the fans at maximum speed - and of course, with that, the video turns off - and I am forced to restart the machine. (see video: https://streamable.com/s3inwa)
This has happened 4x in the last 10 days and it's a little terrifying:
1- First, When I went to install the video card I had this same problem during the system boot - that is, the "100% new" video card, without any use, without entering Windows (at boot) had this problem - AND NO DRIVER INSTALLED.
2- Yesterday - practically idle, watching a YouTube video, however, after a few hours of continuous use.
3- Today - When I decide to update de Bios of my motherboard. (My Bios was OK, on the last stable version) - the video card goes crazy with 80/90% of Bios update. I wait to complete 100% (without video) - wait for auto restart - and done.
4- Today - again - same event of item 2. 1h later.
The video card's power sensor shows that it has been receiving an average of 42W with a variable maximum peak of 200/500W - but it has already registered 725W
Also: 12V / 3.5A
One thing I noticed:
When I installed the video card, when I had the first problem - I THINK (I'm not 100% sure, but there's a good chance) - that there was a Windows update to be done. When I turned off the machine, the update was activated (since there was no video, I didn't see it)
Yesterday, there was an update to be done (this time I saw it being done when the machine restarted)
Today, when I update the BIOS.
But no update in the last event (4).
Any Idea?
===
Another info: When I had the problem during installation, I thought it might be something related to the power supply. So I changed my power supply to a Thor 1600w. Then, the power supply and cables were all changed, and the problem persisted.
===
PC:
GIGABYTE Z890 AORUS MASTER • Intel Core ULTRA 9 285K @ 5.7 GHz • ASUS ROG ASTRAL GeForce RTX 5090 OC • 96GB (2x48GB) CORSAIR DOMINATOR TITANIUM DDR5 6800MT/s • SSD 2TB CRUCIAL T705 NVMe PCIe 5.0 • SSD 4TB KINGSTON FURY RENEGADE NVMe PCIe 4.0 • WATERCOOLER THERMALTAKE TH420 V2 ULTRA EX LCD • 10x FAN LIAN LI TL140 LCD • FONTE ASUS ROG THOR TITANIUM 1600w FULL MODULAR
Windows 11 24H2 26100.3624
Driver Nvidia: 572.83
I'm also having the same issue. The exact same behavior:
Screen goes blank, all GPU fans goes up to 100%, flashing red lights near the connector.
I also had my on-board CPU graphics connected so I could see that the rest of my system is running fine. GPU just 'disappears' from my system.
Can I ask what the position of your BIOS switch when you encountered the problem? I had the on the default side, I've since switched it over and haven't ran into any issues yet. Will report back otherwise.
@Aevi
About your questions:
- My machine is 100% new - built in Dec.24 - Full Specs on main post.
- No software was installed when I have the first crash (at boot during the instalation)
- Then I install Nvdia App and after another crash - GPU Tweak III. NO other software.
- No custom voltage. All settings from factory.
- VBios is the same of yours (but had crashes with the old Vbios too) - all the other infos are on main post.
First, From what I've seen, and I think it's important to keep this in mind, is that crashes is like a protection mechanism for the VGA, it happens with other VGAs too, so we can rule out the issue of it being a hardware bug.
Many experts that I talk about had the same opinion about a driver issue - and recommend to make a D.D.U. and install the best stable version of Nvidia Driver: 572.75
I don't do this yet.
I'm testing all possibilities to discard a hardward mal function first.
- As I said in main post - I rulled out the possibility of a power source issue. I'm installed the Astral with a ROG Loki 1200w - and with the issue - change for a new ROG Thor 1600w - and the issue persists.
- Some experts tell me about the cables - IT'S VERY IMPORT TO YOU TEST THIS TOO - In my case, as I test with a ROG Loki 1200w - then with the new ROG Thor 1600w - I left this test in standby. I chose to test other things - but the next step is to use the adapter that comes with the board itself.
- Now I was testing the video outputs - with the HDMI and DP cables. Why? Because without the drivers installed I was having a lot of screen flickering problems with HDMI. With the driver instalation - the screen flickering problem is gone. But, since the problem starts with the video dropping - I thought it might have something to do with that. But not. The issue still happens.
- So I started to think that it could be a problem with Wireview Pro that I'm using to monitor the GPU. I sent an email to ThermalGrizzly support - to find out if they had any records of Wireview malfunctioning with the 5090 - especially because I'm registering loads of +600w (684w, 730w..)
- With that, I decided to take a picture of the Wireview sensor showing the 684w record. And when I did, the problem occurred. Sent the photos to ThermalGrizzly.
- Thermal Grizzly then asked me to remove the Wireview and try without it. I was going to do that - but since I was testing the HDMI cable - I left it for a few more days - and then, I noticed that the Wireview was showing 795w again and I decided to take another photo. And the issue OCCURRED AGAIN!! WHY????? This was the only time the problem occurred in a "controlled event" - not randomly.
- A photo causing the problem??? No (hehehe) - My PC case is an aquarium and has a glass door that opens with a MECHANICAL button (no electricity). To make the photo look better, I had the idea of opening the Pc Case... both times, when I clicked the button and opened the door - the problem occurred.
- The first time, I didn't think that was it - but the second time, it became clear. I waited for the machine to reboot, enter Windows and PRESSED THE CASE DOOR BUTTON AGAIN... and... CRASHES AGAIN!!!
- Initially I thought it might be an electrical problem... that the button was short-circuiting - something like that...But the button is 100% mechanical. The door is held in place by a pressure clip, and the button only releases the clip.
- Then I remembered a thread I saw on reedit trying to figure out what the problem was, weeks ago: https://www.reddit.com/r/PcBuild/comments/1dd78pv/gpu_maxing_out_fans_and_turning_off_my_screens/
- And what does this have to do with opening the case? When opening the case, the door vibrates the case structure MINIMALLY, exactly like the problem of hitting the table. And in my case there was an aggravating factor. My VGA support was placed ON TOP of the fan (see photo).
Thus, the crash issue as a whole is explained - I had crashes without opening the door.. how explain it?
As the VGA support was over the fan - it vibrated - and passed the vibration to the VGA. "a little more vibration" like opening the case door - was enough to cause the crash. In other words, when the fan vibrated more, even with the machine at Idle, this sometimes happened - we heard the fan making more noise - depending on the vibration added, it was enough to cause the crash.
(ok, it might not be, but I believe it's a plausible explanation to explain all the events.)
Now, who controls this? I think it's the driver. The driver must analyze a VGA displacement "limit" and activate the protection mechanism - to turn it off and avoid damage.
This protection is activated both because of this - and also because of BAD CONTACT of the power cables (that's why it's a good idea to check and test the cables too)
I removed my VGA support and will test it now without it. I'm thinking of buying a horizontal support like ROG Wingwall.
ASUS GPU TWEAK III has this controller:
I don't know why mine are oscillating non-stop. Can you check yours and XYZ's values?
I did a test here too, placing the support tighter - and the vibration, of course, increased. The Sensor then started to oscillate between "OK" and "Could be Better" - Showing that it was the support that was vibrating the VGA more or less.
I'm testing here - Let's continue exchanging information to see if we can resolve this!
Daniel
Hello @Silent_Scone
Yes, the crashes continues without Wireview.
The Machine is 100% new - assembled in dec/24 - and I wait for 5090 to install my first VGA in years (I spent about 10 years using only a notebook).
I test with 2 PSU - first with a Loki 1200w - with the issue - I thought it might be a problem with the PSU and buy another one - now a Thor 1600w. Both new, with original cables. Both I have the crashes.
Astral still with 100% factory settings. NO CHANGES.
As mentioned - I finally discovered on Saturday a way to reproduce the crash consistently - OPENING THE PC CASE. Every time I open the pc, it crashes. something like this:
https://www.reddit.com/r/PcBuild/comments/1dd78pv/gpu_maxing_out_fans_and_turning_off_my_screens/
But it also crashes without opening the case. I imagine the problem is because I placed the VGA HOLDER on top of the case fan. So, when the fan vibrates more - the VGA vibrates more. Until the crash happens. When I open the PC just the minimum vibration of the door opening crashes too.
I remove the VGA HOLDER and testing:
@Silent_Scone do you know the best values for XYZ (GPU TWEAK III):
Mine are like these, but they keep oscillating non-stop (which shows that the VGA is vibrating)
TKS A LOT FOR HELP US!
