Nvidia RTX 2070 Super black screen hdmi output

Hello, I’ve been referred here by Nvidia support because I’m currently running Ubuntu 18.04 LTS. The issue has been ongoing and occurred on both my previous RMAed card which emitted noxious fumes at 70C as well as the replacement which only has this black screen intermittently but consistently.

The problem is the screen will go black for roughly 1-2s and then pop back a few seconds later. During this time the monitor which is an AOC 4K monitor does not show a signal loss, this occurs intermittently 1-10 times a day.

I’ve performed a number of steps troubleshooting this and this only occurs with the HDMI port, there are no issues on the DP port. I’ve re-seated cable connectors on both sides to ensure this is not an intermittent physical connection issue. I’ve borrowed a second high quality HDMI cable and confirmed the issue remains.

My primary use for the card will be in running some basic ML research, but having the display drop gets annoying and shakes confidence as these can be signs of prematurely failing hardware. For the price of the card I’d like to get this working and would appreciate any help you can provide.

To start off, I’ve read the post about asking for help it suggests running Xwindows with logverbose 6, I’m having difficulty finding the procedure for doing this with Ubuntu 18.04 LTS.

The core gnome and xserver components and associated configurations are prepackaged within snap packages which I don’t know how to access or change, nor have I found much documentation on how to move forward to get the needed visibility to run your bug report script.

I’ve attempted changing snap packages like this in the past without much, really any success.

Best Regards, Nathan

Sounds like the the issue is the monitor.

Hi Generix, I had thought of that, but I would expect, if it were the monitor I should be seeing similar issues when I swap out the video cards and the issue does not occur when using an AMD RX 560.

Please run nvidia-bug-report.sh as root after the blackout occured and attach the resulting nvidia-bug-report.log.gz file to your post.

Here are the results.
(upload://gEvRbcN9NrgDAnjQuwUPlwsdOOH.gz) (472.3 KB)

The upload didn’t work.

How about now?

nvidia-bug-report.log.gz (472.3 KB)

Not really anything visible besides a randr requery. Are there any converters/adapters involved?

No, there are no converters or adapters.

Its just a single cable and I’ve swapped the cable out just in case it was a bad cable. No joy.

Its unclear what the default verbosity is set to in these snap packages. The main post mentions having a verbosity of at least 6 but I haven’t been able to find what its being run at since most of the gnome system is filled in two-three interconnected snap packages.

The “verbosity” mentioned just translates to setting the “ModeDebug” driver option, just neccessary if the monitor is detected wrong or not at all. Doesn’t matter or help in your case.
Maybe it’s an electric problem, tried changing the multi-socket computer and monitor are connected to?

Well I had tried changing the monitor to another video card which had a different HDMI socket with the same computer if that’s what you mean. The issue didn’t occur when I did that.

If this is an electrical issue, it seems to be quite common as I had the exact same issue on my previous card (same model) which I RMA’ed about a month ago.

This is very discouraging that I’m having so much trouble with a brand new card that costs over $500, if it is electrical I’m going to have to RMA the card for a second time in a row. I don’t have the necessary test equipment to confirm an electrical issue for this.

My oscilloscope only goes up to 200mhz and TDMS signaling requires at least 225 though for a good picture it would need 4x that, I also wouldn’t be confident in placing my probes given the micro circuitry and earth ground reference of the circuit, nor would I want to void the warranty for an item this costly.

Should I reach back out to Nvidia support for steps for a second RMA?

I guess it’s some general sideeffect of your mainboard/gpu-board/monitor combo so I don’t believe another RMA will help. Was this always the same band/model after replacement?

For the RMA, yes it was the same brand and model. RTX 2070 Super, I was told during the RMA it was Founder’s Edition (it has the full card silver colored heat sink with the transparent sticker that needed to be removed from the reflective area on the bottom where the model name is listed.

Is it possible the HDMI cable may be grounding intermittently to the case, I was reading a post where someone suspected that might be happening since the HDMI port is placed so close to the top of the allowed footprint.

So, I’ve been working with support trying to get this and other hardware issues resolved. At this point I’m unable to return the card to Best Buy for a refund and it looks like the issue is going to be a persistent one, and its unclear what the problem even is with any specificity.

The monitor and card are both less than a year old, the mainboard is a little over a year old, and no hardware issues or defects when using non-nvidia hardware. I’ve gone through the whole gamut of troubleshooting, one RMA process, and now it looks like things are at the stage where I’ll just have to deal with a bad purchase?

What are the next steps for resolving this? Should I reach back out to Nvidia support (since they referred me here), or do I need to get Visa/Capital One involved.