Tripple Monitor Frustrations - System Crashes

I have a 3 console / terminal setup, meaning that I have 3 KVMs (literally) in different rooms that connect back to the same PC using Displayport over MPO fiber extenders. For USB I’m using USB over CAT 6 adapters.

I have 3 monitors, all connected over these Displayport extension cables.

Samsung G95SC
Samsung S95D (TV)
Samsung Odyssey G9 Neo

The issue comes when I want to switch between terminals. At first I was using smartplugs to cut power to the terminals not in use, but after I blew up a power supply in one of the monitors I decided to stop doing that. That method actually worked okay though.

Now what I’m doing is leaving everything powered up and switching the active monitor in Windows 11. This is really cumbersome, but I basically have a fingerprint reader on each terminal’s keyboard, which logs me in with Windows Hello. After presumably logging in, I use Win+P and the up and down arrows + Enter to switch from primary to secondary monitor. This doesn’t account for the 3rd monitor, the TV, which I don’t use very often thankfully. To switch to this display I have to painstakingly navigate the Windows display options to enable it, change it primary, then disable the others.

While cumbersome and annoying, I was willing to live with this arrangement. The problem that I’m having now is that using this system frequently causes the system to lock up when switching monitors. I have no idea why this is happening, and I have no ability to diagnose what’s going on. All I know is that I get no output on any of the displays (they think the system is asleep) and that the num key LED doesn’t respond. That’s my only indication that the system is locked, pressing Num Lock doesn’t toggle the light on the keyboard.

I have narrowed it down to the Odyssey G9 Neo being the likely problem. If I don’t have it connected I don’t have these issues. Conversely, if the OLED displays are not connected and the Neo is the only display connected, I do have this issue.

I cannot replicate the problem reliably, but it’s most likely to occur when all 3 displays are connected. One strange behavior that I noticed was that I wasn’t able to duplicate the display on the G95SC and the G9 Neo (the suspected problem display) despite them being the same resolution and supporting the same refresh rate. I was able to fix this one day (being able to duplicate the displays), but I had to delete some registry keys. I think it had to do with overscan or something like that. Also I think the Neo might have to use display stream compression to achieve 240hz while the OLED G9 does not? Not sure if that’s contributing to this problem or not. This is all I know right now

I know this is an extremely exotic setup, but hey, that’s what we do around here right? I have a lot invested into this setup and I can’t switch between them reliably at this point. It’s a huge issue if I need to jump on a call early in the morning just after I get ready.

I’m desperate at this point. My only recourse right now is to go learn the entire display protocol architecture and try to find some way to get enough diagnostic information to troubleshoot further.

If you guys have ANYTHING I can try I’m willing to give it a shot at this point.

Use Nirsoft’s nircmd to setup the displays the way you want in a rather simple script: https://www.nirsoft.net/utils/nircmd2.html#setdisplay setdisplay {monitor:index/name} [width] [height] [color bits] {refresh rate} {-updatereg} {-allusers}

I cannot replicate the problem reliably

Maybe with a looping script you could try to?

A driver deadlock shouldn’t be your issue to fix. I suppose before you can reliably bug the hell out of Samsungs and GPU’s support you’ll need to replicate it on a fresh Win10 or 11 install with the latest drivers. Even then there’s no promise of getting through Level 1 support. Well maybe this will be enough preparation?

IMO this is a GPU driver bug and you should post your current hardware and software versions here too just for completeness.

I’m desperate at this point. My only recourse right now is to go learn the entire display protocol architecture and try to find some way to get enough diagnostic information to troubleshoot further.

This is not your duty. I fall into this mind trap myself, given how useless support is on average. To the point I don’t regard “good support” as an option/buying decision.

I suppose a lot of support representatives will point out your “exotic setup” along the way, but the fact of this situation is: your system (drivers) should not be crashing under any circumstances. We are long past Windows 95 days. If you had a truly reliable way to trigger this soon-ish, this would add a lot credibility to your case especially with an engineer on the other end.

One last thing to note: I’ve seen myself how Windows “collapsed” 60 Hz in one menu whereas Radeon’s driver showed 59.99 Hz (or vice-versa). If you wanted to find this out, you’d need to dig deeper into actual settings applied per-monitor etc. In my case this discrepancy had also caused slightly higher power consumption (1-3W best case) by the GPU, although both monitors supported 60.00 Hz fine.