Second monitor connected as DVI enters sleep mode during boot (and never comes out of sleep); but works with VGA connection

emacsomancer · 05-20-2019, 03:26 PM

Quote:

Originally Posted by Shadow_7

You might see if that monitors menus has a select input source. And select DVI of course. I've had issues where one DVI monitor would power down to save power and on restore it would select another mode. Although I had two different PCs hooked up to it at the time. Both on. Probably not your issue, but something to look into.

Yeah, I've tried the manual menus. DVI says 'no input'.

Quote:

Originally Posted by mrmazda

Reboot is not necessary, but restarting Xorg is.

I seriously botched that request. This is what it should have been:

Code:

hwinfo --monitor | grep -A2 "Max. Resolution"

So this is the output:

Code:

$ hwinfo --monitor | grep -A2 "Max. Resolution"
    Max. Resolution: 1280x1024
    Vert. Sync Range: 50-77 Hz
    Hor. Sync Range: 24-83 kHz
....

I tried first changing the `HorizSync` of the xorg.conf to `24-83`, and rebooted with only the DVI connection of L1750 plugged in and all other video connections unplugged, but with no change in behaviour. I then changed `VertRefresh` of the xorg.conf to `50-77` and rebooted. Again, no change.

As you said, it's not clear what the path forward would be from here.

I'll have a different video out to test the monitor from later today and will test that, though I'm dubious that that could be issue.

I contacted the place I got the monitor from and told them of the situation and they said their tech people said there was probably a 'small flaw or short in the video driver' of this particular monitor and that they would send out another.

emacsomancer · 05-24-2019, 04:03 PM

Tried now another DVI cable, and a different monitor of the same model, and unfortunately still cannot get the DVI connection to work.

mrmazda · 05-24-2019, 04:15 PM

Quote:

Originally Posted by emacsomancer

Tried now another DVI cable, and a different monitor of the same model, and unfortunately still cannot get the DVI connection to work.

This shows consistency for that HP model, not whether the model has a problem with all DVI outputs, or just with yours.

emacsomancer · 05-24-2019, 04:44 PM

Quote:

Originally Posted by mrmazda

This shows consistency for that HP model, not whether the model has a problem with all DVI outputs, or just with yours.

That's true. Although I have tried plugging into an MSI RX 480 and an MSI RX 470 - similar cards, to be sure, but at least two different actual bits of hardware.

mrmazda · 05-24-2019, 07:06 PM

RX470 and RX480 are so close it's easy to expect the only difference(s) between them wouldn't involve whether or not they completely support DVI specs correctly. Inxi declares they're the same, but so are the HPs, so I still don't think definitive testing is yet complete. I'm leaning on the problem ultimately lying in the HP. https://answers.microsoft.com/en-us/...6-72aa4efa79e2 suggests that HP is less than fully competent absent a Windows driver. https://answers.microsoft.com/en-us/...e-1e6d9e723a88 seems to confirm the HP defective.

There is such a thing as loading a modified EDID into RAM after boot to overcome a defect in a display's EDID report, like loading a kernel module but for Xorg drivers to use. I have no experience with it, but it could be your ultimate salvation if you can't get that HP swapped with some other model known to work with those RX models.

emacsomancer · 05-24-2019, 08:04 PM

Quote:

Originally Posted by mrmazda

RX470 and RX480 are so close it's easy to expect the only difference(s) between them wouldn't involve whether or not they completely support DVI specs correctly. Inxi declares they're the same, but so are the HPs, so I still don't think definitive testing is yet complete. I'm leaning on the problem ultimately lying in the HP. https://answers.microsoft.com/en-us/...6-72aa4efa79e2 suggests that HP is less than fully competent absent a Windows driver. https://answers.microsoft.com/en-us/...e-1e6d9e723a88 seems to confirm the HP defective.

Right, I noticed that inxi declares them to be the same (I think they probably largely are, modulo 'binning'). But, at least, it doesn't seem specific to the individual (token) video card. It does sound like it's an issue the HP monitor itself from the links you include. (A pity, the monitor physically seems very nice.)

Quote:

Originally Posted by mrmazda

There is such a thing as loading a modified EDID into RAM after boot to overcome a defect in a display's EDID report, like loading a kernel module but for Xorg drivers to use. I have no experience with it, but it could be your ultimate salvation if you can't get that HP swapped with some other model known to work with those RX models.

This looks like something worth trying. I mean, it's usable with HDMI->VGA (it looks clear to me), but I'd still *like* to get it working via a proper DVI connection.

I think I was able to get the parsed EDID information out from the monitor via following this: https://unix.stackexchange.com/a/470122/3857 , but I wonder what needs modifying. Do you have any suggestion of where I might look for information on how to proceed from here?

mrmazda · 05-24-2019, 08:43 PM

Quote:

Originally Posted by emacsomancer

Do you have any suggestion of where I might look for information on how to proceed from here?

Other than asking for help on xorg@lists.x.org, I do not. I've never needed to, since xorg.conf always worked around the few times I needed a workaround.

IMO the problem justifies returning HP to vendor as defective/incompatible if a free return or exchange period is available.