Dave Plowman (News) wrote:
In article ,
Paul wrote:
I interpret your description, as a resolution change. You had
one resolution before sleeping the computer, now it's selected
a different resolution, based on something that happened.

If I look at display settings, there is only a choice of two - both
ancient 4:3 ones, with this condition. And instead of the monitor name,
says KVM.


You have to follow the wiring and figure this out.

Take the monitor you hope to use and connect it *directly*
to the computer. Now, use Moninfo, and do a Real Time fetch.
Do you get a proper table from the monitor ?

Some monitors in the past, lacked write protect. It
was possible to overwrite the EDID and fill it with
garbage. "Verifying" the monitor then, is to check
that everything about that monitor is normal.

Now, with your various clever devices, you have
to figure out why a given item is not passing the
table properly. A KVM for example, could scan the
monitor, then present copies of the table on the computer
side. Each "box" in the path, presents a challenge.
What transform does it perform ? Can it foul up ?
What does a failure look like ?

Right away, a monitor name of "KVM" says the KVM could
not get an EDID from the monitor, and it substituted
a nicely formatted, canned EDID table it keeps for
special occasions. Now, you have to figure out whether
the KVM has a config switch or jumper to be doing this.
Or, figure out how the plumbing coming back from the monitor
has ruined the EDID info. If your DVI splitter is
between the KVM and the monitor, then the DVI only
allows one of its two output ports to be driving the
EDID function.

This little adventure is a lot like "buzzing wiring",
except the function involved is a bit more esoteric and
requires more test cases to reach a conclusion. Verifying
the monitor actually has an EDID table, is a necessary
part of this. The KVM cannot "invent" the correct table
for the monitor. and a "1024x768 KVM" table is highly
unlikely to make anyone happy.

Buying one of those boxes (sits in-line in a path where
the EDID isn't working) is certainly an option, but it
helps to understand what is killing your EDID in the first
place.

Paul