I am hoping someone can help here as HP's support has been rather slow. My company recently bought some HP T620 thin clients that run HP's Thin Pro 5 OS. The OS is based of of Ubuntu but shrunken down quite a bit. The Thin Clients only come with 2 Display Ports so we are forced to use VGA adapters which we also purchased from HP. The problem we have encountered is that the display will stop working with about half our monitors. The thin client will boot up and show "Starting Thin Pro" and then the screen flashes and the monitors states that the cable is not connected. We have checked connections and they are all secure. If you power down the thin client and then connect it to a different brand of monitor it works just fine. The monitors we are having trouble with are Mitsubishi DiamandPoint V50LCDs. These are the only monitors that have this problem. What's really odd is that if you image the OS while it is hooked up to the Mitsubishi monitors it will work just fine for a while and then all of a sudden show the same error. Hopefully someone can help!

It sounds like a monitor or connection problem to me. All the Mitsubishi monitors have similar cabling and p[lugs. pins could be sliding back in the plugs, or sockets, or adapters. If the adapter changes temperature, you could lose a connection there. Can you try one on a different box through all the adapters you currently use. Does it stay on.

I am pretty sure it is not a connection problem. It happens on multiple setups. What is also strange is that if I image the machine while it is connected to the Mitsubishi monitors it will work for a while then just stop working. In the Xorg log file there are a bunch of lines stating the vrefresh sync is out of range for every resolution type.

Xorg does strike out some very silly values always - I wouldn't worry about that. It found something, or it would not have started.
Are we agreed something is not right with the monitor? If not what do you suspect? Does the monitor remain lighting after it is "disconnected?"

Yes, I am sure it is not the monitor as it works just fine with our older T610 HP thin clients, although those are connected to a DVI port using a DVI-to-VGA adapter.

OK. So it starts working, stops/goes off /acts disconnected /does something you don't want it to do.
Eliminate the monitor, you say. In my book, you NEVER Eliminate anything until you find the problem.
If you eliminate the monitor, It's power saving on the graphics or it's a pure mystery.

Solve the mystery by defining a test (e.g. connects & remains on for 30 mins) and try
1. the pc and some other monitor
2. The monitor and some other pc
3. The pc and the monitor.

Come back with the results.

Like I said, I know the monitor works, it has been used heavily.
After doing some testing I think I have found the root of the problem but I am not sure on how to fix it. I setup ssh so that I can log in when the monitor is not working. When the monitor works xrandr reports it as DFP1. When the monitor is not working, xrandr reports it as CRT1. Is there a way to force the use DFP1?

There is/was a feature in all CRT monitors known as beam limiting. There is a crowbar like cutout if the HV beam current is exceeded - to protect tubes and the like. Running a very high horizontal rate can trigger it. That might explain why it only happens on one box.

No idea on the DEP1 & CRT1, except that they are obviously different.

I'd like more detail on what exactly the monitor looks like when it's 'dead'

When the monitor isn't working it has the error "cable disconnected" moving across the screen continuously.

That's a hardware error. It may be only one combination of monitor, pc and cable produce the error. That sort of thing happens because logic levels have a 'dud space' in them. On standard 5V logic
Below 0.8V is low (guaranteed)
Above 2.0V is high (Guaranteed)
What happens on the way up between 0.8 and 2V is unspecified. Many chips are high by 0.9V; some don't go high up to about 1.4V. I met one that needed 1.8V once it was brought low. The region can be temperature dependant. 74HC chips sometimes need more, although 74HCT are fine.
What happens on the way down from 2.0V is equally uncertain and different. It is the job of the designer to avoid these areas.

Some driver or input is in poor condition or overloaded, or somebody has failed. A resistance in the cable could also cause it. There is no convenient fix.