Video Wall Blues

So…we have a video wall at the Siebel Center for Computer Science at the University of Illinois, Urbana-Champaign.  And for some time now, it has been a source of annoyance.  We inherited the processing unit that drives the wall, and the eight DLP panels that make up the display itself were donated–all just in the nick of time before the building’s grand opening where the generous benefactor, Tom Siebel, would be making an appearance.  Because the installation was so last minute, we received very little training, and had to figure out a lot on our own.  We’ve found the software that controls the processing unit is extremely clunky and unwieldy.  And in four years, three of the display panels have already broken down–requiring the same replacement parts–and resulting in repairs running about $10K each.  *shudder*

A technician from Chicago came down on Friday to repair the latest malfunctioning panel, and we discovered–upon restarting the wall–that there is now a problem with the central processing unit.  We can no longer display anything on the wall, and the unit itself no longer responds to any manual input commands.  I spent the better part of Friday afternoon trying to solve the problem, in addition to most of yesterday morning.  The result of my troubleshooting was that something is seriously wrong with the central processing unit.

Now, since the wall is comprised of hand-me down and donated parts, there is no budget line for repairs or upgrades…  So we’ll have to be “creative” with an alternative…  With the assumption that we most likely won’t be able to afford a replacement CPU, we’re exploring other ideas–specifically, one my colleague has been contemplating for some time now as a “back-up” plan.

Today, he and I, with some testing and experimentation, (and after discovering that the cabling from the CPU to the panels is an unusual 9-pin to 15-pin VGA cable–*arrgh*), we discovered that each of the DLP displays will show VGA resolutions ranging from 800×600 to 1280×1024–though the latter looks very grainy, but 800×600 looks quite good.  So the plan now is to devise a way that we can drive the central four cubes with one computer with two dual-output video cards, and have the two left cubes and two right cubes driven by repurposed machines.  With this plan, we will lose some of the functionality allowed by the previous configuration that incorporated the CPU, but we will gain some flexibility by having the panels be controlled independently.

We’ll see what shakes out…