Switch Old Apple G4 to Apple LED Cinema Display via ATEN CS782DP

Gefen GTV-DVIDL-2-MDP Dual Link DVI to Mini DP Converter loses Sync after 15 minutes connected as PC2 through ATEN CS782DP KVM switch，but doesn’t lose Sync if connected directly to monitor. Apple 2001 Digital Audio G4 Radeon DVI output is the input to the Gefen GTV-DVIDL-2-MDP; the G4's output disappears from the Cinema Display when the Gefen's Sync LED goes out. PC1 on the CS782DP is connected to an Early 2011 MacBook Pro; the MBP's output shows fine at 2560 x 1440 on the Cinema Display whenever PC1 is activated via the CS782DP's remote Port Selection Pushbutton.

The Gefen converter syncs at 2560 x 1440, as shown by its lighting the second of the three EDID Mode LEDs， but the connection is actually at 1280 x 720 (this is one of the Cinema Display's resolutions, but the Gefen converter doesn't have a mode for it). I can use the remote Port Selection Pushbutton to switch the display and mouse and keyboard back and forth between PC1 and PC2，but—whatever the Port Selection Pushbutton is set to—after 15 minutes or so connected through the CS782DP the Sync LED on the Gefen converter goes dark and I have to reboot the G4 as PC2.

12 April 2016 I tried connecting the Gefen converter directly to the Cinema Display for a half-hour test，and that the display had then not gone black—demonstrating that the problem is between the Gefen converter and the ATEN KVM rather than between the converter and the display. I told this to Jeremy at ATEN, and put that info into ATEN's eSupport system. Jeremy suggested that I swap the computer video and USB cables between PC 1 and PC2; I did so immediately， and reported that the Gefen's Sync LED had turned on and the display—since the KVM was already switched to PC1—had started showing the G4's desktop after a five-second delay. Forty-five minutes later I phoned to tell Jeremy that the Gefen converter's Sync LED had again gone dark and the display—which had been showing the G4's desktop when I left—had again gone black in my absence. However the MBP's desktop displayed promptly when I switched the KVM to the PC2 port.

That afternoon (my NYC time, not ATEN's LA time) I also told Jeremy that I had just tried connecting the Gefen converter directly to the Cinema Display for a half-hour test, and that the display had then not gone black—demonstrating that the problem is between the Gefen converter and the ATEN KVM rather than between the converter and the display. Finally I summarized for Jeremy what my testing had shown: There is some kind of incompatibility between the Gefen converter and the ATEN KVM switch，but this incompatibility only shows up after the booted G4 has been connected through the Gefen converter to the KVM switch for about 15 minutes. The incompatibility is not a bad PC1 or PC2 port on the switch，and there is no incompatibility when the G4 is connected through the Gefen directly to the LED Cinema Display. There is also no such incompatibility between my MBP's Mini DisplayPort and the KVM switch.

Jeremy's hypothesis is that there is a subtle incompatibility between DisplayPort and Mini DisplayPort, which shows up when I use an adapter to go between the Gefen's MiniDisplayPort output and the CS782DP's DisplayPort input, and then use another adapter go go between the CS782DP's DisplayPort output and the Cinema Display's Mini DisplayPort input. I replied that I have no problem if I go from the MacBook Pro's Mini DisplayPort output into the CS782DP's DisplayPort input, and then use another adapter to go between the CS782DP's DisplayPort output and the Cinema Display's Mini DisplayPort input.

I told Jeremy "Instead, gambling on my hunch that the CS782DP is suppressing an EDID 'handshake between the source and the display', I've ordered an ATEN VC060 DVI EDID Emulator." (That order is through KVMGalore.) Jeremy replied "If the VC060 doesn't work, you'll likely have to return everything. I can't say why one set of adapters works fine and the other doesn't, but more adapters add more points of failure that can take communication outside specifications. Display Port based KVMs and non Display Port sources just don't get along all the time，and we have to ensure our devices adhere to Display Port specifications."

Do you think I'm right, and that putting a VC060 between the G4's DVI output and the Gefen's DVI input—with the proper resolution specified on the VC060—will keep the Gefen's Sync LED on and the G4's video output showing on the Cinema Display?

 

The problem with video from my G4 through my CS782DP still existed on 17 April, but I found out early that afternoon that the workaround is a single press on the Shift key. It turns out that, after I push the remote Port Selection Pushbutton to switch to PC2—when the Gefen is connected there—the keyboard connects to the G4 although the Gefen Sync LED remains off and the Cinema Display goes dark. When I then hold down the Shift key, the Sync LED turns on and the Cinema Display shows the G4's screen.

Except the workaround doesn't always work. Experimenting that night, I found that—although holding down the Shift key turns on the Gefen Sync LED when PC2 is activated on the ATEN CS782DP—the Cinema Display doesn't show my G4's screen unless my MacBook Pro attached to PC1 is either awake—not asleep—or shut down.

Then the afternoon of 18 April, as I was walking down Fifth Ave. from a midtown appointment, I further realized that I might also have System Preferences -> Energy Saver -> Put the display to sleep when the computer is inactive for: set to 15 minutes on my Digital Audio G4. I got home, and found it was so. With that reset to Never, the Gefen's Sync LED stays on indefinitely—whether or not the CS782DP is set to PC1 or PC2—so long as the G4 is booted and awake.

I now had no reason to open the ATEN VC060 box when it arrived 20 April and try out the the VC060. I got an RMA from KVMGalore, and mailed it back to them.

My only excuse for this mistake is that I installed OS X 10.3 on a third—SCSI—drive I put into the G4 with difficulty almost a year ago. Since my motivation at the time was solely to run Retrospect 6.1 on it, I didn't look closely at all the default settings in System Preferences. When OS X 10.3 was released in 2003, people were still using CRT displays that needed protection from burn-in and limited life.

However, when the CS782DP's remote Port Selection Pushbutton is used to switch it to PC2, holding down the Shift key does not cause the Cinema Display to switch from a black screen to what the G4 is displaying unless my MacBook Pro—cabled to the CS782DP's PC1 port—is booted and awake. The symmetrical situation is not true for the MacBook Pro cabled to port PC1; I can see its screen on the Cinema Display—provided the Port Selection Pushbutton has been used to set the CS782DP to PC1—whether or not the G4 is booted.

Speaking metaphorically, it's as if the weak G4 is happily playing alone in the playground when the strong MBP walks in. The MBP then says "I'm announcing that I'm the strong guy in this playground, so you can't play if I'm here unless I'm awake."

Nobody I've contacted at ATEN has a cure for this problem. Routing the USB connection from the G4 through a hub on the way to the CS782DP—as the replacement engineer suggested and engineer-1 reiterated—does not fix it; that only results in the mouse pointer being frozen when I use the remote Port Selection Pushbutton to switch the CS782DP to PC2—although the G4's screen appears on the Cinema Display.

A suggestion yesterday by engineer-1, however, ended up disclosing important Mac-related omissions in the CS782DP's User Manual. He/she suggested that I disable the Power On Detection feature of the CS782DP to solve the "weak and strong guys in the playground" problem, as described in the Hotkey Summary Table on page 14. That table says that all hotkeys start with [Scroll Lock] [Scroll Lock], but Mac keyboards don't have a [Scroll Lock] key. The Keyboard Emulation table on page 15 says that F14 is the Mac Keyboard equivalent for [Scroll Lock], but my pressing F14 twice only resulted in bringing up a brightness dialog in the Finder—with any subsequent key-presses per page 14 producing resounding bonking sounds.

Under Hotkey Port Selection on page 13, there is "Note: If using [Scroll Lock] conflicts with other programs, [Ctrl] can be used instead." Under that, Alternate Entering Hotkey Mode has two alternate procedures, the second of which is "Press and hold the remote port selector switch for three seconds. The entering hotkey mode hotkey is now [Ctrl]." After Glen walked me through that in a phonecall today, I was able to disable the Power On Detection feature by rapidly entering [Ctrl][Ctrl][e][Enter], and to then list the current switch settings by rapidly entering [Ctrl][Ctrl][F4][Enter] while in TextEdit. BTW, disabling the Power On Detection feature did not solve the "weak and strong guys in the playground" problem

All ATEN would have to do to save themselves a lot of user dissatisfaction and phonecalls is to change page 13 to say "Note: If using [Scroll Lock] conflicts with other programs, or if you are using a Macintosh, [Ctrl] can be—or must be for a Macintosh—used instead." Glen said today that the Keyboard Emulation table on page 15 applies only to using designed-for-Windows keyboards on Macintosh or Sun machines (or vice-versa, probably), which explains why [F14] [F14] on my designed-for-Macintosh keyboard did not cause my CS782DP to enter hotkey mode; that too should be in the User Manual.

All in all, I'm now happy enough with the CS782DP to keep it and use it. On 23 April I started a new ATEN eSupport case, referencing my old one—which engineer-1 closed out today because he hadn't heard from me since yesterday—to say that the "weak and strong guys in the playground" problem is not solved, and also pointing out the Mac-related omissions in the CS782DP's User Manual.