Apple 23" Cinema Display possible fix for 'short-long-short' error

An explaination of one way to resolve an issue with an Aluminum Apple 23 inch cinema display (circa. 2005-2006) suffering from a 'short - long - short' flashing power light & no visible display. This solution worked for me, it may not work for you, but others have reported similar sucesses, see the forum posts in the links for more info.

Apple discussions…

Page 2 - A1082

The second post by A1082 has the fix, I shall detail it here along with a few other steps.

Thanks for the great info A1082 - he gets credit for this fix.

The issue

The monitor only shows the short-long-short flashing error on the power LED. No other signs of life - the display is black (the 'torch test' shows it is not a failed backlight). Apple System Profiler can see the monitor as a display device when connected to a Mac.

Trying a 150W cinema display power supply failed to make it work (others have said this is a fix for this error on the Apple discussions).

Troubleshooting

I found the 'Service Source' manual (contains take apart & trouble shooting guide for Apple service technicians) & followed it's troubleshooting steps. The end of the process suggests replacing the entire main board, or replacing the display. Apple only reccomend it's technical support replace entire parts instead of troubleshooting thier components. This is helpful if you have 1/2 the original cost of the display available to fix a 5 year old monitor. 

Instead I tried the 'pin hack' but I did so after opening the case. I removed the J1 connector & rewired two of the three pins to the board. See picture 1. To test the hack the LED & side switch panel need to be reconnected.

The monitor worked without the grey center pin wire connected. It seems to be a sensing line to prevent high drain displays destroying the lower rated supplies that can accidentally be connected because they have the same connector. 

The solution

After reading A1082's post on ehMac I measured the voltage of the 1117 regulator. See picture 2.

It was reading 4.2V on the output - not great for a 3.3V regulator. I replaced this with a higher rated version LM1086 & the monitor now works without the 'pin hack'. 

Removing the existing regulator is difficult, I removed the two legs first then managed to separate the casing from the heat sync tab. Because the heat sync is under the regulator body you need quite a lot of heat to desolder it. In the end I left the heatsync in place & connected the new LM1086 to the heat sync. The 1117 casing cleanly pulls of the heatsync.

Because the LM1086 is a larger component it needs to be placed away from the original location. I extended it to above the main board, and also heat shrinked the connections for insulation. Make sure any heat it generated will be dissipated away from other parts.

After a few hours troubleshooting & £12 (I brought 3 LM1086CS-3.3 regulators - just in case) in parts & postage the display now works.