-----Original Message-----
From: Brent Regan <brent@regandesigns.com>
To: lml@lancaironline.net
Sent: Mon, Aug 8, 2011 8:40 pm
Subject: [LML] EFIS "Failure"
When we create the conformal document set (drawings, specifications, etc.) for certified hardware we try to lock down every aspect of the design so that vendors do not "improve" a component and cause an unforeseen problem. One man's improvement can be another man's nightmare. Case in point.
A friend with a Chelton IDU-I manufactured about 10 years ago (by Sierra Flight Systems in Boise) was having a problem where the display would intermittently lose vertical synchronization and start flipping. I disassembled the unit to find the problem. It seems the LCD display vendor, Panelview, who also added the Anti Reflection (AR) coating, decided to add a metalized sticker to the display. Seems innocent enough, right? Well, with time and temperature cycles the sticker started to peal and came in contact with the backside of an inter-board connector that happened to be carrying the V-SYNC signal. With heat and vibration the aluminum layer of the sticker was exposed and shorted the signal pins. The attached pictures clearly show where the pins rubbed on the sticker.
A freaking sticker?!?!?! Really?!?!?! That is one for the books.
It is not the known knowns that get you. It is rarely the known unknowns that get you. It is the unknown unknowns that get you.
It is experiences like this that are the basis for my advice that, despite your best efforts, sooner or later it may all go very wrong. And not for a big obvious reason but for a stupid little tiny sticker of a reason. Ask your self now, in the comfort of your office chair, what is your plan for when it does go very wrong?
Regards
Brent Regan