Marcel Duchamp wrote:
> Light into eprom windows can and usually does affect the operation.  I
> have seen A/D counts vary when light shines into the window.

One of my earlier PIC projects wasn't working.  As I reached over to put a
scope probe on some pin, it started working.  Then I noticed that just
moving my hand nearby made it work.  I thought maybe MCLR was floating and
the capacitance to my hand was making it flip, but MCLR was driven well
enough.  It finally turned out to be that moving the hand between the PIC
and the nearby window made it work because that reduced the light hitting
the die of the JW part.  Since then I always put stickers over the windows,
even if just for short testing.

> And once had a system fail in a spectacular fashion from a camera
> flash
> going off.  In a development environment, the windows are usually left
> open ready for another trip through the uv eraser.  A windowed PIC was
> driving some fets for motor control.  The customer came to see how
> things were going and whipped out a camera and snapped a photo. Bang!
> All fets were turned on at once and shorted the nicad battery pack
> blowing out the fets.

Another PIC project was a rotating LED display.  I had just added the
feature so that a bunch of units could be strung together and all controlle=
d
from a single serial port.  It seemed to work, then everything got shipped
off to a show in Chicago.  At the show things worked until someone took a
picture of the booth.  All the displays went down instantly.

We were transmitting data from the base part to the rotating part by IR thr=
u
the center of the hollow shaft.  This data was just the UART output of the
transmitter.  It turns out there was a little gap just below the IR receive=
r
on the rotating board, and a camera flash was bright enough to get in there
and be detected.  The flash got interpreted as the reset command.

> In a related area, early video camera elements were made from DRAMs
> with out covers over the chips.

I think that was a later development.  I'm reasonably sure using DRAMs as
image sensors was pioneered by Bill Yerazunis of the Mitsubishi Electronic
Research Labs in Cambridge MA in the mid to late 1990s.  You may remember
him as "Crash" on the NERDS team of some Junkyard Wars episodes.

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.