On Wed, 4 Aug 2010 10:39:27 -0400, you wrote:

>On Wed, Aug 04, 2010 at 11:03:33PM +0930, Roger Weichert wrote:
>> If the Dataman is crook, why does the data read off all the chips ok and=
 the=20
>> checksum ok  ...  yet not one of the 14 I've tried have worked in it's=20
>> location.
>
>At a company I worked at many years ago (before I worked there):
>
>A unit was returned because it would not work for the customer.  So, the=20
>unit was opened up by a service technician, but it checked out to be=20
>functioning properly.  It was then returned to the customer to try again. =
=20
>Of course, the unit came back, and the technician tested it again - only=20
>to find that it still worked fine for him.  This time, however, the=20
>technician decided to test the unit after putting the cover back on (hey
>there's a novel idea..).  He found that the unit would no longer work when=
=20
>the cover was in place!  He then discovered that the EPROM, whose window=20
>was exposed inside the sealed box, required light to hit it in order to=20
>work.  So, like any good troubleshooter would do, he placed an LED inside=
=20
>the unit, aimed at the EPROM window, to offer the light that he thought=20
>the EPROM needed.  Amazed that this actually worked, the technician sent=20
>the unit back to the customer.  Several months later, the unit was=20
>returned again because, believe it or not(!), the unit was not reliable in=
=20
>the field.  This time, a different technician looked at the EPROM circuit.=
 =20
>He discovered that the EPROM ground trace had never been connected in the=
=20
>first place!  Yet many other units worked despite this PCB issue (possibly
>getting ground from low address lines?)..
>
>The impact that light had on the EPROM is still not completely understood,
>but someone mentioned something about the light helping to bias the memory=
=20
>cells?  I'm not sure, but I thought this story was funny from beginning to=
=20
>end!
>
>Take care,
>-Chris

ISTR that you could reliably crash a system running from EPROM by firing a =
flashgun nearby - I think
light messes up the read amplifier thresholds.

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