[ale] tech "magic"

[Neal Rhodes]

> Everyone here has had it happen. Something fails and you get called to
> "take a look at it". You get there. take it and it isn't working. So
> you methodically take it apart, inspecting everything alonmg the way
> looking for failure points and find

I keep swearing I really do software, but somehow I've spend a fair
amount of time in factories and other grungy places.   

Yesterday I was installing a new server in a warehouse of vintage
British Motorcycle parts.    Had two Epson printers that were failing to
power up. 

My experience is that dust, especially factory dust, CAN CONDUCT some
electricity. 

With both printers, after the client was ready to take them to a repair
shop,  I removed the covers, blew out the dust really good, checked the
fuse, put it back in, and wiggled the board connector.     Voila.  My
diagnosis - contact resistance and slighty conductive dust bunnies. 

Then the owner's desktop fails to come up, beeping a single long bleat.
Again, owner about to drive it to Windy Hill, the dealer said that was
battery bad.  I said let me take a quick look.  Re-seat the RAM, tested
the button battery, it was fine, wipe it on my shirt, stick it back in,
and voila, it comes up fine. 

As a result, my theory is that if it's not soldered, it's a suspect
connection. 

Some of us old enough to remember Altos boards with 7400 series logic
can remember that the chips would try to crawl out of their sockets at
night, and first thing to do after blowing out the dust is push all the
chips back down.   Glad those days are over. 

Neal Rhodes.