subject: static shock crashes computers
posted: Sat, 04 Mar 2006 11:09:10 -0000


[I did it myself, twice! Zipping around the office at a customer, I
must have worked up a static charge - I returned to the computer room
and went to touch the keyboard - my hand got near the metal shelving
the server sits inside, and zap! I got a static shock and the server
blue-screened. I did it twice in a week. It's been cold here, and
the heaters have been on. I've never really seen static nuke
something like that before, custs lost their unsaved docs and all. I
think I'll start grounding myself before I go near the server in
future. And consider jetting the nylons.. - Stu]

http://computerworld.com/news/2003/story/0,11280,87142,00.html

Shark Tank by Sharky

NOVEMBER 14, 2003 (COMPUTERWORLD) - Flashback to the late 1970s, and
the fall day when a new Texas Instruments 990 minicomputer is
delivered to this company for a laboratory monitoring system. And
it's a beauty, according to the pilot fish responsible for it.

"The new 990 was a sleek system mounted in the pedestal of a desk,
replacing an older system that was rack-mounted," fish says. "It was
installed as part of the monitoring system in late September, and
tested and approved. In October, data entry started."

That's product-testing data, which comes from computers in another
building. "One of the programs on the new system was an analysis
program that provided statistical results regarding product testing,"
fish explains.

"Attempts to link the systems had been rejected because of costs, so
our lab would receive the information on printouts and they would be
re-entered into our computer."

Everything runs fine for the first few weeks, as lab technicians
regularly key in the testing data.

But then the problems begin. "One day, one of the lab techs reported
that the computer had hung, and could I check it out?" fish says. "I
could find no problem on the logs after reboot and put it down as one
of those random occurrences."

Until the next failure. And the next, and the next. The colder the
weather gets, the more often the 990 freezes. "Since the problem
mainly occurred only while the test data was being entered, the
vendor inserted traps and logging points, rewrote various parts of
the program and spent a fair amount of time trying to identify the
problem," says fish.

"During the problem-solving, it was noted that most of the failures
occurred when the two female lab technicians were entering the data.
They were retrained, but the problems continued."

Finally spring arrives, and the crashes happen less frequently, and
then finally clear up almost completely. No one knows exactly what
was causing the problem. But everyone figures that one of the many
vendor fixes must have resolved it. And by late summer, the problem
is almost totally forgotten.

Then, in late fall, it starts again.

"We were crashing at least once a day," says fish. "And 95% of the
crashes were occurring while the female lab technicians were on the
system."

One day, fish is working at the 990's keyboard, on a chair he doesn't
normally use. "The chair had a nylon cover," says fish, "and when I
stood up, I felt a static shock.

"The computer promptly crashed."

Fish restarts the machine, and then slides out of the chair again to
generate another static spark. Once again, the lights on the computer
stop.

And fish gets it. "Even with the grounding of the raised floor, the
dry air during the heating season was generating a static discharge
high enough to shut down the computer," he says.

"The problem with the female techs? Turns out that they normally wore
nylon stockings, in compliance with the company's dress code. The
nylons, unlike the woolen pants most of the men wore, would build up
a charge as they moved around and, once high enough, would zap the
computer.

"My boss approved my emergency purchase of a ground static chair
mat."

---
* Origin: [adminz] tech, security, support -
http://cyberdelix.net/adminz/

generated by msg2page 0.06 on Jul 21, 2006 at 19:03:34

 search: