hull@hao.UUCP (Howard Hull) (11/05/83)
We had an absolutely amazing accident occur here when a contract service
technician was working on a daisy-wheel printer. He claims to have
conscientiously disconnected the RS232 interface cable from the printer
while he was working on its power supply. We believe that was indeed the
case. He also claimed that the AC power cord was disconnected as well,
but we have a witness who saw otherwise. Our guy says that the printer
was "lights on" moments before. If so, then the printer was plugged into
a row of 115 Volt AC extension outlets, along with a nearby video terminal.
Everything was ok until the technician dropped a spacer and a screw into the
printer power supply. He mentioned that he saw a spark as the spacer fell
directly between a filter capacitor terminal and the frame of the printer.
Simultaneously, our PDP-11/70 dropped dead. Real dead. Two days dead.
Visible damage in the computer was for the most part confined to the MC1489A
line receivers on the RS232 multiplexer card. Curiously enough, the damage
was to the port attached to the aforementioned video terminal. The video
terminal was found to be inoperable after the accident. DEC field service
GRACIOUSLY fixed the problems in the processor. [!!!Hooray DEC!!!]
The technician's supervisor claimed that there was "noo waaaay" that the
technician could have caused the problem, since the printer was disconnected
at the time.
I could point out, however, that with the printer turned on and connected to
the AC outlet, a 115 Volt AC short between the AC High terminal of the RFI
filter and AC neutral and/or safety common could elevate the video terminal
AC neutral and/or safety common to as high as 57.5 Volts RMS AC, or more than
80 volts peak. Depending on the care that was used in routing the Pin 1 and
Pin 7 RS232 grounds, as well as the AC neutral and safety ground environment
of the computer, this can overstress the multiplexer line receivers, causing
them to fail in a variety of ways.
Those of you who are considering "2-wire RS232" should realize that you are
vulnerable to a very expensive and dangerous sequence of events in the event
of an AC Power fault. The safety grounds in our equipment are designed to
protect personnel *at the expense of possible damage to the equipment*.
{ucbvax!hplabs | allegra!nbires | decvax!brl-bmd | harpo!seismo | menlo70}
!hao!hullzben@umcp-cs.UUCP (11/06/83)
We have used a 4-wire RS232 connection scheme on the dinosaur (UNIVAC-1100) for years now. For economy we used a Radio Shack non-modular 4-post plug and jack set (remember them) that was much cheaper than DB25s. The four wires connected are transmit data, receive data, signal ground, and data terminal ready (which went only to a LED on the patchboard so you wouldn't accidently pull an in-use line). The patchboard has long since been replaced by a Gandalf PACX switch. You are lucky it was only the interface chip (1489). Then again it was designed to sacrifice itself in just this case. Sometimes it is the only chip socketed in a design (guess why). Before I realized this I always wondered why UARTs didn't just speak RS232 levels... That technician should be yelled at a bit. There have been stories here on the net about techs smoking while bending over an open disk drive. Shudder. Where do we get these people? Ben Cranston ...seismo!umcp-cs!zben zben@umd2.ARPA
padpowell@wateng.UUCP (PAD Powell [Admin]) (11/08/83)
That technician should be yelled at a bit. There have been stories here on the net about techs smoking while bending over an open disk drive. Shudder. Where do we get these people? Ben Cranston ...seismo!umcp-cs!zben zben@umd2.ARPA Most of these people are graduates of the outstanding U.S. High School system. Think about that and shiver. Patrick Powell