Tracy writes:

> I had erronously reported the disconnects after switching
>  to the latest version of QCTerm.  So now I'm also looking
>  for other causes.  One of those things could possibly be:
>
>  I had moved the machine from behind a firewall to outside
>  the firewall.  This was also about the same time I loaded
>  QCTerm 95a, so I may have got my causes mixed up.  I'm
>  beginning to think the former firewall helped maintain
>  connectivity, but how it does it, I'm not sure... do
>  firewalls emit "keep-alive" packets?

The one thing that tends to kill communications capabilities more than
anything else is poor grounding, especially where all of the devices are not
on the same ground and you have substantial impedence between the various
grounds, creating ground loop currents.

I earlier joked about why telephone lines sometimes hum, but that was
probably too subtle a comment. Telephone lines don't hum because they don't
know the words. Rather they hum because of poor grounds.

Cecile Chi mentioned that on rainy days see couldn't talk to a customer 12
miles away or to New Jersey, but everyone else was OK. We had exactly the
same problem here a few years back, with very much the same conditions.
Ultimately, after several months of tracking it down, the problem wound up
that the local central office, only a few blocks from us, had installed new
ESS'es (electronic switches) and had not properly grounded them all. When a
call was routed in one manner, everything worked. When it went through the
adjacent switch, we had a miserable, unusable connection to some local
customers. And the quality of the connection changed with the weather.  When
the problem was isolated and the grounds were properly done, everything
instantly began to work like clockwork.

We had virtually the same problem with a very large customer in upper New
York State, a customer so large that they have their own internal
telecommunications group. They too had a grounding problem and when the earth
froze every winter, we couldn't communicate with their HP3000.

These things are not easy to find, but ratty communications do have a reason.
There is no magic associated with any of this; and the ultimate reasons for
the problems can be found. Moving an HP3000, if only electrically and not
physically, can present a new set of grounds and a new set of problems. It's
entirely easy to hypothesize that your firewall was presenting a more well
grounded portal to the world than your HP3000 is now.

Wirt Atmar

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *