[ih] Fwd: Packet Radio Info (lossiness of Ethernet)

[Brian E Carpenter]

My favourite example was when one of the large experiments at CERN reported
that the Ethernet in their lab was suddenly very, very slow. Investigation
showed a very high rate of bit errors. Further investigation showed that
the screen of every single coax (this was "thin" Ethernet, not yellow cable)
had recently been soldered to ground throughout the lab. Since the other ends
had been grounded at the hub when first installed, this created multiple
ground loops and therefore a strong 50 Hz signal on every wire.

The quite senior physicist who had decided to do this explained to us that
shields should *always* be grounded, so he'd come in early one morning
to "fix" the "problem".

Regards/Ngā mihi
    Brian Carpenter

On 20-Apr-26 01:05, Alexander Schreiber via Internet-history wrote:
> On Sat, Apr 18, 2026 at 05:52:27PM -0600, Craig Partridge via Internet-history wrote:
>> My favorite Ethernet coax disaster was a friend who managed a network, I
>> believe at Northeastern, where they had much higher packet loss.  After a
>> few days they discovered a section of coax (with unrepaired tap holes) had
>> been reused to extend the coax through a janitor's closet -- and the
>> cleaning product fumes had caused wonderful copper+[something] crystals to
>> form in the coax.
> 
> During university, I was part of the group that ran the dormitory networks.
> That was quite a learning experience. The fun things we saw with coax
> Ethernet (10Base-2) e.g.:
>   - New student moves into their room, gets handed a short coax connector
>     cable to wire their machine into the network, shortly after, that
>     segment fails. Analyzer claims the segment (which was slightly over
>     the length limit before) is now way beyond the length limit. WAT?
>     Visit that students room: Instead of the short connector cable, they
>     had used a IIRC 50m roll (why they even had that was unclear) of
>     coax for that purpose. That was unplugged and replaced with the short
>     connector cable, segment works again.
>   - Suddenly, one coax segment becomes intermittently unreliable. We
>     eventually discover one student using the coax cable in their room
>     as a clothes line. The resulting variable mechanical load made
>     for ... variably well working connectors. After some educational
>     talking to, that problem disappeared.a
>   - Removing the terminator when you leave for semester holidays (because
>     "obviously" it's yours (no, it's not) so you take it home with you)
>     will not improve your popularity with your fellow segment dwellers.
> 
> Lots of shenanigans ...
> 
> Kind regards,
>             Alex.