> On Nov 17, 2020, at 7:07 PM, Grant Taylor via cctalk <cctalk@classiccmp.org>
> wrote:
>
> On 11/17/20 3:05 PM, Paul Koning via cctalk wrote:
>> That's true for either kind. Also remember to put the terminators only at
>> the ends. I've seen magazine articles showing a terminator in the middle
>> (on a T connector)!
>
> How?!
>
> The T is inherently three points; up-stream (cable), NIC, down-stream cable).
>
> Did they connect multiple Ts together?
>
> Did they use a BNC barrel on the unused NIC connection and add a terminator?
There was an article in the magazine RSTS Professional -- unfortunately I don't
have it any longer. It mentioned that you can connect thick to thin coax.
That is true; both are 50 ohm cable so you can use them interchangeably,
provided you derate the network to the lower of the two specs. BTW, that also
applies to my comment on transceiver placement. 10Base2 has no placement
rules, which means that a network no larger than 10Base2 limits can be built on
thickwire with placement ignored just as well.
What they got wrong is that they showed the splice being made by a T connector
that had a terminator on the third port. So a terminator -- which by
definition goes at the terminus, the end, of the cable, here was put in the
middle. That will mess up the signal integrity very badly.
You're right, any transceiver connection is a T; or as transmission line people
would say, a stub. Stubs are fine if they are electrically short, i.e., a
small fraction of a wavelength in length, and the end of the stub is high
impedance. Transceivers are high impedance circuits so this works.
A vampire tap is one way to build a short stub. The T connector attached
directly to the 10Base2 transceiver on the bulkhead is another. A thickwire
transceiver connected with an N/N/N T connector would also work; I haven't see
any of those used in practice. The key thing, as others mentioned, is not to
lengthen the stub. You don't put a length of coax between bulkhead and the
10Base2 T connector if you want the network to operate.
paul
