🚀 go-pugleaf

In article <sren28ijnot.fsf@ptavv.es.net>,
Kevin Oberman  <oberman@es.net> wrote:
>For_Email@ddress.see.below_plea.se (Philip J. Koenig) writes:
>
>> The reason that using correct cable is important are as follows:
>> 
>> 1 - Terminators and NIC's are 75 ohm, if the cable is not, you end up
>>     with reflections that play hell with the waveforms and thus, data 
>>     integrity.
>
>Terminators and NICs are 50 ohms for 10Base2 (Ethernet). 

Yep.

What also comes into it, and most people ignore it, is that there are
multiple budgets in thinnet Ethernets:

1.	The total cable length and propagation velocity through both the
	wire and repeaters (the rule being that a transmission of the
	preamble from one end must reach the other end BEFORE the preamble
	length in bit-times expires, or you can't detect a collision).

2.	The length of the *STUBS*.  Each NIC and connector to it is in fact
	a stub.  Each stub generates a reflection of the signal which is, in
	part, based upon its length (and other factors way too complicated
	to worry about).  Even though you don't think about a T-connector as
	being a "stub", it *IS* one, and so is the physical connector on the
	NIC.  Exceed this limit *in total* and the network starts acting
	VERY strange (certain hosts can't see each other, etc).

3.	ANY discontinuities (impedence mismatches) between pieces of wire
	(and there is a small but present one at each connector!) cause
	further reflections due to impedence mismatch.

That's for starters.  

Frankly, thinnet is a pain in the ass.  ThickNet (10Base5) with Vampire taps
at least works reliably.  Thinnet, in my experience, NEVER does across large
installations (hundreds of hosts).  All it takes is ONE person with a
desktop machine that moves it and "stretches" the cable a bit, causing a
loss of good continuity in one of the connectors (an impedence mismatch) 
and you have trouble.  Frequently the trouble will show up somewhere
*different* than where the actual cable problem is.  

The only way to KNOW where the discontinuity happens to be is to have both 
a network map (with cable lengths documented for each piece of wire -
literally) and a TDR.  Given those two things you can find and fix Thinnet
problems.  The TDR has to be used during installation of the cable to get
the lengths (one function of a TDR is to give you the length of the cable
from one end to another, given a proper terminator on one end and the TDR on
the other).  With that, and the use of the TDR when you have trouble, you
can pinpoint the bad connection.

Without those *TWO* things you'll be chasing your tail on thinnet problems
forever.  I know of a few people who have both TDRs and a Thinnet network -
I know of VERY few who took the trouble to map the lengths of all the 
segments from tap-to-tap during installation, and without that information 
a TDR doesn't do much when troubleshooting (ie: the impedance discontinuity
is 57' down that wire - which of the following six workstations is exactly
57' from where the TDR is on the wire right now?!).

These days UTP is so cheap to install that you have to be smoking funny
things to use anything else.

--
-- 
Karl Denninger (karl@MCS.Net)
Voice: 312-803-6271 x219