Path: utzoo!utgpu!water!watmath!clyde!att!rutgers!mailrus!cornell!rochester!ritcv!moscom!jgp
From: jgp@moscom.UUCP (Jim Prescott)
Newsgroups: comp.dcom.modems
Subject: Re: null modem cables
Message-ID: <1261@moscom.UUCP>
Date: 14 Sep 88 07:00:25 GMT
References: <8809030943.AA02215@ucbvax.Berkeley.EDU> <2066@stpstn.UUCP>
Reply-To: jgp@moscom.UUCP (Jim Prescott)
Organization: Moscom Corp., E. Rochester, NY
Lines: 37

In article <2066@stpstn.UUCP> aad@stpstn.UUCP (Anthony A. Datri) writes:
>For many things, a null-modem cable merely means a cable wired thus:
>
>2----3
>3----2
>7----7
>20--20

I think this is bad.  Since null modem connects two DTE devices it isn't a
good idea to tie lines asserted by DTE's together; in the above cable both
sides will be asserting 20 (DTR).  I've heard that having both sides try to
assert the same line results in a greatly shortened lifespan of one or both
sides; anybody know if this is actually true?

Just dropping 20 is probably fine for a minimal null-modem cable.

>A full-blown, general-purpose null-modem cable would be thus:
>1---------1  (always friendly to include the frame ground)
>2---------3
>3---------2
>4---------5
>5---------4
>6--------20
>7---------7
>8---------8  (often varies, often not needed)
>20--------6

Tying pin 8 through isn't as bad since it is DCE but it probably won't have
any effect (both sides are reading the pin, nobody writes it).  I usually
find 6,8 -> 20 in both directions to be best.

This assumes that the devices at least follow the distinction between DTE
and DCE lines.  Are there many devices that don't (ie assert some DTE and
some DCE lines)?
-- 
Jim Prescott	moscom!jgp@cs.rochester.edu
		{rutgers,ames,harvard}!rochester!moscom!jgp