Message-ID: <128@redwood.UUCP>
Date: Wed, 16-Jan-85 23:04:59 EST
Article-I.D.: redwood.128
Posted: Wed Jan 16 23:04:59 1985
Date-Received: Mon, 21-Jan-85 03:35:42 EST
References: <816@pucc-i> <361@ahuta.UUCP>
Organization: [Consultant], Foster City, CA
Lines: 72
+---------------
| Thanks for posting the pin assignments. I'd like to make one small
| correction, not in the pins but the terminology. A "null modem cable"
| is NOT a modem cable without control leads. It's a cable that
| REPLACES a modem, so that two back-to-back terminals (or terminal
| and computer) can converse without benefit of a modem.
| For example:
| ...
| Request to send ------------------------------ Clear to send
| Clear to send ------------------------------ Request to send
| ...
| Dave Tutelman
+---------------
Quite true, but unfortunately you blew it with your example, which does NOT
emulate a pair of back-to-back modems. Request/Clear-To-Send (RTS/CTS) don't
work that way. First, you only need to cross-connect them when emulating a
half-duplex (or "controlled carrier") modem, since a full-duplex modem usually
has clear-to-send true whenever carrier-detect is. Second, CTS is a response
to the LOCAL RTS, not the remote's RTS. To properly emulate half-duplex use
of RTS/CTS, you have to make a local RTS cause both a local CTS and a remote
CD (carrier-detect). (The terms "remote" and "local" are symmetric.) Try it
this way ("X" is a cross-over, "+" a join):
7 GND --------------- GND
2 TD ------\ /------- TD
X
3 RD ------/ \------- RD
4 RTS --+--\ /----+-- RTS (I request to send; you get carrier)
| X |
5 CTS --+ / \ +-- CTS
/ \
8 CD ---/ \----- CD
6 DSR --+ +-- DSR (The dataset is ready when the terminal is)
| |
20 DTR -+ +-- DTR
Alternatively, to correctly emulate a pair of back-to-back full-duplex
modems (of the 103/212 style), use:
7 GND --------------- GND
2 TD ------\ /------- TD
X
3 RD ------/ \------- RD
4 RTS -(n/c) (n/c)- RTS
5 CTS --+ +-- CTS
| |
8 CD --+---\ /---+-- CD
X
6 DSR --+---/ \---+-- DSR
| |
20 DTR -+ +-- DTR
Many variations are possible, especially when non-standard flow-control
methods are used, but hooking RTS on one side to CTS on the other is
not one that normally makes any sense. Usually flow control is done
by cross-connecting DTR and CTS; that is, if the local D(ata) T(erminal)
is R(eady), the remote is C(lear) T(o) S(end).
Rob Warnock
Systems Architecture Consultant
UUCP: {ihnp4,ucbvax!dual}!fortune!redwood!rpw3
DDD: (415)572-2607
USPS: 510 Trinidad Lane, Foster City, CA 94404