home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.dcom.modems
- Path: sparky!uunet!cs.utexas.edu!chinacat!balkan!wrangler!bill
- From: bill@wrangler.WLK.COM (Bill Kennedy)
- Subject: WB<->Intel cultural differences
- Message-ID: <1992Sep1.011410.731@wrangler.WLK.COM>
- Organization: W. L. Kennedy Jr. & Associates
- Distribution: na
- Date: Tue, 1 Sep 1992 01:14:10 GMT
- Lines: 45
-
- I have been beating myself bloody with this one and maybe the collective
- wisdom can confirm/explain what I have arrived at with brute force. I
- have a uucp neighbor equipped with at Trailblazer Plus and an Intel V.32bis
- modem. We communicate flawlessly with my WorldBlazers uucp and (I think)
- cu. About ten days ago we tried to bring up a serial IP link using the
- Morning Star Technogies package. I've never had any trouble with the MST
- stuff using T-2500s, T-1600s, T-3000s, QBs, or WBs amongst each other on
- the Sun SPARC or ISC UNIX (v1.3 TCP/IP please) platforms.
-
- Each time that I would make a connection WB->Intel the login interview
- would proceed successfully and as soon as the serial IP stuff took over we
- would get massive retrains and then a disconnect. On my end originating we
- would get PPP started and I'd send the stuff to negotiate the link but the
- Intel end was mute. If I fell back to SL/IP (not in the same call :-) the
- link would get declared up but the Intel was still mute. Last weekend my
- neighbor decided that he would try it with the Intel side initiating the call
- but he saw the same problems with PPP that I did, but he was able to connect
- and communicate with SL/IP. You're probably asking yourself if I'm posting
- to the right group, read on, please.
-
- Sunday I decided to try the same connection to the Trailblazer but at V.22bis
- 2400bps. It worked like a champ. This evening I decided to try it PEP to
- the Trailblazer and again it worked like a champ, modulo what PEP does to
- serial IP connections. Then for the sheer hell of it I decided to try the
- Intel one more time but this time I stuck an S50=7S94=2 in the dialer to try
- and avoid or omit the retrains that were coming every time. Voila! It worked
- on the first and each subsequent attempt. I had already concluded that this
- was some kind of cultural problem between Telebit and Intel V.32bis (aren't
- standards grand? :-) but this seems to confirm it.
-
- Does this make sense to anyone? Has anyone seen similar problems between the
- WorldBlazer and an Intel V.32bis modem? Is this a viable workaround if you
- have? I'm just as curious as to why serial IP aggravates it when uucp doesn't,
- but that really _is_ for another group. I'm guessing that the WB decides that
- the link is so good that they should renegotiate within the modulation scheme
- but the Intel doesn't follow and all the wheels come off rather quickly. By
- setting S50=7 (insist on 14,400bps) and S94=2 (don't renegotiate anything) I
- have defeated what ever was defeating me. Should I set S94=2 so that when I
- answer a call from a non-Telebit the WB doesn't try to do something with the
- other end that the other end can't do? I should repeat that among Telebit
- modems this doesn't happen at all, ever, since I've been watching. All the
- world is not a Telebit but life seems simpler when it is...
- --
- Bill Kennedy bill@tyche.WLK.COM | Real friends don't let
- | friends drive Novell
-