home *** CD-ROM | disk | FTP | other *** search
- Path: gate.net!not-for-mail
- From: dhaire@gate.net (doug haire)
- Newsgroups: comp.dcom.modems
- Subject: Re: Supra Modems
- Date: 17 Jan 1996 22:40:27 -0500
- Organization: CyberGate, Inc.
- Message-ID: <4dkffb$fuju@navajo.gate.net>
- References: <DL4H42.H2K@news2.new-york.net> <4d9ksu$6fj2@navajo.gate.net> <4dk8sv$h0@news.accessorl.net>
- NNTP-Posting-Host: navajo.gate.net
- X-Newsreader: TIN [UNIX 1.3 950824BETA PL0]
-
- Butt-head (eric@eola.ao.net) wrote:
- : doug haire (dhaire@gate.net) wrote:
- : : brag about. It has what appears to be a problem with not inhibiting the
- : : initiation of a retrain while negotiating error correction. Supra is
- : : currently in denial about this, claiming it doesn't happen, but the
- : : factory default is to inhibit retrain requests (%E0). That means it is
- : : dependent on the remote modem to monitor line conditions and request
- : : retrains. If you set the modem to monitor line conditions and request
- : : retrains (%E1), you then run the risk of it requesting a retrain when it
- : : shouldn't.
- : This is NOT the case. The retrain that occurs when you actually LOSE
- : CARRIER for awhile is different from the retraining that occurs when the
- : modem "falls backward" or "falls forward". The modem will request
- : retrains whether you are set on %E0 or %E1, but they are the fall
- : backward/forward kind, not the complete retrains that sound like when the
- : modems are initially establishing a connection. These take much longer
- : and will sometimes cause timeouts. If you set S192.5=1 on the Supra, it
- : will display an L followed by an up or down arrow when the local end
- : requests a fall-forward or backward, and display R and an up and down
- : arrow when the remote end requests a retrain.
-
- Default settings are %E0 and %G1 and, according to the manual, the %E0
- disables the request for retrain while %G1 "On V.32bis connections, as
- line conditions change, your modem automatically initiates a Rate
- Renegotiation."
-
- Please note the references to "V.32bis". It does not automatically apply
- to V.34 connections and that is where the problem appears. Setting the
- %E1 option, which does not define the V.34 of V.32bis, is said to allow
- the Supra to "monitor line conditions and automatically request a retrain
- if line conditions are bad" (that is the original definition, it is
- slightly different in the newer code but the basic function is the same).
-
- In the connections where this problem occurs, the USR will show a Request
- for Retrain received in the ati6 (call analysis) screen. This means that
- the retrain request came from the Supra. Because of the point at which
- the modem connection was dropped (manually), the request is known to come
- during EC/DC negotiation.
-
- : : Maybe I've just been spoiled by my USR Couriers but I expect a modem to
- : : function properly and not have a factory default which masks a problem.
- : I've found the factory defaults on the USR's (both Couriers and
- : Sportsters) have trouble with a lot of modems, where the Supras here
- : haven't had problems with any.
-
- The USR's have had a terrible time with Rockwell chipset modems. It is
- always Rockwell chipset modems and no other. I have no doubt that you
- accept the factory defaults which limit the Retrain Request in the Supra
- and mask the problem.
-
- This problem reminds me of the Hayes problems some time ago where they
- recommended restricting the Retrain Request. Apparently, that has been
- the "fix" for Rockwell chipsets. USR units do not request retrains at
- improper points in the connection (during setup of EC/DC) and do not need
- this work-around.
-
- I have run (and still run) a BBS using USR's and beta tested the V.FC and
- V.34 and V.34+ codes. The default for the USR (&F1) with only one
- modification (s2=255) worked perfectly. I became aware of the problem I
- reported because I switched from a ZyXel (1496+) to a Supra 288 on one
- node (I needed the CID capability).
-
- I had one caller who repeatedly connected without error correction until
- I switched that node to another USR. His modem is a GVC, if I recall
- correctly. I also found that I ran into the same problem when calling in
- from my office using my USR.
-
- I have switched to the %E0 and now find connections are stable but that I
- regularly get transfer rates for a 26400 rate (receive) even though the
- modem shows a 28800 connect rate (receive) on the display. Does this mean
- that the Supra is falsely reporting a R34/28800 or that it is having
- trouble at that rate and should request a lower speed but doesn't?
-
- FWIW, the Supra routinely gives me lower speed connections to most places.
- For example, a call to my ISP is regularly a 21600/24000 using the Supra
- and never improves but is always a 24000/26400 with fairly regular
- upshifts to 26400/26400 using the USR. This is over the same phone line
- on the same computer using the same software.
-
- Does anyone know for sure what options are designed for V.34 retrains in
- the Supra? ^^^^
-
-
-
-
- --
- "Things are more like they are now than they ever were before."
- [Dwight D. Eisenhower]
-