jeff@markets.amix.com (Jeff Crilly N6ZFX) (05/30/91)
I've been having some problems with a 9600bps V.32 modem (lets call it
brand-x, I don't want to disclose the mfr) when using it to connect to
2400 bps modems (NEC and Avatex).
When I use brand-x to connect to these modems I always get a connection
at 1200 bps. I also tried turning off the auto-fallback feature in
brand-x and the modem would hang up as soon as it hears answer tone.
I have several other modems (Hayes personal modem and practical
peripherals, to name two) that can connect fine to the NEC and avatex at
2400 bps. Also, brand-x can connect to these modems at 2400 bps.
It turns out that the default answer tone for the NEC and Avatex is
Bell 212A. I think the default for the hayes personal and practical
peripherals is CCITT V.22. But I don't have the manuals so I can't
verify it.
So I changed the answer tone (also called protocol in the manual) on the
NEC and Avatex to V.22 (I think this is really V.22 + V.25 answer tone)
and now brand-x can connect fine. And so can the other modems.
Incidently, on the NEC I can select Bell212, CCITT V.22, or CCITT V.22+
V.25. Only, CCITT V.22+V.25 works.
The question is: what is correct? Should the modems be set to answer
with Bell 212A answer tone or CCITT V.22+V.25. Here my understanding of
the standards, correct me if I'm wrong:
1) Bell 212A -standard for 1200 bps modems. Also includes a standard
for answering the line (answering modem generates a
2225Hz tone for x ms.)
2) V.22 -standard for 1200bps modems. Includes answering with
a some tone (which I'm not sure of.
3) V.25 -answer tone of 2100Hz which can be used with V.22 (presumably
to cancel echo supression equipment in the GSTN).
4) V.22bis -standard for 2400bps modems that is based on V.22.
(A V.22 connection is made first, then 2400bps is attempted.)
Is it correct that in the U.S., modems should use the Bell212 standard for
answering the line? The modem should not determine the data rate from the
answer tone, correct?
Also, is the use of 212 answer tone for V.22bis connections incorrect (i.e.
non-standard)? It seems to work with some modems, but at least one modems
(brand-x) and possibly others can't handle this.
Thanks for any info.
Jeff Crilly (N6ZFX)
AMIX Corporation 2345 Yale Street Palo Alto, CA 94306
jeff@markets.amix.com, {uunet,sun}!markets!jeff, N6ZFX@N6IIU.#NOCAL.CA.USAtnixon@hayes.uucp (05/31/91)
In article <1991May29.223758.5121@markets.amix.com>, jeff@markets.amix.com (Jeff Crilly N6ZFX) writes: > I've been having some problems with a 9600bps V.32 modem (lets call it > brand-x, I don't want to disclose the mfr) when using it to connect to > 2400 bps modems (NEC and Avatex). I'm going to assume that this V.32 modem uses the Automode procedures defined by CCITT Study Group XVII for interworking with lower speed modems and with combined V.32/lower-speed modems. The Automode procedure for the calling modem is as follows (much condensed): If the modem hears V.25 answer tone, it sends the V.32 AA signal to let the other modem know that the calling modem has V.32 capability. If the answering modem is a V.32 modem, it will hear this tone and switch to V.32 operation (if the answering modem doesn't send V.25 answer tone, this phase is skipped). If the calling modem hears the Bell 212/V.22/V.22bis "unscrambled binary 1" signal (which is basically the same as 2225Hz), it waits to hear how long it persists. If it lasts longer than 3 seconds, the modem assumes that the answering modem doesn't have V.32 capability, so it sends the appropriate handshake signal to connect with the lower speed. If the calling modem supports V.22bis, it is SUPPOSED to prefix this signal with an "S1" signal, which the answering modem will see and switch to 2400bps. > When I use brand-x to connect to these modems I always get a connection > at 1200 bps. I also tried turning off the auto-fallback feature in > brand-x and the modem would hang up as soon as it hears answer tone. > > I have several other modems (Hayes personal modem and practical > peripherals, to name two) that can connect fine to the NEC and avatex at > 2400 bps. Also, brand-x can connect to these modems at 2400 bps. > > It turns out that the default answer tone for the NEC and Avatex is > Bell 212A. I think the default for the hayes personal and practical > peripherals is CCITT V.22. But I don't have the manuals so I can't > verify it. It's possible that your modem is saying that because the V.25 answer tone was not sent, that the answering modem MUST not support V.22bis, because V.22bis modems (supposedly) _always_ send the V.25 answer tone. There are quite a number of V.22bis modems on market that WILL NOT CONNECT IN V.22bis if they don't hear answer tone first. If they support Bell 212, they will fall back to that speed and never send the "S1" sequence indicating that they support V.22bis. I don't believe that the Automode procedure allows this behavior, but some manufacturers may have carried it forward from older products. > So I changed the answer tone (also called protocol in the manual) on the > NEC and Avatex to V.22 (I think this is really V.22 + V.25 answer tone) > and now brand-x can connect fine. And so can the other modems. > Incidently, on the NEC I can select Bell212, CCITT V.22, or CCITT V.22+ > V.25. Only, CCITT V.22+V.25 works. > > The question is: what is correct? Should the modems be set to answer > with Bell 212A answer tone or CCITT V.22+V.25. It doesn't hurt for a Bell 212A calling modem to hear V.25 answer tone, so I would always set my modem to send it. > Is it correct that in the U.S., modems should use the Bell212 standard for > answering the line? The modem should not determine the data rate from the > answer tone, correct? There's nothing wrong with using V.22 in the USA. It just happens that most 1200bps modems in the USA are Bell 212. In fact, the most-used mode of V.22 is Bell 212A-compatible, and most of the time they connect with each other and operate just fine! As far as limiting the data rate if you don't hear the V.25 answer tone, I don't think its a good idea, but I know that some manufacturers have chosen to do it -- but not Hayes. > Also, is the use of 212 answer tone for V.22bis connections incorrect (i.e. > non-standard)? It seems to work with some modems, but at least one modems > (brand-x) and possibly others can't handle this. Yes, it is non-standard. V.22bis doesn't say that you can send a 2225Hz answer tone. It says to send the 2100Hz V.25 answer tone. -- Toby Nixon, Principal Engineer | Voice +1-404-840-9200 Telex 151243420 Hayes Microcomputer Products Inc. | Fax +1-404-447-0178 CIS 70271,404 P.O. Box 105203 | UUCP uunet!hayes!tnixon AT&T !tnixon Atlanta, Georgia 30348 USA | Internet hayes!tnixon@uunet.uu.net