Ian Smith wrote in a message to Dave Hatch:

DA> Saturday January 20 1996 17:59, Dave Hatch wrote to Deon Attard:

DA>>> Netcomm m34f, it'll normally connect at 26400 if the lines arn't
DA>>> real good and then train its way up.

DH>> Sounds about right.   The Rockwell set has the capability of
DH>> specifying something I'd term "aggressiveness" in line
speed selection
DH>> - good idea, and works a beauty.

DA> Any idea what the command/s register is to enable the
"aggression" :-)
DA> Sounds like an experiment to me..

 IS> I've not seen anything in any Rockwell modem's manual to indicate
 IS> they have this ability - though as Dave says, it must be there,
 IS> deep down anyway. 

 DH> On the Netcomm, the hardware's there.  They do their own
 DH> software, so it probably will take a bit of poking about to
 DH> find what they call it, or if they've implemented the
 DH> manual adjust feature.

 DH> On my Hayes v34, it's in S108, labeled Signal Quality
 DH> Select, range of 0 - 3,
 DH> with 0 putting up with anything, and 3 demanding perfection
 DH> on pain of instant train down to the next lower speed.
 DH> That's what the manual sez.  The practice is a bit gentler
 DH> - under some circumstances you can't see the effect at all.
 DH>  You need a noisy line, under some pretty precise ranges of
 DH> "noisy" before it is an effective tool, IMHO.

 IS> Yes, I don't know why all modems don't allow the 'aggressiveness'
 IS> of trying to stay at a certain speed, rather than fallback more
 IS> readily, adjustable.  Particularly when it can make the difference
 IS> between a stable connect, or not, or save sessions from being
 IS> littered with retrains and/or rate shifts, which can do weird
 IS> things to file transfer protocols sometimes.

this is possible with the microcom deskporte fast and the fast + (i suppose
its the same with the ep and es) thier is a cluster of registers that
allows you to manipulate the timing and sequencing of the "protocol
stack" to achieve a desired profile for a particular line problem
 the only thing that i haven't yet discovered tis how to isable
the v.34 and vfc modulations sperwtely and indipendantly. thier are a
vouple of modems that have no problems with vfc but as soon as i introduce
the v.34 all hell breaks loose. i've spoken to microcom (admitidly here in
sydney) at lenght about his and they cant think of anything either

 IS> We discovered, after digging through the rather cryptic
 IS> descriptions in the manual, that playing with this on the Motorola
 IS> Lifestyle made the difference between getting an error-corrected
 IS> connect or not, albeit at an about 2400bps lower speed, with a rack
 IS> of Netcomm V.34 modems at the local uni.  A solid 24000 with LAPM
 IS> is much better than 28800, without.

agreed .. though most of the people out thier look at the connect rate and
see something other (lower) than 28k8 and complain (bitterly) about being
ripped off.

 IS> I've been playing with this feature for years with my DPX-596; it's
 IS> often made all the difference between complete failure and at least
 IS> limited success on connections to some modems on particularly noisy
 IS> old exchange equipment. 

yes, the 596 .. i don't think its available on the australianised version,
the much cheaper one (496, from memory)

 DH> Still - try.  Nothing to loose.  Snoop in the Netcomm
 DH> online manual, and see if any of the descriptions match
 DH> this function, and play, if they do.

 IS> Any reports on how to do this with the Netcomms (or any other
 IS> Rockwell-chipped modems) will certainly be snipped for future
 IS> reference here .. 

i don't know how rockwell standard the microcom is by the time microcom is
finished byt i'll hunt up the register information and post it to you.

regards ... jonathan

EMail: jmichaels{at}caamora.com.au

---