There was an interesting article in Boardwatch magazine a few months ago
comparing the connect speeds of KFlex and X2 modems.  Overall, the X2 modems
had a significantly higher connect speed.  The article created a flurry of
controversy, but basically the magazine made something like 150,000 calls to
ISPs using both systems.  They used identical equipment for both "standards"
and saw average connect rates for USRs of about 44K and for KFlex of around
37K.  Now, I'm just pulling this out of memory, but I *think* those are the
right numbers.  I know that the USR numbers were quite a bit higher.
Incidentally, it did not seem to matter if the ISPs they called were local
or long distance...they called a LOT of them and got pretty much the same
results.

Also, I believe that they pointed out that there seem to be more X2 ISPs
than KFlex, although that's probably a moot point since the average person
is only interested in what "standard" their ISP adheres to.  Supposedly the
V90 standard will eliminate this issue, but I suspect that things won't be
that easy.  When I worked for a local ISP (RMCI) here in Boise, Idaho, we
upgraded the USR racks to v.90 and got plenty of calls from KFlex users who
said that their v.90 modems would not connect faster than 28.8.  Since I
wasn't involved in any testing or troubleshooting, I can't really speculate
on wha the problem might have been...only that there was one.

Anecdotally, my Sportster X2 modem always connected at 53K and never
renegotiated below 50K while I was using it, except for a two day period
when the #@%! phone company put a splitter on my lines.

Something that Boardwatch pointed out was that there are a few different
companies making KFlex host systems, the digital modems on the ISPs end.  I
didn't know that...I thought that Livingston was the only one.  Apparently
some KFlex modems connect better to certain hosts than others.  It's the old
compatability issue that used to plague us with v.34 and v.fast when it was
new.

On the other hand, the article got an enormous amount of attention from
Rockwell and Lucent, so it may give them incentive to work on refining their
chipsets.

I suspect that the article is on their web site...it's generated so much
mail to the magazine that I'm sure that they must have posted it.  It's very
good reading.

My 2 cents!

Drew Dunn
[log in to unmask]

> There are other factors to also consider before upgrading to a 56K
> modem.  Some of which may account for the problems you experience.
> >
> > I have a friend with a software upgradable zoom modem. (33.6 to
> > 56k)he wants
> > to pay the $60 to upgrade but I told him I bought a 56k modem and
> > my connect
> > speeds almost never get past 33.6.
>