Which External Voice-Modem Should I Buy?
Bennu Oing
56K V.90 voice/fax/modem. I want one that is reliably fast and has good
(broad) voice and fax support. The Tiger Direct catalog lists the CNet
and Newcom for $59.99 (US) and $89.99 respectively. I'm guessing the
"CNet" company has nothing to do with the Commodore-related BBS software,
nor the TV show. I'll be using the PCMCIA Surf-Squirrel SCSI/fast-serial
hyrbid interface.
While both are listed as V90 modems, they also support the two previous
competing semi-standards: CNet's is X2 and Newcom's is K56flex. If V.90
is the agreed standard, why do they even mention these other (retired)
protocols? Should this sway me? I'v e heard that V.90 is mostly
Kflex-derived, and that V.90 has relatively little relationship to X2.
Does this mean that a KFlex/V90 modem would (or may) have more reliable
connections, or attain higher speeds than an X2/V90 modem? Even a
reference to ben chmark speed/reliability tests in some unbiased
publication would help.
Voice support is VERY important to me. My pal got burned with his
voicemodem. It used some 3rd-string chipset so there was almost no
support for it's voice features. What are the top modem chipset standards
anyway (beside Rockwell and Cirrus)? I know Rockwell and Cirrus are two
major ones, and think there at least one or two other competitors. I'd
like to be able to use STFax's voice features, as well as try the
voice-mail programs on AmiNet. Did any voice standard for AT-like
command-sets ever evolve? What are the issues in voice handling in
modems anyway? I reviewd an old (1990?) Zyxel manual and it seemed very
straightforward to write programs for.
Are there any models or brands that are known to be problematic? I know
winmodems suck and i can't use them anyway. Are there any *external* V90
voice/fax/modems that are really good deals? Experienced buyers only,
please.
BTW did Cirrus-based modems only use X2 while Rockwell-based modems used
only 56KFlex? Or were the chipsets able to go either way? What about the
other chipsets?
My current fax program is GPFax 2.343 (Class2 5.4.94). The CNet modem
has "fax compatibility for ITU-T, V17, V29, V90, v27ter, and V.21 ch2,
Group 3 fax suport, Class 1 and more." Will it work with my version of
GPFax? The Newcom modem doesn't have much listed for it, but it does
have vague marketroid hype like "Digital Answering Machine" (gee, is this
self-contained or does one need special software?).
Please don't insult me by assuming I'm running an x86 box. Ignore this request after April 22nd 1999.
Thanks!
Bennu
PS Technical treatise should be cc:'d to bo...@waste.org
Countdown: computing changes forever, December 1999, and it has nothing to
do with Y2K! It's baaaack, and boy is it ticked off!
John Navas
Bennu Oing <bo...@waste.org> wrote:
> I looking for qualified advice on how to choose a really good external
>56K V.90 voice/fax/modem. ...
Diamond Supra.
--
Best regards,
John mailto:Use...@NavasGrp.Dublin.CA.US http://www.aimnet.com/~jnavas/
28800-56K Modem FAQ: http://www.aimnet.com/~jnavas/modem/faq.html
Marcel DeVoe
> I looking for qualified advice on how to choose a really good external
> While both are listed as V90 modems, they also support the two previous
> competing semi-standards: CNet's is X2 and Newcom's is K56flex. If V.90
> is the agreed standard, why do they even mention these other (retired)
> protocols? Should this sway me? I'v e heard that V.90 is mostly
> Kflex-derived, and that V.90 has relatively little relationship to X2.
> Does this mean that a KFlex/V90 modem would (or may) have more reliable
> connections, or attain higher speeds than an X2/V90 modem? Even a
> reference to ben chmark speed/reliability tests in some unbiased
> publication would help.
> Voice support is VERY important to me. My pal got burned with his
> voicemodem. It used some 3rd-string chipset so there was almost no
> support for it's voice features. What are the top modem chipset standards
> anyway (beside Rockwell and Cirrus)? I know Rockwell and Cirrus are two
> major ones, and think there at least one or two other competitors. I'd
> like to be able to use STFax's voice features, as well as try the
> voice-mail programs on AmiNet. Did any voice standard for AT-like
> command-sets ever evolve? What are the issues in voice handling in
> modems anyway? I reviewd an old (1990?) Zyxel manual and it seemed very
> straightforward to write programs for.
Huh? I thought Cirrus was a CPU used in a computer. ;-)
Actually it's Rockwell and Lucent. Rockwell started with X2 and Lucent
with 56Flex.
> Are there any models or brands that are known to be problematic? I know
> winmodems suck and i can't use them anyway.
Well, whether Winmodems suck is not the issue. They can have the same chip
sets in them and work just as well but just don't have the software
programmed in ROM and can't think without Windows software support, so
they don't even work with PC's trying to run MSDOS applications.
> BTW did Cirrus-based modems only use X2 while Rockwell-based modems used
> only 56KFlex?
No, the other way around. ;-)
> Or were the chipsets able to go either way? What about the
> other chipsets?
V90 was actually a compromised standard and somewhat less than either for
compatabilities sake.
In either case, your best connection to your provider, if you intend to
stay with him, is to get a modem that has the same chipset as what he is
using. You'll get the best handshaking, if you will, that way.
Once you find out that, then you're only face with half the problem of
which is best.
As Kirk pointed out, the Supra is the best with the Lucent chip, and hence
the best 56kFlex one.
I can't comment on the X2 though. I know some like the USR "do everything"
vmodems (and that definately excludes the Sportster ones).
> Please don't insult me by assuming I'm running an x86 box. Ignore
> this request after April 22nd 1999.
You're a strange person.
> PS Technical treatise should be cc:'d to bo...@waste.org
Well, we expect you to come back and read follow-up postings to your OWN
post to this group.
We frown on "posters and runners".
--
Marcel J. DeVoe - mde...@shore.net - Team *AMIGA*
A4000/060 CyberStorm MKII overclocked 66mhz - see "How to Overclock!"
and "DIY A4000 Tower for $45" at http://www.shore.net/~mdevoe
A4091scsi CV64 64 megs CDROM M1764-17" Catweasel FUSION/Emplant
Rick Collins
wrote:
>Huh? I thought Cirrus was a CPU used in a computer. ;-)
Cirrus Logic also makes modem chipsets.
>Actually it's Rockwell and Lucent. Rockwell started with X2 and Lucent
>with 56Flex.
Rockwell and Lucent both developed K56Flex; USR developed X2.
>V90 was actually a compromised standard and somewhat less than either for
>compatabilities sake.
Nonsense. While V.90 is the result of the ITU's "committee" process
there's no "proof" that it's "somewhat less than either".
-----------------------------------
Rick
Marcel DeVoe
> On Mon, 22 Mar 1999 07:24:22 GMT, Marcel DeVoe <mde...@shore.net> wrote:
>>Huh? I thought Cirrus was a CPU used in a computer. ;-)
> Cirrus Logic also makes modem chipsets.
Hmm, never heard of a Cirrus chipset. Can you name some brands that use
it?
>>Actually it's Rockwell and Lucent. Rockwell started with X2 and Lucent
>>with 56Flex.
> Rockwell and Lucent both developed K56Flex; USR developed X2.
Yes, you're right. Lucent and Rockwell developed 56kflex in cooperation.
And I WORK for LUCENT! :-}
I believe USR uses Motorola chips for it's modems.
>>V90 was actually a compromised standard and somewhat less than either for
>>compatabilities sake.
> Nonsense. While V.90 is the result of the ITU's "committee" process
> there's no "proof" that it's "somewhat less than either".
Nevertheless, you're best to try to "match" you ISP's chipset, otherwise
you'll get less than maximum performance. I've seen it stated several
times.
To quote the April issue of PC World on page 30;
"Ask your ISP what chip set it uses in the modems at it's facilities.
Consider sticking with x2 or K56flex if your ISP supports it. Buying a
desktop modem that uses the same chip set will reduce the likelihood of
incompatibilities and may boost performance."
I believe that V90 was mean't to be a "bridge" between the two
technologies. But you're best staying on your own side of the bridge if
you can.
Neil Bothwick
> On Mon, 22 Mar 1999 07:24:22 GMT, Marcel DeVoe <mde...@shore.net>
> wrote:
>>V90 was actually a compromised standard and somewhat less than either for
>>compatabilities sake.
> Nonsense. While V.90 is the result of the ITU's "committee" process
> there's no "proof" that it's "somewhat less than either".
In fact, feedback from users and server logs shows higher throughputs
and more stable connections with V90 than either K56flex or X2.
Neil
--
Neil Bothwick - http://www.wirenet.co.uk icq://16361788
Connected via Wirenet,The UK's first Amiga-only internet access provider
--
Clap on <clap clap> Clap off <clap clap> NO CARRIE���¼
Rick Brice
Check this...
http://www.cirrus.com/news/corp98/news-corp28.html
then this....
http://www.basiscomm.com/
According to Cirrus....Modems are *Legacy* products...
Regards,
Rick
Marcel DeVoe wrote:
>
> In comp.sys.amiga.datacomm Rick Collins <ab...@issc.debbs.ndhq.dnd.ca> wrote:
> > On Mon, 22 Mar 1999 07:24:22 GMT, Marcel DeVoe <mde...@shore.net> wrote:
>
> >>Huh? I thought Cirrus was a CPU used in a computer. ;-)
>
> > Cirrus Logic also makes modem chipsets.
>
> Hmm, never heard of a Cirrus chipset. Can you name some brands that use
> it?
<Big Snip>
Rick Collins
wrote:
>In comp.sys.amiga.datacomm Rick Collins <ab...@issc.debbs.ndhq.dnd.ca> wrote:
>> On Mon, 22 Mar 1999 07:24:22 GMT, Marcel DeVoe <mde...@shore.net> wrote:
>
>>>Huh? I thought Cirrus was a CPU used in a computer. ;-)
>
>> Cirrus Logic also makes modem chipsets.
>
>Hmm, never heard of a Cirrus chipset. Can you name some brands that use
>it?
Not off hand, though there is likely someone here who could provide
that information. Not that it matters overly much.
>>>Actually it's Rockwell and Lucent. Rockwell started with X2 and Lucent
>>>with 56Flex.
>
>> Rockwell and Lucent both developed K56Flex; USR developed X2.
>
>Yes, you're right. Lucent and Rockwell developed 56kflex in cooperation.
>And I WORK for LUCENT! :-}
>
>I believe USR uses Motorola chips for it's modems.
3Com/USR uses TI products in their modems.
>>>V90 was actually a compromised standard and somewhat less than either for
>>>compatabilities sake.
>
>> Nonsense. While V.90 is the result of the ITU's "committee" process
>> there's no "proof" that it's "somewhat less than either".
>
>Nevertheless, you're best to try to "match" you ISP's chipset, otherwise
>you'll get less than maximum performance. I've seen it stated several
>times.
That's bafflegab. What matters is how well V.90 was implemented in
silicon on both ends of the link. A lousy implementation will likely
work OK with the same lousy implementation on the other end, but the
goal is, of course, proper implementations and interoperability.
Since the ISP is concerned with maintaining customers, you can expect
that any decent ISP will strive to have a fully-compliant
implementation. So, buying for quasi-compatibility _now_ may leave
you high and dry when the ISP updates - especially if your update
route is less than assured. A well-defined update route and a
manufacturer with a proven track record are far more important that
what product the ISP is using on his end.
My X2/V.90 modem connects with my ISP's K56Flex/V.90 modem usually at
49333 but often at 50666, and I've had connects of up to an hour with
no BLERS, Link NAKS, retrains, or other maladies.
>To quote the April issue of PC World on page 30;
>
>"Ask your ISP what chip set it uses in the modems at it's facilities.
>Consider sticking with x2 or K56flex if your ISP supports it. Buying a
>desktop modem that uses the same chip set will reduce the likelihood of
>incompatibilities and may boost performance."
Which is why I don't consider such sources particularly useful as a
source of recommendations. With flash updates, the ability to update
through a known, assured path is a critical consideration. Ask anyone
with a K56Flex modem by some unknown manufacturer who cannot find an
upgrade. Any update at your "same chipset as me" ISP to correct V.90
implementation errors can leave the user high and dry if he cannot
find an appropriate upgrade for his modem. That applies, too, of
course in cases where the ISP is _not_ using the same chipset - which
is why the assured update route is so important.
>I believe that V90 was mean't to be a "bridge" between the two
>technologies. But you're best staying on your own side of the bridge if
>you can.
Reality is K56Flex and X2 were proprietary, interim protocols. You
will not see another upgrade to either one - they are orphans. There
are already some ISPs who have dropped "K56Flex" and are sole V.90
sites. My bet is this trend will continue.
-----------------------------------
Rick
Jeff Grimmett
> >> Cirrus Logic also makes modem chipsets.
> >Hmm, never heard of a Cirrus chipset. Can you name some brands that use
> >it?
> Not off hand, though there is likely someone here who could provide
> that information. Not that it matters overly much.
My wife's PC has an internal USR with a Cirrus chipset, made in 1997. I've
seen others, usually no-names, as well.
> My X2/V.90 modem connects with my ISP's K56Flex/V.90 modem usually at
> 49333 but often at 50666, and I've had connects of up to an hour with
> no BLERS, Link NAKS, retrains, or other maladies.
What is UP with that, anyway? I've become accustomed to modems always using
multiples of 300 for the connect rate -- first time my Zoom connected at
37666, I almost called the Pope.
--
// Jeff Grimmett . ICQ 17300370 . http://www.ald.net/dlg
// --------------------------------------------------------
// Digerati Dreams : DLG Pro BBOS : Skimmer : Nag Plus
// : Addiction FTP : ASTG :
// --------------------------------------------------------
Greg Tallent
>In article <36f7c5d6...@news.ndhq.dnd.ca> ab...@issc.debbs.ndhq.dnd.ca
>(Rick Collins) writes:
>> My X2/V.90 modem connects with my ISP's K56Flex/V.90 modem usually at
>> 49333 but often at 50666, and I've had connects of up to an hour with
>> no BLERS, Link NAKS, retrains, or other maladies.
>What is UP with that, anyway? I've become accustomed to modems always using
>multiples of 300 for the connect rate -- first time my Zoom connected at
>37666, I almost called the Pope.
My Zoom usually connects at 50666. V.90 is quite a bit more stable in it's
connections than V.34(+) ever was and it doesn't downgrade it's connection like
K56 does either.
<tsb>
Greg Tallent |Amiga2000 GVP 040/33mhz/3.1 72 megs,4 gig/OpalVision|
gwt @ gte.net |Zoom 56k, Syquest Syjet, SCSI-Zip, Picasso II 2Meg|
Everyone is weird. Some of us are proud of it.
Bennu Oing
On Mon, 22 Mar 1999, Marcel DeVoe wrote:
> Huh? I thought Cirrus was a CPU used in a computer. ;-)
That's Cyrix.
Somebody else already replied to you, but I want to, too.
> Actually it's Rockwell and Lucent. Rockwell started with X2 and Lucent
> with 56Flex.
Uh, I'm pretty sure Rockwell was 56Kflex.
> Well, whether Winmodems suck is not the issue. They can have the same chip
> sets in them
Oh do they now? Is that why they need a high-speed host processor?
>and work just as well but just don't have the software
> programmed in ROM and can't think without Windows software support, so
WinModems AFAIK lack full DSP's. They rely on some host processor to do
some of the work. That's why crapboxes like the eTower 333MHz can't
sustain a 53kbps connection if you have more than one net program open (or
multitask).
> > BTW did Cirrus-based modems only use X2 while Rockwell-based modems used
> > only 56KFlex?
>
> No, the other way around. ;-)
I'm darn sure X2 was not used by Rockwell.
> > Please don't insult me by assuming I'm running an x86 box. Ignore
> > this request after April 22nd 1999.
>
> You're a strange person.
I was crossposting to a group frequented by Wintel heathens.
> Well, we expect you to come back and read follow-up postings to your OWN
> post to this group.
>
> We frown on "posters and runners".
My ng netfeed is a bit iffy. As is my storage situation.
Marcel DeVoe
> On Mon, 22 Mar 1999, Marcel DeVoe wrote:
>> Huh? I thought Cirrus was a CPU used in a computer. ;-)
> That's Cyrix.
> Somebody else already replied to you, but I want to, too.
Well, not just one. ;-)
>> Actually it's Rockwell and Lucent. Rockwell started with X2 and Lucent
>> with 56Flex.
> Uh, I'm pretty sure Rockwell was 56Kflex.
Evidently you didn't read the followups, including mine.
Bad form on your part. ALWAYS read followups before replying.
>> Well, whether Winmodems suck is not the issue. They can have the same chip
>> sets in them
> Oh do they now? Is that why they need a high-speed host processor?
Hmm? Some 166mhz users are using Winmodems.
>>and work just as well but just don't have the software
>> programmed in ROM and can't think without Windows software support, so
> WinModems AFAIK lack full DSP's. They rely on some host processor to do
> some of the work. That's why crapboxes like the eTower 333MHz can't
> sustain a 53kbps connection if you have more than one net program open (or
> multitask).
Thanks for the info. But even on my Amiga if I run IBrowse twice and
and click both on a link at the same time they are each both going to be
reduced to downloading web pages to half speed.
>> > BTW did Cirrus-based modems only use X2 while Rockwell-based modems used
>> > only 56KFlex?
>>
>> No, the other way around. ;-)
> I'm darn sure X2 was not used by Rockwell.
Read followups.
>
>> > Please don't insult me by assuming I'm running an x86 box. Ignore
>> > this request after April 22nd 1999.
>>
>> You're a strange person.
> I was crossposting to a group frequented by Wintel heathens.
You certainly got that right. ;)
>> Well, we expect you to come back and read follow-up postings to your OWN
>> post to this group.
>>
>> We frown on "posters and runners".
> My ng netfeed is a bit iffy. As is my storage situation.
Get a new ISP? ;-)
Well, you must have got this one quite sometime later, so you should see
the followups also.
And drives are cheap these days.
John Navas
Bennu Oing <bo...@waste.org> wrote:
>WinModems AFAIK lack full DSP's. They rely on some host processor to do
>some of the work.
No, "WinModems" (a 3Com trademark) have a (data pump) DSP -- they only rely
on the host CPU for control functions. Host-based ("software") modems lack
a DSP. See my FAQ.
>That's why crapboxes like the eTower 333MHz can't
>sustain a 53kbps connection if you have more than one net program open (or
>multitask).
I'm afraid that's just not true.