cec@cup.portal.com (Cerafin E Castillo) (11/30/90)
OK, the inside story...
This was my personal experience in regards to the PC Magazine
review of TELEBIT's T2500 modem. I have expressed my own personal
opinions herein. I am not presenting this as an official TELEBIT
representative, but only as a TELEBIT modem and UNIX user.
I assisted in the Telebit side of aid to these PC Magazine technicians
conducting the review. I should make it clear that PCMag was very
thoughtful and thorough in contacting Telebit for assistance in
conducting a fair review. Those of you who saw the MacUser High-Speed
Modems Review, this year, know of the bad review Telebit modems
received, as well. Telebit was never contacted for this review and
the evidence was the satellite transmission results for V.32 modems,
PEP was never tested. When PCMag contacted us at Telebit, we where
very surprised and happy to get a chance at a 'good' review.
Now, the reality. In so far as those of us techies involved in
assisting the reviewers, they were conducting legitimate tests
and doing so properly. Telebit Systems Engineers and a couple of
Tech Support Reps. were involved in assisting with the review
over the phone. No on-site assistance was ever given to PCMag.
This may have been Telebit's biggest mistake. The technical people,
along with Marketing Communications, at Telebit, were all disappointed
with the results, as well.
Here's what leads to this disappointment. The V.32 tests, and only the
V.32 tests were quite accurate and administered properly, as per Telebits
review of PCMag test procedures. Telebit has just announced the T1600
modem which addresses some of the missing criteria used in the PCMag
review. The T1600 has a 38.4 kbps serial interface which Telebit is
promoting as "TRUE" 38.4 kbps. I don't quite know what they mean by
this, yet. The T1600 is using Telebit's own DSP for V.32 instead of
the Rockwell Chipset used by the T2500 and T1500. The Rockwell chipset
proved to be a quality assurance nightmare, while I was at Telebit.
It's default settings often caused unstable connections. While Telebit
was able to compensate for these problems in their firmware (GE6.01)
the inability of the Rockwell chipset to do V.32bis (14.4 kbps), in
its present form, caused Telebit to develope the T1600 DSP. Cost was
obviously an issue as well. The T1600 is listed at $795 US. The
T1600 can not do PEP, at the moment, and may not ever.
Speaking of PEP...
PEP's failure in the impaired line tests was due to the lack of use, by
PCMag, of the hidden registers in the modem firmware (S120/J6S36). These
registers affect packetization for better survival of connection and
throughput over adverse line conditions ONLY. It was not in Telebits
best interest to start talking about or publishing this kind of information.
Only recently has Telebit Tech Support introduced a document detailing
the use of these registers. Thus, in default PEP packetizing mode, the
failure of PEP was legitimate.
PEP also manifested the drawback many of us Telebit modem users have
encountered, the "jerky" response of PEP's half-duplex (aka Adaptive
Duplex) modulation. This is the half-screen redraw or delay in
character echo experienced over interactive sessions. I disagree with
the gentleman who stated that PEP was only an issue in "simple typing"
sessions and not with dial-up IP connections. I have used SLIP/CSLIP/PPP
extensively, and not even the Van Jacobson Header Compression Algorithms
have made the use of PEP any more efficient. V.32 is still preferred.
It's no wonder that the TELEBIT NetBlazer is meant to be used with the
T2500/T1500/T1600 V.32 modems exclusively, and not necessarily with PEP.
All my testing and that of the NetBlazer project concluded that this was
the case, even in the light of IP Inverse Multiplexing used by the
NetBlazer.
RTFM, in particular TFM, is also a failure point. I never read through
all of the over 200 pages in the TELEBIT Commands Reference Manual. NFW!
As the UNIX Communications Specialist, my charge was to write system
specific configuration guides for UNIX systems and applications. This was
because of the fact that the ability to configure your Telebit modem
hinged on your ability to deal with your O/S or data comm software. Telebit
does not have a "Plug & Play" modem, yet. I don't believe that any
modem is ever fully understood by its user and installed with such
understanding of both modem and system. This is part of the market that
Telebit and most modem manufacturers try to address for greater sales and
market share. This is also the readers that PCMag is trying to educate
or, at least, bring up to speed. "Plug & Play" means money. The T1600,
in its planning stages, was meant to have "user macros" which would
auto-configure the modem for the users specific system or application.
"AT&0-9" or use of the modems front panel lights would supposedly make
modem set-up fool-proof and fast, thus "Plug & Play". Unfortunately,
my experience with this idea, at least for UNIX systems, was that Tech
Support came up with two "optimum" UNIX initialization lines to be used
as macros. No one in Telebit Tech Support is a Sys Admin level UNIX user.
Thus, the macro proposal I saw was nowhere near accurate. I tried to
convince the developers to use a VERY generic, semi-"fool-proof" init
macro for UNIX systems and SLIP/CSLIP/PPP applications. The objective
should be to get the modem to accomodate the O/S functions of the system,
as opposed to doing "optimum" performance, which is different from
system to system. Just remember, of the over 75 Telebit modem registers,
only a dozen or so are necessary for basic RS-232 and PSTN configuration.
And Tech Support...
Great bunch of guys! Highly professional and knowledgeable, except for
UNIX! This is my biggest beef. There is no one at the Tech Support level
that I can qualify as being UNIX experienced enough to handle the complex
O/S issues involved with UUCP or dial-up IP configuration. There is
basic knowledge and documentation, but any "stickies" might be met with
a shoulder shrug. E-mail support, from Telebit, is all but abandoned.
Tech Support handles this, now. If they can't get an automatic return
path from their mail system, you will not get an answer to your e-mail.
Forget postings or responses to netnews, this part of the net support
is not even being monitored. I am biased here because of the fact that
I was laid-off in September '90 by Telebit due to their lack of "making
their numbers" in that quarter and bad forecasts for the remainder of the
year. I was the guy in charge of e-mail and netnews support. Luckily,
my new employer, a TELEBIT distributor, allows me to continue my netnews
activity in support of TELEBIT modem and NetBlazer users. This is
fortunate for me because I consider my self part of this community, as
opposed to just an industry rep..
I hope that this info is of help to understanding the issue at hand.
===============================================================================
Cerafin E. Castillo || //\\ ||\\ ||
Network Consultant || //__\\ || \\ || Los Altos
Los Altos Networks || // ---\\|| \\|| Networks
340 Second St. #6 ||___// \ | \ |
Los Altos, CA 94022
(415) 941-8031 UUCP: {apple,sun,uunet}!portal!cup.portal.com!cec
INTERNET: cec@cup.portal.com
"...No hay mal que por bien no venga..."
===============================================================================ddl@husc6.harvard.edu (Dan Lanciani) (12/03/90)
In article <36349@cup.portal.com>, cec@cup.portal.com (Cerafin E Castillo) writes: [ ... ] | PEP's failure in the impaired line tests was due to the lack of use, by | PCMag, of the hidden registers in the modem firmware (S120/J6S36). Well, gee, that seems perfectly reasonable. After all, if the registers are hidden and undocumented then the typical end-user can't use them either... | These | registers affect packetization for better survival of connection and | throughput over adverse line conditions ONLY. It was not in Telebits | best interest to start talking about or publishing this kind of information. Considering all the worry about the review, maybe it would have been in their best interest to document the registers after all... | Only recently has Telebit Tech Support introduced a document detailing | the use of these registers. Thus, in default PEP packetizing mode, the | failure of PEP was legitimate. How does one get obtain document? Dan Lanciani ddl@harvard.*
Alvin@cup.portal.com (Alvin Henry White) (12/19/90)
A friend and I bought Telebit T2500 modems about three months ago. We have never gotten them to work anywhere near what they are supposed to do. He is a retired doctor and an avid computerist. I am 48 years old and an avid computerist. I do not know anyone personally that uses telebits. I bought it based on written materials and word of mouth statements. I spent about 14 hours a day seven days a week for three weeks trying to get it to work. Best results are about 4800 baud telebit to telebit. If I call U.S.Robotics Dual Standard I always get 9600. Hayes ultra connect in v32 at 9600. Telebit to telebit no luck. Called customer service and asked if there was a bulleting board that I could call to practice on. They said no, they used to have one but shut it down. I think the report of the pc magazine is right on the money for a person who is thinking of purchasing one for home or small business. Mine is home. It someone knows someone who will gaurentee that they use one and that they will come to your house and set yours up to talk just to there's then I don't doubt that the telebit may be incomparable under certain line conditions. My experience was that you read the manual and found that you have about a hundred switches, who ever you are calling has a hundred switches, you need to guess how they have their switchs set and then guess which of your switches need to be set. Essentially, all wrong combination = failure. I read a lot on the net, and it is apparent that if one can find the right combination the machine is one of the best on the market for its purposes. alvin Alvin H. White, Gen. Sect. G.O.D.S.B.R.A.I.N. Government Online Database Systems Bureau for Resource Allocations to Information Networks [ alvin@cup.portal.com (OR) ..!sun!portal!cup.portal.com!alvin ]
bill@bilver.uucp (Bill Vermillion) (12/19/90)
In article <37045@cup.portal.com> Alvin@cup.portal.com (Alvin Henry White) writes: >A friend and I bought Telebit T2500 modems about three months ago. We have >never gotten them to work anywhere near what they are supposed to do. He >is a retired doctor and an avid computerist. I am 48 years old and an >avid computerist. > My experience was that you read the manual and found that you have about >a hundred switches, who ever you are calling has a hundred switches, you >need to guess how they have their switchs set and then guess which of your >switches need to be set. Essentially, all wrong combination = failure. All you have to do is go through the manual one page at at time. Read the register descriptions, and then the register to the value needed. You don't need to know what the other person is doing with the exception of doing spoofing, then you set register s111 to 255 when being called. Did you REALLY read the manual. Look at table 5-1. You do ~&F, then S55=4 (or appropriate) then s66=1. Then you run. That works for most cases. You mention you are and avid computerist and that you are 48. You age probably has a lot to do with, when you put on a few more years and get to be as old as I am you should be able to get on up an running in 1/2 hour or so. :-) !! Chalk your problems up to immaturity :-=) :-) :-) My 2500 joins my prevous TB+, and I wouldn't have anything else. The HSTs I connect to have no problems with the v.32 or v.42. -- Bill Vermillion - UUCP: uunet!tarpit!bilver!bill : bill@bilver.UUCP
pj@pnet51.orb.mn.org (Paul Jacoby) (12/21/90)
[comments regarding difficult in setting up a T2500] That's odd; when we installed two Telebit T2500's on our UNIX machines, I think we only had to change two parameter settings and about 4 registers. We found the easiest way to configure the modem was to start with AT&F, and then change one thing at a time. Having a target system to dial into helps loads, but as long as you hit both modems with factory defaults, then make small, testable changes, all goes pretty smoothly. After a week of '10 minutes here, 5 minutes there' configuration, we had both T2500's jammin with PEP between them, V.32 to a couple 9600 boards, and 2400/MNP5 to appropriate destinations. To say nothing of the work it took to set the silly /usr/lib/uucp/* files! That was far more challenging. Of course, your mileage may vary. .-----------------------------------------------------------------------------. | UUCP: {crash tcnet}!orbit!pnet51!pj | RTFD = Read The Silly Doc! | | INET: pj@pnet51.orb.mn.org | | `-----------------------------------------------------------------------------'
mikes@iuvax.cs.indiana.edu (Michael Squires) (12/25/90)
I've been running a TB+ for several years now, and recently purchased a TB2500. The setup took about fifteen minutes, but without having set up several other TB, TB+, and TB2500's before it would have been much more difficult. If I remember correctly the T2500's V.25 mode was a "gotcha", as is S50/S92 for negotiating a PEP link correctly. I'm 45 :-)... Both modems are working fine, best modems I've owned (also have a USR Dual Standard). -- Mike Squires (mikes@iuvax.cs.indiana.edu) 812 855 3974 (w) 812 333 6564 (h) mikes@iuvax.cs.indiana.edu 546 N Park Ridge Rd., Bloomington, IN 47408 Under construction: mikes@sir-alan.cica.indiana.edu