|Home • Why 56k=v.Unreliable • Thhe Original 'Why 56k will NEVER work reliably"|
It has been almost a year since US Robotics (now 3Com) released its x2 technology. They have never been able to work out the bugs that plague a large (but unknown) number of users. Then came 56Kflex. While I have less experience with this 56k protocol, it has even more problems relating to multiple vendors with their own implementation of the protocol in addition to many of the same Telco network incompatibilities of x2.
It has been 8 months since USR Courier support told me my problem was being handled by a special group that would "definitely" call me back. I have to this date, never received a call back from USR.
My x2 non-experience began in June 97. Since then, GTE placed a new switch in the office serving several of the ISP providers, and I can now get x2 connections at up to 48k. 37.3k is most common. But, sometimes the connection is solid, and other times it is so bad that the modem will retrain down to as low as 2400bps and/or disconnect. I strongly suspect the problem is modem firmware incompatible with the digital facilities used by GTE in part because the speedshift occurs on the receive side only - ie, my rate can become 2400/28800! On these lousy connections, the ARQ light on my external Courier will flash almost constantly, and the MR light will go out for 5-10 seconds as the modem retrains and data flow is suspended.
I get the same results on all the ISPs located in Hilo served by the new DMS100 switch. The ISPs located in Kailua-Kona are still served by a GTD5 switch, and it is impossible to get an x2 or Flex connection when I call these numbers. I get poorer throughput with Flex than x2: to the Flex ISPs in Hilo, I get 32000 or 28.8 (GTE!) and 34000 (Interpac). 34000 is the highest Flex rate I've ever seen (using an LTWinmodem);. 48000 is the highest x2 rate I've seen.
I wrote the following diatribe in July 97. I focused on the obvious digital pad problem. Since then, I have learned that the GTD5 switch does not use digital padding - and now conclude that in addition to the digital padding, there are still more unidentified problems preventing 56k connections. In addition, shortly after I first posted the text in a newsgroup, 3Com totally changed their 800# Courier Support line, and the message referring to 3/31/97 disappeared with the change. Also, something I hadn't considered may be in play here: while there have been hardly any updates to the client modem firmware, there have been many x2 server firmware updates. As I understand it, USR equipment requires each modem in the ISP's racks to be flashed individually - creating a situation where it is more than possible that you can get modems playing from a different set of rules from call to call dialing the same ISP!
At the end of this document, you'll find links to other sites with more information. I also plan to add a section on some things that a user may be able to do (short of moving) to alleviate 56k problems. Also, I have links for GTE and US West's network disclosure which can help you find out what kind of telco switches you are dealing with. The Telecom Act of 1996 requires public disclosure, and I so far have been unable to find sites for the other local telcos. If you know of such site(s), please let me know so I can update the list.
In June, the "experts" answer to non-x2 connections was "You must have more than 1 A/D conversion on your line." Their tune has changed, but they still are clueless.
Why x2 will NEVER work (reliably)(reliably)
***** CAUTION ****
This is a pre-release "beta" version for use at your own risk. By using this version, you agree that the author cannot be held liable for any damage, direct or indirect, caused by your use of this version. There may be a number of incorrect facts and assumptions scattered anywhere herein. The author makes no claim of being an expert in the operation of the PSTN (public switched telephone network), and may only know enough to be dangerous.
Version 3.2.1 7/20/1997
First there was Watergate, then Whitewater, now X2GATE is unfolding before your very eyes --- it's becoming very clear now why some people are having problems out-of-the-gate (that supposedly shouldn't) with x2, and why all those "lucky" ones that can get x2 today are at risk of loosing that capability at any moment.
The proof comes from USR's own published documents.
Today, some unknown number of users are unable to achieve x2 connections with local ISPs. The focus has been on the "strict" requirements of the facilities used by your local telephone company monopoly that provide you with access to the PSTN. But, it is obvious that there are a large number of users who have the necessary local loop (customer premises to telco central office) facilities to achieve x2, but can only achieve x2 on TOLL calls. This occurs because a different DIGITAL PAD is used on the toll call versus the local call, and the local call is using a PAD which USR has NOT identified. (See page 3 of the first link where USR states, "It is possible that we have not identified all the possible pads." Further, pay attention to the statement in the section regarding Problems with the network, that "The most common issue is a digital pad where the digital bits (PCM) are manipulated to reduce the volume. This causes problems for x2, but it has been overcome using complex mathematical computations within the modems at each end.."
USR states that digital pads, (which alter the digital data stream) can cause problems, but they compensate for that with complex computations.
USR admits that there may be PADS they have not been identified so it follows that no compensation for these pads is provided in the firmware. (That would cause failure of x2 where it should be successful.) If you still don't believe, see Appendix A of the troubleshooting guide for an explanation of the "SERIOUS" errors - "Unspecified Negotiation Failure".
Any user achieving x2 today is at risk of loosing it at any time - the digital pads are controlled by software, and the operating TELCO monopolies have no obligation to make USR aware of any new software they might employ. The TELCOs are free to change the software at any time - creating the potential for massive disruption of what were successful x2 connections. (Source: local telco)
If you look carefully at their wording, they use words like "most likely" or "quite possible, if not likely" when describing their ability to provide compensation for the network problems they have not yet identified. They do NOT say they WILL. Because they cannot! Now, this document is a technical document, and I don't think marketing quite understands the implications of what has been released here. And while I think that USR must have some "kickback" from long-distance calls to their BBS (although I don't know how that could be legal or possible to arrange), why else would they make customers of their product pay for a TOLL CALL to their BBS to obtain a firmware update to fix the problems they are successful at identifying and correcting, and NOT MAKE IT AVAILABLE VIA THE INTERNET???????
This company, now a part of 3COM has REALLY SERIOUS problems in front of them. If you've ever tried "dealing" with a TELCO, you come to understand they hold a lot of power - they're the only game in town. And while the intent of recent legislation is to change that, they're not going to give it up without a fight. Do you really think we'd still have an infrastructure where you PAY for touch-tone service as opposed to rotary - which ALL the switches in ALL the local TELCOs still support if there was competition? Do you think that a TELCO could get away with charging $7. a month for CALLER-ID service if anybody other than them could provide that same service to you. (To enable it, all that's necessary is somebody punching in a few keystrokes. The only real cost involved with touch-tone and call waiting (and a host of others) is the accounting and control costs involved in limiting the service to those who are willing to pay the price. They are not market-driven. What incentive do all the major telcos have to cooperate with USR when their attitude typically is "we provide voice service, if you've got a data problem, that's not our problem". Or, as I was recently told by one "front-line" local TELCO employee, "oh data... We support up to 9600bps on our lines." Can you imagine what you'd run into if you wanted to get ALL the telcos to cooperate with you, and have some control over their implementation of software in their digital pads? And do you really think a company as arrogant and insensitive to its customers will have the skills necessary to pull off a deal to get all the telco's to "cooperate"?
Reference:: US Robotics x2 Troubleshooting Guide for ISPs - http://totalservice.usr.com/x2shoot4.pdf
My shouting for help from USR has yielded nothing more than automated responses from robots.
Do you think you'll fare any better if you lose your x2?
One of many messages to USR: (if you can find a way to get a message to them.... See more below)
It's been over 3 weeks since I got my Courier V.everything x2'ed, and I'm
almost at the end of my "free" trial with IBM's ISP service - the only one
offering a local access # for me - a # that, if x2 is enabled on the modem,
provides me with a 12.2 - 21.6k connection, while if it is DIS-abled
(s58=1), I can get reliable 28.8k connections.
It's been over 3 weeks since your TOTALLY AWFUL support told me that because
I'm getting "Unspecified Negotiation Failure", my call would be escalated,
and someone would get back to me. TOTAL JOKE.
I'm still going around in circles with IBM's help desk - they haven't got a
clue. I'm back to the "you probably have more than one A/D conversion....",
something that had (I thought) been TOTALLY eliminated.
Well, it's pretty clear to me now that I've found the necessary technical
information buried in your TOTALLY AWFUL web-site, that the issue on page 3
of your x2 Troubleshooting Guide for ISPs (which, I guess most ISPs haven't
seen, read or used): "It is possible that we have not identified all the
possible pads...." is the reason that I, as well as people from all over the
country using Telcos including GTE (mine), Bell Atlantic, BellSouth,
Southwestern Bell, PacTel, and more are unable to achieve an x2 connection
on a local access #. (I'm sure the marketing department hasn't seen this
document, or it would never have made it to the web.)
My attitude towards USR/3COM right now is I hope you burn in hell. Or at
least, go bankrupt. Or lose a class action lawsuit. -OR- repent and mend
your ways and BE HONEST and more customer-oriented.
To those that insist there's more than 1 A/D conversion if you have an x2 problem:
Wake up and smell the coffee.
From USR's web site "Troubleshooting x2 problems" document for ISPs. http://totalservice.usr.com/x2shoot4.pdf
It says, in part:
THE FOLLOWING ARE SERIOUS ERRORS. IF YOU SEE THEM PLEASE CALL U.S. ROBOTICS TECHNICAL SUPPORT.
Conditions that produce these errors are extremely rare. They will most likely be resolved in future code updates.
Unspecified negotiation failure
x2 should have worked, but didn t. Follow the client troubleshooting steps and see if this problem exists when calling different, known x2-capable, hosts. Collect as much data as possible and call U.S. Robotics Technical Support! It will be very valuable to help engineering root out the cause if they know whether you are never able to achieve x2
or if the problem arises only when calling certain destinations!
I don't know what they mean by extremely rare, but try do a USENET Alta Vista Search on usr x2 problem. Plenty pages.
What USR is saying here is that all the criteria they think is necessary for an x2 connection are met, but for a reason they don't know, it doesn't work. Try and find that spending a few minutes on their web site.
When I bought this Courier, v.32 was the standard, and USR had "lost" with their early HST high-speed technology that was first to market with near giveaways to BBS operators in an effort to gain market share. I got it with a coupon for a free upgrade to 28.8 (firmware), and the promise of future upgradability. Until this x2 fiasco (I think it's gonna get
out of control), they released products that pretty much WORKED, and if they didn't, it usually was a hardware problem that could be FIXED with a repair or replacement. This non-working (in "extremely rare" cases) problem is a different story. They say it "will most likely be fixed" with future firmware updates. I think they've got the Mars rover
computer working on this one, and it may take a little longer due to unexpected glitches and 'resets'.....
I registered my USR Courier external modem in Feb 97 in time for a "free" upgrade to x2 - originally promised for March 97. If I had plunked down the $60 bounty they want now, I would be STEAMING. Just today, 7/18, USR
posted their first "update" to the x2 firmware - and it DOESN"T FIX the problem. IBM is the only provider on this island "up" with x2, but that means a LOWER speed connection with a handshake that lasts over 60 seconds before carrier unless you DISABLE x2. USR "escalated" my trouble call of June 23 to a "special group" that "believe it or not calls everyone back" that is working on the problem. They just can't tell you when that might be. This year? Shoot, you can call the USR Courier "TotalJoke Support" TODAY (800-231-8770) and you'll hear instructions on what to do IF
you registered for the Courier upgrade and don't receive it by March 31 of this year - as if that date hasn't yet arrived. These guys are so out of touch and have such a lousy web site that it will take you a little while to realize that your ONLY option for support is to call their 800# M-F from 3am-noon hawaii time and WAIT ON HOLD F O R E V E R. Any promising e-mail address you might find on their support site will only be answered by a robot. And God forbid if you have a problem with a Sportster, because your only option then becomes to wait on hold forever while YOU pay the toll charges, or call a 900 number. And if they finally do come out with some firmware to fix the product you plunked down good money for, you'll be able to download the DUMBEST wizard I've ever seen from their web site, which will FORCE you to make a TOLL CALL to their BBS to download the actual firmware. You cannot get it on the web through your ISP account. Does USR have a deal to get long-distance revenue kickbacks? There isn't another computer vendor I deal with that doesn't offer their fixes through the web. And, if you've got the x2 incompatibility that causes your modem to connect at 12.2 or maybe up to 21.6 instead of the 28.8 you could have if you disabled x2, your toll call will last twice as long because the wizard is smarter than you and will take care of all the modem settings - and it knows better than you and will enable x2. And unless they've changed the latest release, it won't tell you until after its spent well over an hour on that toll call that you already have the LATEST CODE!!!! It
doesn't tell you how long you've been on-line, how much data has been transferred, or how much is left to transfer. If you have an internal modem, you don't even know if anything at all is coming down, because all it says is Downloading....." Gee. And that Sportster I bought on June 24 from Wal-Mart had an insert giving me a special 800# for installation assistance through APRIL 30, 1997. Wow. But, Wal-Mart will take it back if it doesn't work, right? Well.... They tried to tell me it was SOFTWARE and subject to their "no refund" policy on opened software.
Right. Fight. The assistant manager finally gives in after I refuse to take another one because I know it not gonna work either. Heck, I tried it on more than 1 phone line in more than 1 computer at different locations. So, it's the local phone company's fault that x2 doesn't work when they give you "crystal clear" all-digital local voice connections that meet ALL the requirements USR says you need, and it happens in areas served by: GTE, Bell Atlantic, Bell South, NYNEX, SouthwesternBell, PacTel..... and......
Please. Get real.
I saw some USR propaganda that says they expect that x2 will not work in about 5% of the cases. My suspicion is that they severely underestimated this - but, I guess I just haven't found the right place to look to find their TOTALLY OPEN discussion of these issues and how many users are actually affected.