Re: [hylafax-users] Occasional Total and complete failure

["Paul A. Franz, P.E." <paul@eucleides.com>]

At 02:57 PM 4/8/2001Lee Howard sez:
>For the most part, faxes should occur at 14400 or 9600 baud.  If it's
>slower that that, there's probably a communication problem of sorts.  2400
>baud communication is indicative of an extreme problem, and honestly I'd
>say that it would be a good idea to set HylaFAX up to terminate faxes that
>go so slow if they're common at all.

This problem seldom occurs but is 100% repeatable when talking to specific 
modems. Guessing about 1 in 40 or 50 have a problem talking to my Telebit 
WorldBlazers.

When the problem occurs it is always a Phase B/C training error and the 
resulting reported info is 2400 2-D MMR. All normal completed faxes go as 
9600 1-D MR with a rare connection at 7200.

> >No apparent settings for those last two (Class2HFLOCmd, and Class2SFLOCmd).
> >Do I need them?
>
>I don't know.  You need them if you need them, and you don't if you don't.
>It depends on your modem and what the default settings are.  Since you're
>working from a faxaddmodem-made-from-prototype file, I'm assuming that the
>creator believed that the default settings for the modem were sufficient,
>and that Class2HFLOCmd and Class2SFLOCmd were not necessary.  There are no
>HylaFAX defaults for these commands.

The last line of this sample config is truncated. You can tell because the 
comment is incomplete. Maybe Class2HFLOCmd and Class2SFLOCmd needed to be 
assigned? Do they influence anything to do with Phase B/C training?

>You'll need to review your manual if you want to provide them, since they
>can vary from modem to modem.  And many modems do not have values for these
>settings.

Years ago I had a stack of these modems connected to a Novell Netware NACS 
and I remember hearing the same handshaking sequence then too once in a 
while. At that time I attributed the failure to cheap fax machines as the 
Telebits were around $1400 each and some fax machines were done to $45. I 
don't know though. As is the case now, I don't yet know what machines I 
cannot connect with or even if this is an dial in or dial out only problem.

The problem also was occasionally evident with WinFAX. WinFAX uses Class 1 
only. And until very recently I ran my modems as Class 1 on HylaFAX too. No 
apparent difference.

> >Any suggestions on how to proceed?
>
>Try Class 1 rather than Class 2, or try xonxoff rather than rtscts.

Same with Class 1 and Class 2. Xon/Xoff would only have an effect after 
linkup. The problem occurs in Phase B/C training.

>If you find that certain remote machines prefer one arrangement while
>others prefer another, then you may need to see what you can do about it
>via destcontrols, and if that's no good, then you need to look into another
>modem, perhaps in addition to the current one.

I could use a USR modem as Class 1 as they are known to be bullet proof. 
But I want to make these old Telebit modems to work. BTW, I have a Web 
Server that is a 486/66 EISA with 128 MB of RAM and all SCSI and all EISA 
cards too. It is quite nice actually. Runs, the Firewall, serves a few 
thousand web pages everyday, is a print server, runs Squirrel Mail, runs a 
bunch of mail lists and handles all the company mail and secondary DNS. So 
you're talking to a guy with a lot of reluctance to abandon expensive but 
older equipment.

So, I am fishing for what causes problems in Phase B/C training? Maybe it 
is my hardware but it only happens to some modems and to those it is 
repeatable.


Paul Franz

PAF Consulting Engineers       |  427 - 140th Ave NE
(425)641-8202 voice            |  Bellevue, WA 98005
(801)749-8480 fax              |  <http://blackdog.bellevue.wa.us/>



____________________ HylaFAX(tm) Users Mailing List _______________________
 To unsub: mail -s unsubscribe hylafax-users-request@hylafax.org < /dev/null