Re: [hylafax-users] Unable to initialize modem

[Lee Howard <faxguy@deanox.com>]

At 11:31 PM 11/14/01 +0100, Edward van der Jagt wrote:

>But now things stop being fine.
>Faxq now spawns a faxsend process. But this process ends and the syslog shows
>a message about the modem being wedged.
>If I look at the modem, I can see that DTR gets dropped (light goes out) and
>nothing else happens from then on.
....
>Seeing as the problem appears to start when DTR is dropped, I've tried
changing
>the related command to "" in the config file, but DTR still gets dropped and
>the problem persists.
>Changing the flow control between xon/xoff and rts/cts doesn't change
anything
>either.
>
>
>Can anyone give my any tips on how to get this working ?
>I'm getting pretty desperate, no make that frustrated, from all of this.

In HylaFAX 4.1rc1 (in CVS as of 20 May 2001) there was a correction made to
the code (which I think applied to both Class 1 and Class 2/2.0 setups, but
I cannot be sure without looking deeper, but I KNOW that it applied at
least to Class 1) to correct the DTR drop timing.  It also created a config
option ModemDTRDropDelay.

You may want to try using Class 1 and/or changing the settings for
ModemDTRDropDelay.

From 'man hylafax-config':

       ModemDTRDropDelay
              The time, in milliseconds, to pause between placing
              DTR OFF and DTR ON while resetting a  modem.   This
              value should be at least equal to the time-value of
              the analogous S-register, usually S25.

The default for ModemDTRDropDelay is 75 ms.  However, you should check your
S-register (probably S25) to see what the required length is for your modem
and ensure that ModemDTRDropDelay is longer than it.  You may also just
want to start increasing it to see if there is any result.

The other timing factor at this moment is ModemResetDelay.  The default for
that is 2600 ms.  You may also want to adjust that setting to see if it helps.

So, you should see something like this (under Class 1):

Oct 31 12:16:18.05: [29075]: SESSION BEGIN 00004837 15106084511
Oct 31 12:16:18.05: [29075]: HylaFAX (tm) Version 4.1
Oct 31 12:16:18.06: [29075]: SEND FAX: JOB 3861 DEST 1 510 608 4511 COMMID
00004837
Oct 31 12:16:18.06: [29075]: DEVICE '/dev/ttyS1'
Oct 31 12:16:18.06: [29075]: MODEM set DTR OFF
Oct 31 12:16:18.06: [29075]: DELAY 75 ms
Oct 31 12:16:18.14: [29075]: MODEM set DTR ON
Oct 31 12:16:18.14: [29075]: DELAY 2600 ms
Oct 31 12:16:20.74: [29075]: MODEM set baud rate: 38400 baud, input flow
RTS/CTS, output flow RTS/CTS
Oct 31 12:16:20.74: [29075]: MODEM flush i/o
Oct 31 12:16:20.74: [29075]: <-- [4:ATZ\r]
Oct 31 12:16:20.87: [29075]: --> [2:OK]
Oct 31 12:16:20.87: [29075]: <-- [5:ATE0\r]
Oct 31 12:16:20.88: [29075]: --> [4:ATE0]
Oct 31 12:16:20.88: [29075]: --> [2:OK]

You can see a log example of the "wrong" thing (and the problem that was
fixed) at:
http://www.hylafax.org/archive/2001-11/msg00170.php

If you are indeed using 4.1-release, then you should know it by 'faxstat
-i'.  Anything prior to the release candidates did not disclose version
information this way.  (The info file is created during faxsetup, so if
you've upgraded since you ran faxsetup last it may differ.)

If this doesn't help please re-post a session log or server log with the
tracing level at 0xFFF.

Lee.



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