Re: [hylafax-users] EOP errors, Was: Little question

[Lee Howard <faxguy@deanox.com>]

Failure to receive a response to EOP/MPS is a difficult problem to diagnose
because of the situation.  The last communication from the remote occurred
before we even began sending the page image data... in some cases this can
be in excess of 60 seconds or so.  Consequently, there's a long period of
time in which something can go wrong.  Under normal circumstances these
errors can be that the remote runs out of paper or buffer space, the user
turns off the remote fax, the line disconnects, or whatever.

However, I have found that all USR/3Com modems (Sportsters and Couriers
alike) as well as all modems using Lucent chipsets (i.e. Multitech) have
very long internal delays after issuing commands such as +FTH=3.
Consequently, we lose accurate timing due to a lack knowledge of when the
transmission carrier ceases.  Because the length of delays are set in stone
by T.30, this can cause transmission problems and can result in failure to
receive a response to EOP/MPS when the remote is strictly obeying T.30
timing requirements.

If we fail to get a response to EOP/MPS, we really do not know if the error
occurred before the page was sent or after.  In my experience the error
usually occurs beforehand, but it's impossible to tell without having
control of the remote system, also.

Anyway, the way to fix the problem with USR/3Com and Lucent-chipset modems
is to stop using software pauses for timing after +FTH=3 and other similar
commands in lieu of using +FTS=n, which is the more appropriate procedure.
+FTS=n has not been used historically in HylaFAX because some chipsets will
respond with OK/CONNECT following +FTS=n after a period of time that is not
n*10 ms.  This actually is expected behavior and occurs because the modem
is waiting for the transmission carrier to cease before the silence begins.
 (I've also seen it to be less than n*10 ms, also, for the corresponding
reason.)

So, to conclude, failure to receive a response following EOP/MPS can be
cause by a number of things, but one of the things to do is to apply the
patch which Giulio pointed out:

http://bugs.hylafax.org/bugzilla/show_bug.cgi?id=197

This has not yet made it into CVS, but I suspect that it will eventually
make the 4.1.1 release.

Lee.



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