HylaFAX The world's most advanced open source fax server

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [hylafax-users] [hylafax-devel] !!! USR/3COM modems and the latest CVS !!!



"Bernd Proissl" <news@proissl.de> writes:

> > They were entirely software-related (i.e. they were the bugs in Hylafax
> > Class1 driver). Most problems, known to me, has already been fixed. 
> These fixes are in current CVS? 
> Or are they only fixed in your local environment?

Most of the fixes has already been commited to CVS.

> > But now I use
> > Class 1 in the production environment (with Zyxel 1496), and it 
> > works quite
> > good.
> Send only? Or send and receive?

Send and receive, although the incoming traffic is rather low (I think 1
incoming fax per 10 outgoing ones).

Try Class 1 now and tell me your impression. If you find a bug (a
problematic session, that can be reproducted), let me know. We'll try to
fix it :-))

Hope to hear from you soon,
Dmitry

P.S. There is an issue that still needs some research -- percent of "No
response to MPS/EOP" faults IMHO is higher than it have to be. There is a
person who has "secrect knowledge" on this subject -- Dr. Harald
Pollack. Once he told me:

[---cut---]
>Currently they complain even more because of "no responce to EOP/MPS". Also 
>just ignore these mesagges and think that the page has been sent successfully?

Oh! This is quite an other problem, which can NOT besolved easily in class
2/2.0, because modem reacts with +FHS <> 0 in this case.

I know this behaviour (no response to EOP or MPS) very well, but I see no
solution. It mainly arises with Unix fax programs, but also slightly with
Windows fax programs (but much less with OS/2 FaxWorks and my FSEND).

My diagnosis is, that MAYBE the communication between DTE and DCE (PC to
modem) is not as fast as neccessary. This problem is quite clear, how it
works:

fax sender transmits fax data in phase c (V.29) and after RTC, it make a
pause of 75ms and switches to V.21 and sends EOP or EOM or MPS signal. Than
sender waits for reaction of receiver, which should send MCF or RTP or
CSI+DIS in case of EOM. 

If transmitter does not receive any response in a welldefined time window,
it repeats this post page message twice.

I know at least three different scenaria, which could happen that sender
gives up because no response was detected (response was even there, but at
the wrong moment).

But with class 2/2.0 it is impossible, to adjust anything in timing
behaviour of sender (that's why I prefer class 1 so much :-)

Zyxels developer have tried to fix some 'bugs' in timing after phase c, but
success was marginal (you often read in release notes, that one or the
other fax machine now will be work better, but that's not always the truth
:-)

All I can say is, better use class 1 ....
[---cut---]

As you see, Harald claims that timings can be adjusted to get better
results. One day I will ask him how to adjust them and then ... :-))




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




Project hosted by iFAX Solutions