$Header: /bsdi/MASTER/BSDI_OS/contrib/flexfax/MODEMS,v 1.1.1.1 1994/01/14 23:09:26 torek Exp $

Modem Bugs and Gotchas
----------------------

    Copyright (c) 1990, 1991, 1992, 1993 Sam Leffler
    Copyright (c) 1991, 1992, 1993 Silicon Graphics, Inc.

    Permission to use, copy, modify, distribute, and sell this software and 
    its documentation for any purpose is hereby granted without fee, provided
    that (i) the above copyright notices and this permission notice appear in
    all copies of the software and related documentation, and (ii) the names of
    Sam Leffler and Silicon Graphics may not be used in any advertising or
    publicity relating to the software without the specific, prior written
    permission of Sam Leffler and Silicon Graphics.

    THE SOFTWARE IS PROVIDED "AS-IS" AND WITHOUT WARRANTY OF ANY KIND, 
    EXPRESS, IMPLIED OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY 
    WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.  

    IN NO EVENT SHALL SAM LEFFLER OR SILICON GRAPHICS BE LIABLE FOR
    ANY SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND,
    OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS,
    WHETHER OR NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND ON ANY THEORY OF 
    LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE 
    OF THIS SOFTWARE.

This file contains information about modem bugs and gotchas that have
been gleaned from experiences making the FlexFAX code work with the
modems.  These problems are based either on my experiences or on
experiences of others that have worked with the software.  One should
always checked the version of the firmware listed.  Note also that if a
modem is not listed here then I have NOT tried it.  This list should be
sorted in alphabetical order.

Class 1 Modems
--------------

AT&T Paradyne		DataPort 14.4		122

o received HDLC frames do NOT include the FCS bytes (fixed in rev 133).
o carrier is sometimes dropped w/o notice during receive (rev 133)
o the command "ATH+FCLASS=0;A" does not do the right thing because the
  command parser does not parse the ";A"

Digicom			Scout+			2A19/2931

o +FMFR?, +FMDL?, and +FREV? all return "3" in some versions of firmware
o S17=2 can be used to force 1300 Hz CNG tones to be sent (the
  register is not documented)
o RTS/CTS is not honored while sending page data; software flow
  control must be used (*F2).  This is fixed in rev 3507/3506 firmware
  by adding a register S24 that can be used to control flow control
  in facsimile mode: 0 means use XON/XOFF, 1 means use RTS/CTS, and
  2 means use whatever is specified by *F.  Note however that S24
  is not permanent; it is reset after each disconnect.  On power on
  S24 is 0.  S24 is unchanged by &F.
o receiving at 4800 baud v.27 does not work correctly; received
  data is garbage		[fixed in firmware revision 3225/3310]
o commands must not be sent to the modem too fast, no more than
  every 50ms (or more) depending on what the modem is doing
o the modem gets confused by receivers that assert a data
  carrier before sending CED; a +FCERROR result is returned before
  usually returning "NO CARRIER"; this can sometimes be overcome
  by sending 1300 Hz CNG tones instead of the normal 1100 Hz tones

Hayes			Optima 2400+Fax96	TR00-J260-001 XXX

o receiving is unreliable; I was never able to get the receipt
  of TCF to work correctly in the short time that I had the modem
  (no data was delivered after establishing carrier before timeout)
o see also below under Class 2

Nuvo			Voyager 96424PFX	AF-C2500-E0

o RTS/CTS is not honored; use software flow control
o factory configuration of &C1 causes problems; do "AT&C0&W" to set
  DCD handling before running the server
o beware of battery drain

Supra			SupraFAX v.32bis	V1.200-H TR14-Jxxx-001

o the +FAE command that can be used to enable auto-detect of calls from
  data modems is only accepted in Class 0 (data mode); I've never been
  able to get it to work correctly (when setup the modem is either unable
  to establish carrier, or the modem locks up after notifying the host)
o RTS/CTS is not honored while sending/receiving page data; use &K4 for
  software (XON/XOFF) flow control
o see also below under Class 2

Twincom			144/DF			V1.270 TR14-Jxxx-001

o RTS/CTS is not honored while sending/receiving page data; use &K4 for
  software (XON/XOFF) flow control
o see also below under Class 2

USRobotics		 Sportster		Supervisor 4.1 (04/22/93)
						DSP 10 (03/23/93)

o timing problems have been observed with older firmware revisions;
  if you have problems, be sure to verify the firmware revision
o when sending DCD is raised when carrier is first received, but it is
  then dropped from the first HDLC receive until the point where the
  initial training is sent (from that point on DCD is asserted); this
  can be a problem because there is no way to override this behaviour
  which means that the host must be capable of ignoring DCD for sending
  to function correctly
o in testing the modem locked up during transmission to a Multi-Tech 1432BA
  modem and failed to respond to the host dropping DTR

Class 2 Modems
--------------

Everex			EverFax 24/96D		911027

o zero-padding in the EOL codes of received facsimile data is often
  not zero when the Phase B+D bit ordering is set to "direct"; a
  workaround is to use +FBOR=2 and do bit reversal in the host

Hayes			Optima 2400+Fax96	TR00-J260-001 XXX

o many commands (e.g. ATZ) generate annoying clicks & pops from
  the speaker
o using '@' in the dialing string causes the modem to return
  "NO ANSWER" when calling many modems
o the modem does not do a good job of holding reliable high speed
  carrier--when sending and receiving, data is frequently corrupted
  at both 9600 and 7200 baud (the phone line did not appear
  to be the reason as other modems worked fine in back-to-back
  testing)

Multi-Tech		MT1432BA		0307

o there are problems communicating with Sharp model 6200 fax machines;
  this is reportedly fixed in rev 0309 firmware
o the EC (error correction) subparameter reported in +FDIS: result
  strings can contain invalid values (2)

Practical Peripherals	PM14400FXMT		2.17

o defaulted subparameters in +FDIS and +FDCC commands are wrongly
  treated as zero
o +FDIS commands to set session parameters after establishing a
  connection with a remote machine are accepted but ignored; this
  makes it impossible to select/change data format (1D vs 2D encoding)
  and vertical resolution (98lpi vs 196lpi); since it is not in general
  feasible to set all session parameters before dialing the telephone,
  this problem makes it possible only to support the minimum set of
  Group 3 capabilities (1D data, 98 lpi)

Supra			SupraFAX v.32bis	V1.200-H TR14-Jxxx-001

o changing DCD handling can cause the modem to drop DSR; if done
  during initialization this can cause the fax server to get confused;
  be sure that the setting loaded from NVRAM on reset agrees with
  the value specified in the config file (&C1 by default); one
  possible way of doing this is to specify &F2 for the ModemResetCmds
  parameter
o the line counts in the +FPTS result code during a reception are
  total nonsense; the only way to deduce the number of lines in a
  received page of data is to parse the data on the host
o facsimile reception is locked at 19200 baud
o adaptive answer is not reliable
o the training algorithm can loop forever; client software must
  timeout +FDR requests
o RTS/CTS is not honored while sending/receiving page data; use &K4 for
  software (XON/XOFF) flow control

[Firmware rev V1.80-02 TR14-Jxxx-001 is a significant improvement over
 previous firmware; it is highly recommended that you upgrade your modem
 if you do not have firmware at or near that rev level.]

Telebit			T3000A			Version LA7.02

o when HDLC frame tracing is enabled, the modem sometimes does not
  send the +FCON status message to the host
o when carrier is dropped during the training phase that follows a
  +FDT command, the modem sometimes violates the Class 2 spec by
  returning "NO CARRIER" and not a +FHNG status string

Twincom			144/DF			V1.270 TR14-Jxxx-001

o changing DCD handling can cause the modem to drop DSR; if done
  during initialization this can cause the fax server to get confused;
  be sure that the setting loaded from NVRAM on reset agrees with
  the value specified in the config file (&C1 by default);
o instead of the normal &D3 setup string for RC144AC-based modems,
  &D2 should be used (note that this configuration must be done
  manually as faxaddmodem is incapable of distinguishing between
  RC144AC-based modems)
o the line counts in the +FPTS result code during a reception are
  total nonsense; the only way to deduce the number of lines in a
  received page of data is to parse the data on the host
o the modem has frequently been observed to just lockup during message
  receive (i.e. +FDR); it dropped DCD to the host but did not return
  <DLE><ETX> and/or a +FHNG code
o in testing the modem locked up during transmission to a Multi-Tech 1432BA
  modem and failed to respond to the host dropping DTR

Ucom A/S		Traveller 96		SSX996/1.32, 120691,SSC

o remote TSI strings are returned without matching quote marks;
  as a result the current software truncates the TSI strings to be null
o AT+FCLASS=2 causes the host-modem interface to be locked at 19.2KB
  irrespective of the modem configuration and baud rate in use prior
  to the command

Zero One Networking	ZyXEL 1496E		U1496E  V 5.02 M

o +FMDL? and +FREV? return the same string, the concatentation of
  the modem model and the firmware revision.
o some prom revisions do not include a trailing "OK" in response to
  +FDCC=?
o DC2 must be sent to the modem to trigger the transfer of Phase C
  data from the modem to the host after a +FDR command; this reflects
  revision 2388-B of the Class 2 spec and is different from the DC1
  required by 2388-A (and accepted by most modems)
+ the modem advertises that it supports the 2DMR data format, but it
  does not; this "advertisement" is both in its result string for a
  +FDCC=? query and in the DIS/DCS sent to clients
+ sending 2DMR data to the modem can cause it to lock up leaving CTS
  low to the host; some versions of firmware also send an infinite
  stream of trash to the remote side when this condition occurs
o in some prom revisions the result string from a +FDCC=? query does
  not conform to any syntax listed in either 2388-B or 2388-A;
  furthermore, it requires a context sensitive parser!
+ +FDIS commands to set session parameters after establishing a
  connection with a remote machine are accepted but ignored; this
  makes it impossible to select/change data format (1D vs 2D encoding)
  and vertical resolution (98lpi vs 196lpi); since it is not in general
  feasible to set all session parameters before dialing the telephone,
  this problem makes it possible only to support the minimum set of
  Group 3 capabilities (1D data, 98 lpi)
o sending non-Class 2 commands to the modem after establishing a
  connection often causes the modem to drop carrier; for example,
  ATM0 before a +FDR command will do this;
o the line counts in the +FPTS result code during a reception are
  total nonsense; the only way to deduce the number of lines in a
  received page of data is to parse the data on the host
o in later versions (5.05 for example) the syntax of the +FPTS:
  result code violates 2388-A and/or 2388-B
o there is no way to abort a session (i.e. the +FK command is not
  recognized)
+ when sending, the modem appears to immediately drop DCD when it
  loses carrier; this usually occurs after a final +FPTS is sent
  to the host, but before the requisite OK and and/or +FHNG strings;
  this action appears to be independent of whether or not DCD is
  configured to follow carrier (i.e. &C0 or &C1)
+ when receiving with a negotiated 14400 signalling rate, the +FDCS:
  status string indicates a 9600 baud signalling rate (br=3)
+ the reception of multiple documents appears to cause the modem to
  hangup (i.e. an EOM post page message causes the modem to hangup)
+ the modem sometimes sends received data to the host with the
  wrong bit order
o the scheme used to do adaptive-answer asserts a data carrier before
  falling back to a fax carrier;  this can confuse a few fax modems and
  make it impossible to receive facsimile without disabling adaptive
  answer (+FAA=0)
o AT+FLID? returns OK but not the current value of the local identifier.

+ = This problem is known to be corrected in firmware revision 6.01.

Other Modems
------------
Abaton			Interfax 24/96

o nothing of interest
