Brother MFC-5895CW driving me crazy

[jggimi]

We've had this multifunction device as the family printer for about a year. I'm pleased with everything about it ... except the ability to use it from OpenBSD. I've tried plain text, Postscript, and even GDI via foomatic's hl7x0 driver (via a similar printer's PPD). None of these services print anything, ever. All I ever see is communication with the printer on port 515. The printer lists these LPD queues:

Local> show service

 * 1 Service Name : BRN001BA967B8D7    Port Name : P1
       Filter=0     BOT=1    EOT=1  Option : BQ
       Protocol :  IP    TCP Port : 9100

   2 Service Name : BINARY_P1      Port Name : P1
       Filter=0     BOT=1    EOT=1  Option : BQ
       Protocol :  IP    TCP Port : 9100

   3 Service Name : TEXT_P1    Port Name : P1
       Filter=1     BOT=1    EOT=11    Option : BQ
       Protocol :  IP    TCP Port : 9100

   4 Service Name : POSTSCRIPT_P1     Port Name : P1
       Filter=0     BOT=10     EOT=4    Option : BQ
       Protocol :  IP    TCP Port : 9100

   5 Service Name : PCL_P1     Port Name : P1
       Filter=0     BOT=9    EOT=4  Option : BQ
       Protocol :  IP    TCP Port : 9100

   6 Service Name : BRN001BA967B8D7_AT       Port Name : P1
       Filter=4     BOT=1    EOT=1  Option : BQ
       Protocol :  IP    TCP Port : 9100

All these test printcap entries have failed to produce any printed output:

foo|test of MFC8600 hl7x0 driver:\
    :lp=:rm=192.168.2.56:rp=BRN001BA967B8D7:\
    :sd=/var/spool/output/lp:lf=/var/log/lpd-errs:\
    :af=/etc/foomatic/Brother-MFC-8600.ppd:if=/usr/local/bin/foomatic-rip:\
    :sh:

ps|test of PS:\
    :lp=:rm=192.168.2.56:rp=POSTSCRIPT_P1:sd=/var/spool/output/lp:\
    :lf=/var/log/lpd-errs:sh:

text|test of text:\
    :lp=:rm=192.168.2.56:rp=TEXT_P1:sd=/var/spool/output/lp:\
    :lf=/var/log/lpd-errs:sh:

bin|test of MFC8600 hl7x0 driver via another service:\
    :lp=:rm=192.168.2.56:rp=BINARY_P1:\
    :sd=/var/spool/output/lp:lf=/var/log/lpd-errs:\
    :af=/etc/foomatic/Brother-MFC-8600.ppd:if=/usr/local/bin/foomatic-rip:\
    :sh:

All that /var/log/lpd-errs ever shows is that lpd is restarted. Logging of packets with tcpdump(8) shows that there is a conversation, the right queue is requested, the data is sent ... and nothing ever prints.

Every couple of months, I Google this printer to see if anyone has ever had success with the printer without Brother's proprietary software. Nothing, to date.

Any suggestions?

[shep]

It looks like you are trying to use the built in wireless for network printing.

The manufacturers spec sheet indicates that it supports hp jet direct on port 9100

Supported Network Protocols (IPv4) ARP, RARP, BOOTP, DHCP, APIPA (Auto IP), NetBIOS/WINS, LPR/LPD, Custom Raw Port/Port9100, DNS Resolver, mDNS, FTP Server, TELNET, SNMPv1, TFTP, Scanner Port, LLTD Responder, Web Services (Print)

The printcap syntax that worked for me was

hp5l|lp|ljet4;r=300x300;q=draft;c=gray;p=letter;m=auto:\
    :lp=9100@192.168.1.2:\
    :rp=raw:\
    :if=/etc/apsfilter/basedir/bin/apsfilter:\
    :sd=/var/spool/lpd/hp5l:\
    :lf=/var/spool/lpd/hp5l/log:\
    :af=/var/spool/lpd/hp5l/acct:\
    :mx#0:\
    :sh:

Other options are to try cups or apsfilter. In gnome/xfce4 network printers can be "discovered' in cups

[J65nko]

Could it be a end of line incompatibility? In other words, it may expect a CR/LF pair (0x0d/0x0a) instead of a single the LF (0x0a). Or horror of horrors maybe it needs a Control-Z.

Just a shot in the dark

[jggimi]

Originally Posted by shep

It looks like you are trying to use the built in wireless for network printing.

Yes. How could you tell?

The manufacturers spec sheet indicates that it supports hp jet direct on port 9100...

Yes. I'd tried 9100@, but not rp=raw and not with apsfilter. Will try later today and report results. Thank you for the printcap example.

Other options are to try cups or apsfilter. In gnome/xfce4 network printers can be "discovered' in cups

I dislike CUPS primarily because of lack of diagnositics. I've used it before; if I'm forced to use it again, I will.

Originally Posted by J65nko

Could it be a end of line incompatibility? In other words, it may expect a CR/LF pair (0x0d/0x0a) instead of a single the LF (0x0a). Or horror of horrors maybe it needs a Control-Z.

Just a shot in the dark

Thanks, but if Brother did that they'd be violating RFC 1149 and they have other printers that are known to work with LPD just fine. My thought was they might be using one of their standard operating systems for a less than completely functional printer, and that the OS doesn't know the capabilities aren't actually present. For example, nothing in the specs mentions Postscript or BR Script, but the printer lists on operational Postscript service via Telnet.

[jggimi]

Same result, Shep. No output. I'll try CUPS, next. From a test environment. I'm not a fan of it.

[jggimi]

CUPS found the "BINARY_P1" interface only. The rest of the services were not found. But after configuring the found printer, CUPS was unable to find the queue to send a test page. I manually set the connetion to socket://192.168.2.56:9100, but that failed to send anything due to format errors.

I hate CUPS.

[shep]

I utilize a single printer with a parallel port print server. Like you I prefer the simplicity of lpd/printcap. I use Arch linux on a laptop that is not compatible with OpenBSD and access the printer via Cups. In Arch/Cups I do not have to use the HP direct 9100 printer port and use ipp://192.168.1.2/P1 or http://192.168.1.2/P1 or lpd://192.168.1.2/P1.

I would also try 192.168.2.56:9100/P1 and socket://192.168.2.56 from what I read here

socket://192.168.2.56:9100

may be redundant since socket automatically "happens" on port 9100

[roddierod]

Originally Posted by jggimi

Thanks, but if Brother did that they'd be violating RFC 1149 and they have other printers that are known to work with LPD just fine. My thought was they might be using one of their standard operating systems for a less than completely functional printer, and that the OS doesn't know the capabilities aren't actually present. For example, nothing in the specs mentions Postscript or BR Script, but the printer lists on operational Postscript service via Telnet.

I have a brother laser printer (hl-4040cdn), so maybe this doesn't apply to the inkjet...but I had to go into the configuration page and set treat LF as CR/LF to print correctly.

CUPS only worked sometimes for me so I switched to using LPD. I finally got LPD to work by making a custom filter to translate PS to PCL (again, not sure if this applies to the inkjet world). But anyway, here's the link to how I can not print without problem, it's FreeBSD but should work on OpenBSD.

http://www.wonkity.com/~wblock/docs/...dprinting.html

[jggimi]

Originally Posted by roddierod

...but I had to go into the configuration page and set treat LF as CR/LF to print correctly.

My issue is different -- I get no output at all, ever. And it's a guess, but I'm guessing this change you made was for text content, not RFC 1149 LPD commands. Even so, I'll look for a CR/LF setting, just to check.

I finally got LPD to work by making a custom filter to translate PS to PCL...

I had been testing PS -> Brother GDI using Ghostscript's HL7x0 driver. I'll retest with Ghostscript's ps2pcl, and report back.

Thank you for the suggestions.

[jggimi]

There's no place to set LF/CRLF from the printer's little interface, nor from Telnet. One can set EOT characters by service, though they appear to be standard (ctrl-A or ctrl-D, as noted in the top post).

Using ljet4 PCL, per the ps2pcl script, had no change in symptoms. No print, printer "receiving data" and the file long since considered sent, and out of spool. Tried with both the BINARY_P1 LPD port, and the 9100@ Jet Direct device.

Bleh.

[jggimi]

I even tried the proprietary Linux driver; all its binary needs is already in emulators/fedora_base. Same results -- clears spool immediately, nothing printed.

I shall continue to use the device from non-OpenBSD platforms: Windows, Linux, and Android (they have a nice app).

[roddierod]

Did you try changing the device in the ps2pcl script? I had pretty much the same problem as you were the spool would clear...I even got a blinking light on my printer meaning it was warming up to print but nothing ever printed (had this with CUPS and even using the lpd instruction from brother and there binary driver).

my ps2pcl looks like this:

#!/bin/sh
/usr/local/bin/gs -dSAFER -dNOPAUSE -q -sDEVICE=pxlcolor -sPAPERSIZE=letter -sOutputFile=- -

Here is my printcap just in case it helps: brother is defined in my /etc/hosts as 192.168.1.3

lp:\
      :lp=:\
      :rm=brother:\
      :rp=raw:\
      :sd=/var/spool/lpd/brother:\
      :mx#0:\
      :sh:\
      :if=/usr/local/libexec/ps2pcl:\
      :lf=/var/log/lpd-errs:\

[shep]

If you haven't been driven crazy yet you could also break the task into smaller parts:
1. Hook printer directly to usb and work out print cap.
2. Printer via NIC cable to add networking
3. Finally printer via wireless networking - I looked at the online manual which talked about WEP keys for the printer. Did not say anything about WPA

A quick test would be to ping 192.168.2.56 and watch the router lights
My Trendnet print server also has a web interface
From

lynx 192.168.1.2
   Change IP Address
   
    IP Address:
   (*) Manually Assign
   IP Address      :   192.168.1.2_______
   Subnet Mask     :   255.255.255.0_____
   Default Gateway :   192.168.1.1_______
   ( ) Automatically Assign              
   
       Save   Cancel

[jggimi]

Originally Posted by roddierod

Did you try changing the device in the ps2pcl script?

I had not. I will try the hl7x0 driver and then the pxlcolor driver. Thank you for that tip.

I even got a blinking light on my printer meaning it was warming up to print but nothing ever printed (had this with CUPS and even using the lpd instruction from brother and there binary driver).

Yep. I have an LCD display which either lights up or otherwise displays it is receiving data (forever). Depends on which of the many broken methods I've tried which I get.

Originally Posted by shep

If you haven't been driven nuts yet you could also break the task into smaller parts:
1. Hook printer directly to usb and work out print cap.

Thanks. Unfortunately, I am certain that the issue is not networking per se. I can reach the device with Telnet -- the service list in the top post came from a Telnet session. Also, tcpdump shows end to end conversations via TCP port 515 or 9100. All that stated, the problem source could just be a bug in the printer's control program, where incoming WiFi traffic for LPD or Jet Direct is just dropped. We have plenty of systems that print via WiFi connection via proprietary driver or proprietary Android app. Even so, I will attempt a direct connection to rule WiFi out as a possibilty.

[jggimi]

*&^%$#!!!!

All new tests failed, including USB interface.

[roddierod]

Thats sucks.

The only other suggestion I can possible suggest is try using netcat to print a text file, just to see if it produces any output at all.

[shep]

Did you find this web site?

The site describes linux emulation, installation of the *.rpm based printer drivers from Brother and CUPS. The end result looks bloated and antithetical to OpenBSD but it is reported to work.

[jggimi]

I've used Telnet with port 515, attempting to print stuff via LPD protocol, to no avail. But that could be due to my misunderstanding of the protocol.

[jggimi]

Shep, no, I used the proprietary LPD driver under compat_linux, not CUPS. That, I suppose, will be my next (and last) attempt.

[shep]

An archived OpenBSD Brother thread with multiple options.

I have a feeling jggimi will eventually pound this into submission.