$Id: lexmarkprotocol.txt,v 1.6 1999/03/26 21:20:03 henryk Exp $
(C) 1999 Henryk Paluch
Lexmark 7000 - documentation of undocumented protocol
-----------------------------------------------------

This document describes communication protocol used by Lexmark 7000 printer.
The key problem is, that Lexmark consider the protocol top secret &
copyrighted, so all information bellow was guessed! Use on your own risk!



All Escape sequences begin with
<- *
1B 2A

<- *  m  .  @
1B 2A 6D 00 40 ....                   56h bytes   -- Init

<- *  .  s  0
1B 2A 07 73 30
-------- EPSON (?) or PCL (?) sequence - length - various

<- *  .  c
1B 2A 07 63
-------- again EPSON (?) -- various length


1B 2A 6D 00 42 ...
-------------- various init ? -- length - various

<- *  .  .  .
1B 2A 03 04 EA     -- 5 bytes -- 4EAh means offset
-------- paper shift

<- *  .  .  .  .  .  .    .  .  .   .   .   .   .  .  
1B 2A 04 00 00 L1 L2 6x?? N1 N2 HS1 HS2 HE1 HE2 00 00 

   "  3  D  U
   22 33 44 55 01 N1,2x<packets>
-------- print pixels
L1,2  - complete length of the sequence (including data) BE order
N1,N2 - number of packes, BE order
HS1,2 - horizontal offset start
HE1,2 - horizontal offset end
Should be valid: N = (HE-HS)+1

packets

every packet define exatly one vertical line, but it contains information
for both left & right inkjets!

3? FF - where number of zero bits in ?x2 means number of data bytes e.g.
3F FF - no byte follows
3B FF - 2 bytes
37 FF - 2 bytes
36 FF - 4 bytes
35 FF - 4 bytes
32 FF - 2 bytes
31 FF - 2 bytes
30 FF - 8 bytes

1F ?? - like 3? FF
1F FE - 2 bytes
1F E6 - 6 bytes
1F F6 - 4 bytes

NOTE: 2? ?? packets behave very strange. I'm using only 3? ?? packets...

2F FF - 2 bytes ?? why ??

       <--------byte1----------> <--------byte2---------->
2F FF | L1 R1 L2 R2 L3 R3 L4 R4 | L5 R5 L6 R6 L7 R7 L8 R8 |

L? = left ink
R? = right ink

ink jets layout:

         L1    
               R1
         L2    
               R2
         ........
         L8    
               R8

Horizontal resolution is always 600dpi.
Vertical resolution is 300dpi for one series.
Using either left & right series we got 600dpi vertically.
But we must take account the distance betwen two series
(about 16 - visible when doing align test)



3? ?? <bytes 2x zero bits in ?) byte format same as above (Li Ri)
For Example:

3F FC AA AA A0 00
3    F    F   C    LRLRLRLR LRLRLRLR LRLRLRLR LRLRLRLR
0011 1111 11111100 10101010 10101010 10100000 00000000
     ^           ^ highest jet  1= white 0=two bytes
     lowest jet

We should get (hopefully)
     0       -
           0  \
    .........   - number of 1 * 16  10 * 16 = 160 zeroes
     0      - /
           0
     1      -
           0 \
     .......   - 10 * Left black
     1       /
           0\
     .......  - 6 * white
     0      /
          0

    96  + 96 = 192 ink jets

3F sequence can drive 12 * 8 * 2 = 192 ink jets in head


<- *  .  e
1B 2A 07 65 ...  Note: various bytes after sequence  ?? why ??
----------- Paper Eject 

Colour printing
----------------

There is difference in 6x?? packets:

 00 02 01 01 1A 11  Black & White printing (Left cartridge)
 00 03 01 00 18 11  Colour printing (Right cartridge)
          ^^--- cartridge selection 01 = left (Black/Photo)
                                    00 = right (Color)
    ^^--------- motor speed 02 = slower (?) 03 = faster

It is possible to use same data packets for colour and Black & White
packets. 
Colour printing is real nightmare, because there is vertical space
between colors (ooouuuuggggghhhh!). It looks like this for one pass:
O        -| 
     O    |
O         |  64 Cyan Ink Jets
     O   -| 
  x
  x          small vertical space (!)
O        -| 
     O    |
O         |  64 Magenta Ink Jets
     O   -| 
  x
  x          small vertical space (!)
O        -| 
     O    |
O         |  64 Yellow Ink Jets
     O   -| 

The big problem is just in "small vertical space". Try to print
"stairsc.prn" and you'll find that...
 


Notes for Lexmark 5700:
-----------------------
It seems to have only two significant differences:
Different prologue (107 bytes long).
12th byte in main sequence 1B 2A 04 00 00
is set to 01 instead of 11...