[sane-devel] Plustek Scanfx /w Proprietary ISA card

[Fred Stevens]

Gerhard Jaeger wrote:

>Hi Fred,
>
>On Friday 31 December 2004 22:29, Fred Stevens wrote:
>  
>
>>Hello,
>>
>>I have a Scanfx with an ISA adapter card.  The card is mapped to a
>>jumper selectable range of addresses (0x220,230,260,270,320,330,360,370)
>>and IRQ settings(3,4,5,10,11,12).  After looking at the card, I
>>determined that it is similar to a parallel port with the exception that
>>the data bus is bi-directional.  There are only a handful of LSttl chips
>>(74LS245, 74LS244, 74LS86, 74LS74, 74LS174 and a 16?? pal) and one PAL
>>on the board so it can't be doing too much.  The address range of the
>>card is four bytes (I think) and they are mirrored in the next four
>>bytes since address line A2 isn't connected on the card edge.  The A0
>>and A1 lines are connected, as well as A3 through A9.  Nothing above
>>that along with  ~iow and ~ior only being used from the ISA bus.  IRQs
>>3,4,5,10,11,12 are connected and are jumper selectable.  The connector
>>attached to the cardis a high density 25 pin Dsub style.  In the scanner
>>unit the LSI thatis connected to the interface (normal 25 Dsub on the
>>scanner side) is a Plustek 92001.  I am going to do some more hardware
>>debugging to see what the card does when I have some time.
>>
>>Is this information useful to anyone?  The scanner is a paperweight to
>>me unless I can use it under Linux since I stay away from that other
>>operating system for intel platforms.  Even so, the drivers available
>>for Win95/98 aren't very good.  My dad gave me the unit when he was done
>>using it under Win98 for that reason.  It's rather well made though and
>>could be useful attached to my print server.
>>    
>>
>
>the description is more or less useless, because you need to know how this
>ASIC works. I think its a predecessor of the 96001 ASIC, which is used in
>later Plustek parallelport scanner and supported by the plustek_pp backend.
>As this ASIC if pretty old, I don't have any technical documentation on that.
>
>Ciao,
>Gerhard
>
>  
>
Well,

I got the beast to talk to me.  After some debugging under dosemu, I 
came up with some init command strings and some more data on what 
happens when a scan is initiated and terminated.  The latter I am still 
processing.  So far, commands are sent and received at location 0x271, 
scan data is received at location 0x270 and interrrupt is acknowledged 
at location 0x273.  I couldn't find any access to location 0x272 so 
far.  2544 bytes are received during a scan and then an interrupt is 
generated.  The dos driver seems to indicate that it is emulating an HP 
scanjet 1200.  The following text represents a cold initialization of 
the unit followed by placing the date and time on the front panel LCD 
(16x2 ) display.  After this process, the keys on the unit function and 
produce scan codes available at 0x271.  There is no interrupt generated 
when a key is pressed so polling must be used to detect a pressed key. 

// Initialize unit
271 < 0  // Write to unit
// Need to wait a while here
271 < 1b // Write to unit
271 > 81 // unit busy
271 > 80 // unit ready
271 < 56 // Write to unit
271 < 0
271 > d2 // Read from unit, not sure what this code is
271 > d2
271 < 0// Write to unit
271 > d2
271 < 0 // Write to unit
... //check to see if we are still busy or d2 code output??
271 < 0 // Write to unit

271 < 1b // Write to unit, check if busy in between these.
271 < 43 // Write to unit
271 < 5a // Write to unit
271 < 0  // Write to unit

271 < 1b  // Load scanfx.exe utility into memory 1b 54 writes to LCD 
display on front panel.
271 < 54  // 1b 4d is the next line on the display.  Status is checked 
between writes.
271 < 30  // The date and time are written to the display here.  The 
first two bytes are commands, the rest are ascii.
271 < 32  // The rest of the "gibberish" below is doing the same as this 
set is.
271 < 30
271 < 30
271 < 35
271 < 0

271 < 1b
271 < 54
271 < 31
271 < 30
271 < 31
271 < 2f
271 < 30
271 < 34
271 < 0

271 < 1b
271 < 54
271 < 32
271 < 31
271 < 35
271 < 3a
271 < 33
271 < 31
271 < 0

271 < 1b
271 < 54
271 < 33
271 < 35
271 < 32
271 < 0

271 < 1b
271 < 4d
271 < 32
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 0

271 < 1b
271 < 4d
271 < 33
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 20
271 < 0

I tried these out by exercising the port and found that things work the 
same as the dos driver and the scanfx.exe utility, at least for 
initialization.  When I have more time I will see what else I can find 
out.  Thanks for your response by the way, I hope that I'm doing 
something useful here.  I took a look at the code for the mustek scanner 
with a proprietary card but couldn't find anything out from that about 
how to create a back end for this thing.  I will look at it again though.

Take Care,

Fred