[sane-devel] Re: Canon 4400F

Maximilian Fabricius mfabricius at gmail.com
Mon Jan 1 15:24:38 CET 2007


On 12/31/06, Pierre Willenbrock <pierre at pirsoft.dnsalias.org> wrote:
> Jaagup Repän schrieb:
> > Jaagup Repän wrote:
> >> Hello.
> >> I have Canon 4400F scanner. This is not supported, but it must to be
> >> added to genesys backend. I am trying to add, but I need help. I
> >> have to find SENSOR, DAC, GPO and motor type, shading and search lines,
> >> start of scan area, white strip and black mark.
> >>
> > I forgot to say that I have USB sniffer log
> > file(http://jrepan.pri.ee/usbsnoop.log.decode) and I found some
> > information from
> > http://lists.alioth.debian.org/pipermail/sane-devel/2005-December/015750.html
> >
> >
> > I got frontend:
> >     {{0x02, 0x20, 0x30, 0xdc}
> >    , {0x32, 0x00, 0x03}
> >    , {0x10, 0x32, 0xc8}
> >    , {0xda, 0x00, 0x00}
> >    , {0xf0, 0x00, 0x00}
> >
> > Is this correct?
> >
>
> >From your log i get this:
>
>  {{0x00, 0x23, 0x24, 0x2f}
> , {0x00,0x00,0x00} //never set in log
> , {0x60,0x60,0x60}
> , {0xfa,0xfa,0xfa}
> , {0x00,0x00,0x00} //never set in log
>
> look for this sequence:
> wrote 0x51 <frontend register address>
> wrote 0x3b <high bit of data>
> wrote 0x3a <low bits of data>
>
> those are frontend register writes(at least in your log).
>
> Regarding the Motor struct, this will probably work(without looking at
> your log)
>
> {
> 2400,
> 4800,
> 1,
> {
>   12000,
>   12000,
>   1,
>   1.0
> },
> {
>   12000,
>   12000,
>   1,
>   1.0
> }}
>
> This should lead to very slow movements of the scanning head, but it
> should move if the scanner is correctly initialized(the Canon LiDE
> 35/50/60 are a bit tricky, and the 500F seems to be similar).
>
> As you can see, i use half of the real resolution here. The reason is,
> the backend is not (yet) capable of using something other than 2400,
> 1200 or 600 as sensor resolution. In the meantime, we can make double
> the scanner dimensions, to accommodate for the real 4800dpi.
>
> The gl841 part of the backend also lacks support for non-LiDE color sources.
>
> Hope that helps,
>   Pierre
>


On 12/31/06, Pierre Willenbrock <pierre at pirsoft.dnsalias.org> wrote:
> Jaagup Repän schrieb:
> > Jaagup Repän wrote:
> >> Hello.
> >> I have Canon 4400F scanner. This is not supported, but it must to be
> >> added to genesys backend. I am trying to add, but I need help. I
> >> have to find SENSOR, DAC, GPO and motor type, shading and search lines,
> >> start of scan area, white strip and black mark.
> >>
> > I forgot to say that I have USB sniffer log
> > file(http://jrepan.pri.ee/usbsnoop.log.decode) and I found some
> > information from
> > http://lists.alioth.debian.org/pipermail/sane-devel/2005-December/015750.html
> >
> >
> > I got frontend:
> >     {{0x02, 0x20, 0x30, 0xdc}
> >    , {0x32, 0x00, 0x03}
> >    , {0x10, 0x32, 0xc8}
> >    , {0xda, 0x00, 0x00}
> >    , {0xf0, 0x00, 0x00}
> >
> > Is this correct?
> >
>
> >From your log i get this:
>
>  {{0x00, 0x23, 0x24, 0x2f}
> , {0x00,0x00,0x00} //never set in log
> , {0x60,0x60,0x60}
> , {0xfa,0xfa,0xfa}
> , {0x00,0x00,0x00} //never set in log
>
> look for this sequence:
> wrote 0x51 <frontend register address>
> wrote 0x3b <high bit of data>
> wrote 0x3a <low bits of data>
>
> those are frontend register writes(at least in your log).
>
> Regarding the Motor struct, this will probably work(without looking at
> your log)
>
> {
> 2400,
> 4800,
> 1,
> {
>   12000,
>   12000,
>   1,
>   1.0
> },
> {
>   12000,
>   12000,
>   1,
>   1.0
> }}
>
> This should lead to very slow movements of the scanning head, but it
> should move if the scanner is correctly initialized(the Canon LiDE
> 35/50/60 are a bit tricky, and the 500F seems to be similar).
>
> As you can see, i use half of the real resolution here. The reason is,
> the backend is not (yet) capable of using something other than 2400,
> 1200 or 600 as sensor resolution. In the meantime, we can make double
> the scanner dimensions, to accommodate for the real 4800dpi.
>
> The gl841 part of the backend also lacks support for non-LiDE color sources.
>
> Hope that helps,
>   Pierre
>


I habe a similar log, not decodes as nicely though.
I try to get the Plustek OpticSlim 2420 with GL842 to work.
I find in the log that the frontend registers are modified many times.
How do I know which values to take for the frontend?
I scanned through the snoop log
(also uploaded here: http://www.physik.tu-berlin.de/~mxhf/OS2420/PreviewScan.rar
SnoopyPro 0.22)
by hand and found
following writes to the frontend registers:

URB#	REG(from 51)	Values (from 3b, 3a zero alway)
------------------------------------------------
Plugging scanner in
...
42 :	00-03		00 23 24 07
...
URB 48 was the last one after plugging the scanner in,
started preview scan.
...
88 : 	00-03		00 23 24 07
...
150 :	00-03		00 27 24 0f
	04-06		00 00 0f
	20-22		80 80 80
	28-2a		4b 4b xx
...	
152 :	28-2a		xx xx 4b
...
192 :   29		4b
	2a		4b
	28		4b
	21		7f
	22		7f
	20		7f
...	
250 :	21		00
	22		00
	20		00
...	
322 :	21		00
	22		00
	20		00
...	
366 :	29		4b
	2a		4b
	28		4b
...	
413 :	29		d2
	2a		c8
	28		c4
...	
417 :	21		05
	22		06
	20		06
...	
451 :	21		00
	22		00
	20		00
...	
459 : 	21		01
	22		01
	20		02
...	
789 :	29		d2
	2a		c8
	28		c4
	
stopped there reading here....



Max



More information about the sane-devel mailing list