[sane-devel] Some testing of HP Scanjet 2400 scanner (genesys backend)

stef stef.dev at free.fr
Tue May 10 06:37:18 UTC 2011

Le vendredi 6 mai 2011 06:35:04 Alexey Osipov, vous avez écrit :
> Hi, dear developers of SANE!
> I have `HP Scanjet 2400` scanner. Serial NO: CN648SR2H7. Regulatory
> Model: GRLYB-0209.
> $ sane-find-scanner
> found USB scanner (vendor=0x03f0 [Hewlett-Packard], product=0x0a01 [hp
> scanjet scanner], chip=GL646_HP) at libusb:004:002
> I have made some tests of my scanner under brand new Ubuntu 11.04, which
> seems to have latest version of sane-backends (1.0.22).
> I was using `scanimage` program to do tests like this:
> $ scanimage --mode <mode> --resolution <resolution>  > ~/Desktop/image.pnm
> where <mode> and <resolution> varies in ranges:
> <mode>: Color, Gray, Lineart
> <resolution>: 50, 100, 150, 300, 600, 1200
> Results are:
> 50dpi:
> output is garbage in all modes (color example: image50.tiff)
> 100 dpi:
> all modes works good (color example: image100_c.tiff)
> 150dpi:
> all modes: result image squashed vertically, scan head run beyond
> allowed scan borders with awful sounds, then surprisingly correctly
> returns back. (color result example: image150.tiff)
> 300 dpi:
> all modes works good (examples: image300_c.tiff, image300_g.tiff,
> image300_la.tiff)
> 600 dpi:
> `Color` mode: scanner hangups after few movements. Note that first
> couple of this movements are happens with 'correct' sound and the
> second couple - with slightly different sound, which seems to me not
> quite 'correct' (it's like screwing something). But this is definitely
> NOT the sound when head moves beyond allowed borders, that's different.
> If in this 'hangup' situation I hit "ctrl+c", the head correctly moves
> in it's parking position and I can do other scans in other modes just
> fine.
> `Gray` and `Lineart` modes: They don't hang, scanner head goes through
> all area and then correctly returns back. But the result isn't correct -
> there are two images squashed horizontally (see examples:
> image600_g.tiff, image600_la.tiff).
> 1200dpi:
> In all modes scanner just hangups, with no movements and sounds.
> "ctrl+c" not working, next scans failed with "No SANE devices found"
> until scanner reconnect.
> In 100 and 300 dpi modes, if I hit "ctrl+c" during scanning, scanner
> stops and then hangup. After some time scanimage exits with error
> message like "Error while scanner I/O". Next scan attempts failed with
> message "No SANE devices found" until scanner reconnect.
> `--preview=yes` switch doesn't change anything in all modes and
> resolutions.
> All examples archived here:
> http://lion-simba.github.com/scanjet2400_samples.7z
> I hope this info helps you guys finish support of this scanner and
> remove "Untested" mark from it. :)
> If you need some additional feedback, feel free to ask me. I'm a
> programmer so I have enough skills to apply patches, compiling, etc.
> Best regards,
> Alexey Osipov.
> PS. I think it's not quite good to require mail-list subscription prior
> to allow posting messages to it. ESPECIALLY in the list, to which users
> supposed to post feedback. Note the output of scanimage:
> $ scanimage
> [genesys] WARNING: Your scanner is not fully supported or at least
> [genesys]          had only limited testing. Please be careful and
> [genesys]          report any failure/success to
> [genesys]          sane-devel at lists.alioth.debian.org. Please provide as
> many [genesys]          details as possible, e.g. the exact name of your
> [genesys]          scanner and what does (not) work.

	HP2400/G2410 support completion comes quite often. What is needed is test 
and tweaking of the values defined in genesys_gl646.h for the 
CCD_HP2400/MOTOR_HP2400 in the mode table definitions. For instance the 
vertical image distortion and head going too far may be due to using single 
step instead of half step or half instaed of quater.
	The process should be run a test with full debug log like :

scanimage -d genesys --resolution 50 --mode Color 2>try.log >try.pnm

	then compare the values (mostly register settings and slope tables) with 
logs recorded under windows, and fix what's needed. This process requires you 
have a device to test with. There isn't much code (if any) to touch.

	I repost the scripts I use to decode UBS logs (done with usbsnoop from 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: scripts.tar.gz
Type: application/x-compressed-tar
Size: 14589 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/sane-devel/attachments/20110510/2678674e/attachment-0001.bin>

More information about the sane-devel mailing list