Over the last days I've been very busy trying to get my Microtek ScanMaker
X6 to work with Linux / SANE. I think I've seen about every problem that
can possibly occur (according to this mailing list...) -- and now it _is_
actually working. Since I've read nearly every `article' on the list, I
think it is only fair if I describe the solution which appears to get
around the problems (at least with my configuration...).
Just for reference, I'm now using
o the new kernel version 2.2.1,
o an SCSI adapter AVA-1505 (which came with the scanner),
o a Microtek Phantom 636 scanner (Microtek says it's actually a ScanMaker X6),
o the kernel driver aha152x and
o the SANE package sane-pre1.01-3
First of all, the adapter card should have been PnP (it _was_ detected by isapnp), but this
approach didn't work; in fact, the situation was exactly the same as Mr Rogers described in
his article -- see for example
http://chandra.ph1.uni-koeln.de/linux/lg/issue31/rogers2.html
After the kernel had recognized the card, I tried to use an older version of SANE (which
came with S.u.S.E. Linux 5.3). Since it had no support for the new Microtek scanners, this
was doomed to failure...
Consequently, I upgraded to SANE 1.00; the scanner was now (should now have been?)
supported, but I just got to the point where it _should_ have started scanning (preview
scan with xscanimage). Then the whole X server got slowed down and still the scanner
wouldn't budge...
Another upgrade to SANE pre1.01-3 produced (after I manually edited part of the makefiles
to have it find _all_ of my SCSI and JPEG header files) xscanimage started producing
images, but they all had that greenish tint which is also described in some of the
`articles' on this list. The strange thing is: the preview scans were OK, just the `real'
scan was all green (and red -- in fact, only the blue parts appeared to be missing...).
Since I had read about these problems, I then installed on of the `patches' for the
microtek2 backend, namely the file
microtek2-pre0.6.120399.tar.gz
which produced a more consistent result, i.e. _all_ scans (preview included) got green.
Installation of the old microtek2 backend from version 1.00 (just the microtek.? files)
slightly improved the results, but caused one further problem (also described on the
list): the first scan is OK, all further ones remain green.
Now comes the good ending: I replaced the microtek2.? files with those from another patch
microtek2-pre0.6.291298.tar.gz
and this combination, strangely enough, is working _and_ produces good color scans (24
bit), not just one, but as many as I like...
Of course I don't really know _why_ (or how) this combination solves the problems -- being
somewhat of a Linux novice myself -- but maybe my report could help others who experience
related problems. And then, since I've been studying this mailing list rather extensively,
it is only fair to report the success.
Bye, Christian
P.S. If someone can find out why the backend behaves the way it does (especially where the
greenish colors come from), please mail me...
-- Source code, list archive, and docs: http://www.mostang.com/sane/ To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com