[sane-devel] need help with Microtek 35t+ on SPARC/Solaris

Matto Marjanovic maddog@mir.com
Wed, 14 Aug 2002 12:41:11 -0400

 >Several months ago I asked on this list who maintains the microtek
 >backend, and nobody appears to do so.

Hiya... that would be me, and I most probably have that email in the backlog
 somewhere.  I've been busy with other stuff for a while, but recently went
 and touched the code again (and I was out of town for the last two weeks).

And, it won't be until this weekend that I get some time to look at your
 patch and how it fits in with the "microtek world".

 >> Hi, I am having problems making SANE work with a ScanMaker 35t Plus.
 >I have 35t+ with linux and sane 1.0.7 didn't work either (I have not
 >tried with 1.08).

Which version of the microtek driver is included in 1.0.7 (probably the
 same as 1.0.8)?
>From my changelog (http://www.mir.com/mtek/changelog.html), it looks like
 "v0.11.0" should have some success.
(Hmm, I need to update the changelog, too.)

 >However, as coded it couldn't load 10 bit gamma tables. I don't know
 >what is wrong, but I got it "working" by forcing 8 bit gamma tables by
 >modifying the "microtek.c" with the attached diff.
 >HOWEVER! It definitely not correct, I think it results somewhat odd
 >gamma table, but at least it scans!

From the debugging trace in the parent message of this thread, it looks
Status: O

 like perhaps the scanner doesn't really accept a 10-bit gamma table.
 I remember something like this happening with the Scanmaker III.... hmm.

In case you haven't already (actually, if you have, please do it again),
 please send me a debugging log of a scan attempt (as instructed in the
 'sane-microtek' manpage).  Preferably, I'd like to see two, one from
 before your patch, and one from after.

(Note:  Following discussion applies only to the "old" Microtek interface.
        I have no idea whether the "new" interface (handled by the microtek2
        backend) is easier to work with.)

 >I could FIX things, if I had the 35t+ programming info. Has anyone
 >actually managed to get those from Microtek?

I applaud your youthful vigor --- but the "programming info" won't help
 you fix anything.  I have their scanner interface specification, a couple
 of different versions of it (none of which is specific to the 35T+). 
 Microtek used to have it online at one or more of their FTP sites.  

None of their scanners completely follow the spec.  Each scanner has some
 exceptions or restrictions or extra stuff.  For examples, the E6 uses
 undocumented commands to deal with calibration --- this is handled by
 the backend, after I reverse-engineered it with the help of another user.
 The same scanner models sometimes use different firmware revisions which
 also require different handling.

At this point, the backend pretty faithfully implements the spec, along
 with many, many special cases to handle points at which the various 
 scanners in this old line appear to diverge.  So, the only way to make
 progress is through reverse-engineering.

(But, I'll see if I still have the digital copy of the spec somewhere if
 you still want to read it.)

 >I'm fairly pissed off by these scanner makers hiding the info. Is
 >there *ANY* scanner maker that gives programming info? If so, my next
 >scanner will definitely be of that brand!!!! (I won't buy anything
 >from microtek any more).

Yeah, I agree with you about being pissed.  But, like I said above, 
 perhaps the newer line of Microtek scanners aren't so bad.  As far as
 I do know, the interface for that line is supposed to follow as SCSI-2
 spec for scanners....

-matt m.