plustek: adding support for additional lm983x based scanners
Subject:
Re: adding support for additional lm983x based scanners
From:
"Allan N. Hessenflow" ####@####.####
Date:
28 Jan 2002 19:05:02 -0000
Message-Id: <20020128190139.A909@kallisti.com>
Hi Gerhard,
> > I'm trying to make a Umax 3420 scanner work with SANE. It is a USB device
> > with an LM9832. I have made it turn the lamp on and off and do some scans
> > (at least a lot of the time; sometimes it doesn't seem to be driving the
> > stepper correctly), but so far the image data, while showing a recognizable
> > shape, is way off. So I'm trying to figure out how the calibration works,
> > and how the sensor gain/offsets are set. It looks like, besides finding
> > the calibration areas for the scanner, I also need to know the CCD type,
> > but do not know how to identify it. I've looked and it is unmarked. It's
> > in a 22 pin DIP, in case that narrows it down any.
>
> In general, it should be possible to add the UMAX 3420 to our device list.
> If this scanner works partially for you, you might have all the necessary
> information to open a new entry in the file plustek_devs.c.
> As the Plustek code contains all of the calibration stuff, there's no need
> to reinvent the wheel again...
>
> As further information source you can have a look at:
> http://scanjet2200c.sourceforge.net/ns-twain/
>
> There are sample sources for the LM983x chipsets and ini stuff for
> the UMAX 3400...
I have an ini file for the 3400; I setup most of the fields for
plustek_devs.c based on that, and made some changes to another file to
tell it how to turn the 3400 lamp on and off (it is controlled by
one of the misc io pins). I'm having two problems with the device.
One is that the stepper does not always work (it just makes noise
without moving the head), so I'd guess that either the speed/acceleration
parameters in the ini file are wrong or the driver isn't setting the
stepper controls correctly (it usually, but not always, works fine for
color scans, and so far has never worked for greyscale scans - the clock
the stepper is controlled by is different in those two modes, so that
makes me suspect the driver isn't setting something correctly, but the
color scans also occasionally failing suggests the ini file numbers aren't
right).
The other is that the image data is wrong. The red channel looks like a
very dark version of the original, and the other channels don't look
related to the original - they contain a bunch of horizontal stripes.
Physically, the scanner has a fairly large white band in it that looks
like it starts at the home position. It also has a thin black band between
that and the scan area; does the calibration function need to be able to
find that band? If so, and it missed, that would certainly cause some
very strange curves. Also, the ini file gives different integration
times for each resolution, while plustek-devs.c only has one pair of
integration times. I haven't looked at that enough to determine which
integration times the driver wants.
allan
--
Allan N. Hessenflow ####@####.####
