plustek: Thread: Genius Vivid configures OK, fails to read lines

Hello,

I'm struggling with a Genius Colorpage Vivid III parallel scanner of
friend of mine.

Under kernel 2.4.16, with precompiled sane 1.0.8-4 from RedHat 7.3 and
with the 0.44.10 of pt_drv, the scanner seems to communicate, using SPP,
but when it gets to actually reading the image data (be it in the
xsane preview mode or during the actual scanning), it fails. I'm
including the relevant (I hope it is) part of my /var/log/messages.
The number of lines that are read successfully is not constant --
something it says ReadOneImageLine() failed at line 57!, sometimes at
line 7, sometimes at line 63.

Once it happened that the preview read the whole preview and displayed
it, so I believe that the scanner is technically OK. But I was never
able to repeat that success. I also fiddled with the timeout (the log
below shows timeout of 10 seconds) and with the _DODELAY in the
imageP98003ReadOneImageLine look, but it doesn't seem to have any
impact.

The scanning head starts moving and stops after a while, before
reaching the other end of the scanner. I also tried to log a message
after each line scanned successfuly -- right after the head starts
moving, the lines received are reported, then the head stops, a couple
more lines gets reported (but that may be a delay between the syslog
and my tail -f), and then it waits for the timeout.

I feel like I might be overlooking something -- but then, the scanner
it recognized and communicates, so it should as well just send me my
picture, right? :-) Infortunately, I run out of ideas.

The machine has 128 MB RAM and has no load on it. But I thought that
the scanner might be sending the data too fast and get its buffers
full or something and just stop sending any more. So is there any way
of telling it that it should send it at slower pace? Some settings of
the parallel port? Or any way ot logging even more information to see
what the status of the scanner is?

I went through the mailing list archives for a couple of months back
but couldn't find description of problem similar to mine. I really
envy you guys who are struggling with stripes in your images -- so far
I have troubles getting any image at all. :-)

I'd appreciate any comment you might have about the situation. Of
course I can upgrade/downgrade anything you might think would help, or
do other tests and send any other logs.

Thank you.

Well, and here's the /var/log/mesages excerpt:

Oct 20 09:48:30 charybda kernel: dacP98003AdjustRGBGain()
Oct 20 09:48:30 charybda kernel: dacP98003AdjustRGBGain() - all loops done!!!
Oct 20 09:48:30 charybda kernel: dacP98003AdjustDark()
Oct 20 09:48:30 charybda kernel: dacP98003AdjustShadingWaveForm()
Oct 20 09:48:43 charybda kernel: DacP98FillShadingDarkToShadingRegister()
Oct 20 09:48:43 charybda kernel: MotorP98003BackToHomeSensor()
Oct 20 09:48:46 charybda kernel: LineCtrl=80, XStepTime=80
Oct 20 09:48:46 charybda kernel: p12SetupScanningCondition()
Oct 20 09:48:46 charybda kernel: P12SetGeneralRegister()
Oct 20 09:48:46 charybda kernel: MinReadFifo=1024, MaxReadFifo=1235
Oct 20 09:48:46 charybda kernel: bRefresh = 2
Oct 20 09:48:46 charybda kernel: RD_ScanControl = 0x15
Oct 20 09:48:46 charybda kernel: RD_Origin = 87, RD_Pixels = 423
Oct 20 09:48:48 charybda kernel: MotorP98003PositionYProc()
Oct 20 09:48:49 charybda kernel: MotorP98003PositionYProc() - done
Oct 20 09:48:49 charybda kernel: dwAppLinesPerArea = 584
Oct 20 09:48:49 charybda kernel: dwAppBytesPerLine = 1269
Oct 20 09:48:49 charybda kernel: dwLinesToRead = 584
Oct 20 09:48:58 charybda kernel: Timeout - Scanner malfunction !!
Oct 20 09:48:58 charybda kernel: Waiting for Sensor to be back in position
Oct 20 09:48:58 charybda kernel: p12InitP98003(0)
Oct 20 09:48:58 charybda kernel: PCB-ID=0x90, CCD-ID=0x01, DAC-TYPE=0x05
Oct 20 09:48:58 charybda kernel: Scanner has 2 Buttons
Oct 20 09:48:58 charybda kernel: Seems we have a Genius Colorpage Vivid III V2
Oct 20 09:48:58 charybda kernel: SAMSUNG 8531
Oct 20 09:48:58 charybda kernel: CCD-3799
Oct 20 09:48:59 charybda kernel: - done !
Oct 20 09:48:59 charybda kernel: ReadOneImageLine() failed at line 57!
Oct 20 09:48:59 charybda kernel: read aborted!
Oct 20 09:49:00 charybda kernel: ioctl(_PTDRV_STOP_SCAN)
Oct 20 09:49:00 charybda kernel: Waiting for Sensor to be back in position
Oct 20 09:49:00 charybda kernel: - done !
Oct 20 09:49:00 charybda kernel: ioctl(_PTDRV_CLOSE_DEVICE)
Oct 20 09:49:00 charybda kernel: *** cleanup buffers ***


-- 
------------------------------------------------------------------------
 Honza Pazdziora | ####@####.#### | http://www.fi.muni.cz/~adelton/
      ... all of these signs saying sorry but we're closed ...
------------------------------------------------------------------------

On Sun, Oct 20, 2002 at 10:39:31AM +0200, Honza Pazdziora wrote:
> 
> Under kernel 2.4.16, with precompiled sane 1.0.8-4 from RedHat 7.3 and
> with the 0.44.10 of pt_drv, the scanner seems to communicate, using SPP,
> but when it gets to actually reading the image data (be it in the
> xsane preview mode or during the actual scanning), it fails. I'm
> including the relevant (I hope it is) part of my /var/log/messages.
> The number of lines that are read successfully is not constant --
> something it says ReadOneImageLine() failed at line 57!, sometimes at
> line 7, sometimes at line 63.

In the meantime I've upgraded to kernel 2.4.19, sane-backends 1.0.9
compiled from src.rpm and pt_drv 0.44.10. Overall the scanner
behaviour is weird. The scanner seems to only send a couple of lines,
where "couple" ranges from 0 to like 120. But never it gets to read
those 584 which is the announced (by DEBUG=all of pt_drv)
dwLinuesToRead and which would be needed to get the whole preview in
xsane.

Also in xsane, if I select a rectangle at the bottom of the scanner,
the scanner never seems to move its head to that part of the page. In
messages, it says MotorP98003PositionYProc() but the head really reads
the lines from the top of the page. If I select a rectangle that is
"low" enough, with only a couple of lines, xsane returns a white image
-- 25 lines the scanner usually gets to read them happily. But from
incorrect part of the page, and the result is pure white, not the
colourfull page I'm trying to test with.

I've also found out that it doesn't matter if I use the SPP or EPP
mode (via forceMode), and that if I enable ECP in BIOS and
/proc/sys/dev/parport/parport0/modes says
PCSPP,TRISTATE,COMPAT,ECP,DMA, I cannot use EPP -- only SPP
(forceMode=0 or 1 gives SPP, according to /var/log/messages) works.

I'd appreciate any suggestion that you might give me to try to get
this Genius Colorpage Vivid III V2 working. It there anybody with
the same model around? What other parameters might be relevant, in the
parport/parport_pc land or otherwise? Is there some other utility that
the high level xsane that could be used to check various functionality
of the scanner?

Thank you,

-- 
------------------------------------------------------------------------
 Honza Pazdziora | ####@####.#### | http://www.fi.muni.cz/~adelton/
      ... all of these signs saying sorry but we're closed ...
------------------------------------------------------------------------