Re: sane-0.65 released

Stefan Luettringhaus-Kappel (stefan@uran.informatik.uni-bonn.de)
Sun, 5 Oct 1997 20:09:57 +0200 (MEST)

On Sun, 5 Oct 1997 becka@rz.uni-duesseldorf.de wrote:

> Hi !
>
> > has there been any progress with respect to the MFS-12000SP problems which
> > were investigated in David's survey? Last version I tried was 0.62.
> We have tried to investigate it, but need more test-data.
>
> The problem with the Lineart mode hanging seems to be timing related and there
> is a hack for it by M. Ritzert. (R U reading ? Maybe you should send it to the
> list).
>
> The other problem with the occasional hangs occuring for some parameter sets
> is not yet solved, though I have a faint idea about the cause.
>

Well, I never had any hangs. My problem, which was reported by others as
well, is missing scan lines in color and gray mode. I'll try any patches,
but I really have no time to have a deeper look into the sources. I include
my answer to David's survey, just in case.

Regards,
Stefan

-- 
Dr. Stefan Lüttringhaus-Kappel  <><><><><><><><><><><><><><><><><><><><><>
Institut für Informatik III     Phone +49 (228) 73-4536
Universität Bonn                Fax   +49 (228) 73-4382
Römerstraße 164                 stefan@uran.informatik.uni-bonn.de
D-53117 Bonn (Germany)          http://www.informatik.uni-bonn.de/~stefan/
<><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><><>

8< ------------------------------- (snip)

- What version of SANE are you using?

0.62 (similar problems with previous versions)

- How was SANE configured? Did you use the --disable-ld-fix option?

no, --disable-ld-fix makes ugly colored shadows everywhere in the image.

Also, I have to comment out ... dev->flags |= MUSTEK_FLAG_LD_MFS although my scanner is only version 1.00. Otherwise the 360dpi image gets single reddish lines at scsi buffer boundaries.

- What SCSI adapter are you using?

Asus SC-200 (NCR53C810), the scanner is the only device on this host:

stefan@piano ~: cat /proc/scsi/ncr53c8xx/1 General information: Chip NCR53C810, device id 0x1, revision id 0x1 IO port address 0xe400, IRQ number 11 Using memory mapped IO at virtual address 0x4843000 Synchronous period factor 25, max commands per lun 4 Profiling information: num_trans = 1 num_kbytes = 0 num_disc = 0 num_break = 1 num_int = 18 num_fly = 2 ms_setup = 0 ms_data = 10 ms_disc = 0 ms_post = 0

- Does grayscale scanning work fine at various resolutions and without dropped lines (e.g., try to scan a diagonal line to detect dropped lines). If not, does it work for certain resolutions or do other settings (such as "force backtracking") affect the result?

No, it doesn't work fine. several scan lines are dropped at intervals corresponding to the scsi buffer size, both at 32k and 128k. This happens at all resolution, both below and above 300 dpi. The Rest of the image is fine (apart from the missing lines).

(By the way, using 128k instead of 127.5k never resulted in any problems, neither with the scanner, nor scsi disks, cd-rom nor dat streamer.)

AND, STRANGELY (?): scanning the same area takes 30secs at 300dpi, but 84 secs at 302dpi. Something is fundamentally different at > 300dpi.

- Does color scanning work? If not, does it work for certain resolutions? Describe the kind of errors you see.

Quite similar to Gray, but here a range of lines is repeated. I think I had missing lines, too, but I can't reproduce it now. At the boundaries color error appear in addition to the extra lines.

Errors occur at 300dpi and below but not above; at 302dpi and 360dpi, for example, everything is fine. But it again takes much longer.

- Include the output of command: SANE_DEBUG_MUSTEK=3 scanimage -h -d mustek >/dev/null

[sanei_init_debug]: Setting debug level of mustek to 3. [mustek] sane_init: strip-height set to 1 inches [mustek] attach: opening /dev/scanner [mustek] attach: sending INQUIRY [mustek] wait_ready: sending TEST_UNIT_READY [mustek] attach: firmware revision 1.00 [mustek] attach: found Mustek scanner model MFS-12000SP (flatbed scanner), 1-pass

--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com