canonscanner-devel Mailing List for Canon scanner backend (Page 3)
Status: Alpha
Brought to you by:
volodya
Menu
▾
▴
canonscanner-devel
You can subscribe to this list here.
| 2001 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(6) |
Jun
(3) |
Jul
(2) |
Aug
(1) |
Sep
(2) |
Oct
(23) |
Nov
|
Dec
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2002 |
Jan
|
Feb
(4) |
Mar
(9) |
Apr
(13) |
May
(16) |
Jun
|
Jul
(3) |
Aug
(2) |
Sep
(26) |
Oct
(17) |
Nov
(7) |
Dec
(1) |
| 2003 |
Jan
|
Feb
(1) |
Mar
|
Apr
|
May
|
Jun
|
Jul
(2) |
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
|
From: <ki...@kh...> - 2002-09-11 23:34:23
|
Christian, I am no expert on programming, but I do have a Canon N670U, and I remain interested in getting its support beyond the alpha stage. I have downloaded the new stuff from Johann Philipp just now, but I did not find your patch there, at least as far as I can see. I wonder if you would be willing to share this? Theodore Kilgore On Wed, 11 Sep 2002, Theodore Kilgore wrote: > > > ---------- Forwarded message ---------- > Date: Wed, 11 Sep 2002 11:49:18 +0200 > From: "Christian [iso-8859-15] R=F6=DFner" <c.r...@ro...= > > To: CanonScanner <can...@li...> > Subject: [Canonscanner-devel] CanoScan N670U - minor problems > > Hi, > > after some discussions with Johann Philipp, I got his SANE-API working fo= r the > CanoScan N670U. > > I sent him a diff file (canon_usb.c.diff), which contains a different _IO= R > call and a expanded switch/case construct in function sane_get_devices(..= =2E) > > After his last email to me, I changed every explicit dpi resolution from = 1200 > into 600 and changed all HZ_RES_DIVER_x to half of its values > (HZ_RES_DIVIDER_4 =3D> HZ_RES_DIVIDER_2). These changes were made in > canon_usb_mid.c > > There are still some problems: > > a) > The previewscan only scans DIN B5 in y direction and about 20cm in x > direction, although the scanarea was chosen "full area" in xsane-0.87. > > b) > If I choose an area from the previewed picture, the scanned picture often= does > not show that part after scanning. > > c) > Sometimes the scanned picture looks phasing-shifted. > > I think, these problems occur, because I took the canon_usb_mid.c from N1= 240U > and have not changed everthing to the N670U model, yet. > > I am not very familar to the code and the sane-api, so I do not really kn= ow > where to look for. Maybe Johann Philipp (hoffentlich liest Du das hier ;-= ) ) > has still some ideas for me. > > Any ideas? > > Regards > > Christian > > > > ------------------------------------------------------- > In remembrance > www.osdn.com/911/ > _______________________________________________ > Canonscanner-devel mailing list > Can...@li... > https://lists.sourceforge.net/lists/listinfo/canonscanner-devel > |
|
From: Christian <c.r...@ro...> - 2002-09-11 18:45:12
|
Hi, I got the values for the x and y range parameters for CanoScan N670U These are: 8.48 and 12.00 Preview seems to be alright now. What are MINX, MINY, MAXX and MAXY? pixel, mm, inches, per what. What do they mean? I hope to find answers for the y range problem to a selected frame if I k= now=20 what these "really constant" values are standing for. Thanks. An other thing: Phasing-shifted pics seem to have problems with odd x ranges. I can not s= ay if=20 it is the pixel-value or the cm-value? As my cousine said (he wrote a scs= i=20 scanner driver), it has to do with the fact, that x-ranges need to be=20 divideable by 2. I do not know, if that is right. Maybe it helps you. I d= o=20 not think that I can figure that out. Please help me with the above min, max-stuff Thanx Christian |
|
From: Christian <c.r...@ro...> - 2002-09-11 09:48:54
|
Hi, after some discussions with Johann Philipp, I got his SANE-API working fo= r the=20 CanoScan N670U. I sent him a diff file (canon_usb.c.diff), which contains a different _IO= R=20 call and a expanded switch/case construct in function sane_get_devices(..= =2E) After his last email to me, I changed every explicit dpi resolution from = 1200=20 into 600 and changed all HZ_RES_DIVER_x to half of its values=20 (HZ_RES_DIVIDER_4 =3D> HZ_RES_DIVIDER_2). These changes were made in=20 canon_usb_mid.c There are still some problems: a) The previewscan only scans DIN B5 in y direction and about 20cm in x=20 direction, although the scanarea was chosen "full area" in xsane-0.87. b) If I choose an area from the previewed picture, the scanned picture often= does=20 not show that part after scanning. c) Sometimes the scanned picture looks phasing-shifted. I think, these problems occur, because I took the canon_usb_mid.c from N1= 240U=20 and have not changed everthing to the N670U model, yet. I am not very familar to the code and the sane-api, so I do not really kn= ow=20 where to look for. Maybe Johann Philipp (hoffentlich liest Du das hier ;-= ) )=20 has still some ideas for me. Any ideas? Regards Christian |
|
From: Johann.Philipp <joh...@we...> - 2002-09-08 21:41:24
|
Hi guys, good news: I was able to knit a sane backend for my N1240U. It works fairly well and is basically my testprogram ct_N1240U.c with the SANE-API. I hope this will be a good starting point for merging our knowledge on the different models. Also I hope for some feedback... Just have a look at the cvs! Have a good week! Johann |
|
From: Vladimir D. <vo...@mi...> - 2002-08-20 06:08:09
|
Try looking in national datasheets - this should help with turning the
light on.
Vladimir Dergachev
On Sun, 18 Aug 2002, Andy U wrote:
> Hi everyone,
>
> the test program for ct1220u is working, scanner is moving etc., but the
> light is not turned on - so the resulting picture is completely black.
> Someone else posted this problem - does anybody got any idea?
>
> Is it a different register for turning the light on, or a different value?
> I tried to set all other bits in register 0x29, but that did not work ;-(
> I think I do not really want to snoop the usb bus.
>
>
>
> What I did? I just took ct_1220U.c from the cvs and changed the
> device=0x2208 (plus the height, the D660U seems to interpret the height
> differently, height=4 was too much for it, 3 is better to avoid hitting
> the end of the scan bed), compiled, worked. It's just a little to dark ...
>
> Thanks, Andy
>
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by: OSDN - Tired of that same old
> cell phone? Get a new here for FREE!
> https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390
> _______________________________________________
> Canonscanner-devel mailing list
> Can...@li...
> https://lists.sourceforge.net/lists/listinfo/canonscanner-devel
>
|
|
From: Andy U <an...@uh...> - 2002-08-18 12:51:51
|
Hi everyone, the test program for ct1220u is working, scanner is moving etc., but the light is not turned on - so the resulting picture is completely black. Someone else posted this problem - does anybody got any idea? Is it a different register for turning the light on, or a different value? I tried to set all other bits in register 0x29, but that did not work ;-( I think I do not really want to snoop the usb bus. What I did? I just took ct_1220U.c from the cvs and changed the device=0x2208 (plus the height, the D660U seems to interpret the height differently, height=4 was too much for it, 3 is better to avoid hitting the end of the scan bed), compiled, worked. It's just a little to dark ... Thanks, Andy |
|
From: Vladimir D. <vo...@mi...> - 2002-07-14 22:25:46
|
On Fri, 12 Jul 2002, [ISO-8859-1] St=E9phane wrote:
> I've got a problem with a N670U ?
> usbview found it but it sane doesn't.
> I'm in trouble?
Try asking on sane mailing list.. But possibly this scanner is not
supported yet.
Vladimir Dergachev
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Gadgets, caffeine, t-shirts, fun stuff.
> http://thinkgeek.com/sf
> _______________________________________________
> Canonscanner-devel mailing list
> Can...@li...
> https://lists.sourceforge.net/lists/listinfo/canonscanner-devel
>
|
|
From: <ste...@cl...> - 2002-07-12 20:40:27
|
I've got a problem with a N670U ? usbview found it but it sane doesn't. I'm in trouble? |
|
From: Johann.Philipp <joh...@we...> - 2002-07-04 19:21:22
|
Dear Stephan, the answer to your question in the subject is: my N1240 is working under Linux. But: Unfortunately I haven't been able to work on my program for some weeks now, and I didn't publish it yet. The reason is: I went a quite dangerous way in coding the functionality, by formulating the way I think the scanner might work into C-code. The reason I did that is that I thought the way cannon programmed their drivers is quite messy, and therefore not easy to understand. What I can do is to comment some more what I have done so far (actually my program scans some lines...) and publish it this weekend. I hope to have some time for programming again in August. To come to your questions: All NxyzU scanners seem to be very similar, having more or less the same chip inside. But unfortunately a program for the N1220U will most probably do nothing or something completely wrong with a N1240U. Some of the register settings are very sensitive. To communicate in any way with the scanner was for me no real problem. I have a SuSE 7.3 system. All I need to do is to load the scanner.o kernel module via "modprobe scanner" and start my program. The file /etc/modules.conf should have a line "options scanner product=0x220e vendor=0x04a9" But my experience is, that if something failed in a previous run, the scanner (or the kernel...?) can be in a very confused state. Usually this is solvable by disconnecting the scanner cable (and thereby taking away the power), unloading the kernel module, replugging the scanner and finally reloading the kernel module. I'm not sure, but I think there were occasions, when I had to restart my whole machine. In any case, be prepared to unplug the scanner on unnatural noises while testing. Sane doesn't help, even, if there are some usb scanners supported. I hope all this helped more than it did confuse... Johann ------------------------------------------------------- > Stephan Rudlof wrote: Dear Johann, I have good news: accessing the scanner works now. It was just doing /dev/MAKEDEV usb . But now the bad ones: the initialization fails with attached log.txt (interrupted) and many NAKs. I have found this after playing with the program *after* a failed initialization (so this is not so interesting anymore): ##################### karl:~# /var/local/sane/canon/canon-test-1240U Starting conversation with the scanner Statistics: sizeof(init_data1)=2552 Optical resolution: 300x300 dpi Vendor id=0x04a9 Product id=0x220e CanonScan 1240U detected. Assuming the following hardware parameters: memory_unit_size=2048 bytes total_memory=524288 bytes max_height=11 inches max_width=8.5 inches Write warning: 4 bytes requested, -1 written REG_VERSION returned unexpected value ----- In a root term without X there is slightly more: ----- karl:~# /var/local/sane/canon/canon-test-1240U Starting conversation with the scanner Statistics: sizeof(init_data1)=2552 Optical resolution: 300x300 dpi Vendor id=0x04a9 Product id=0x220e CanonScan 1240U detected. Assuming the following hardware parameters: memory_unit_size=2048 bytes total_memory=524288 bytes max_height=11 inches max_width=8.5 inches [c97be0c0] link (097be092) element (097bf210) 0: [c97bf210] link (00000001) e0 SPD IOC Stalled CRC/Timeo Length=3 MaxLen=3 DT0 EndPt=3 Dev=2, PID=e1(OUT) (buf=098db000) scanner.c: write_scanner: NAK received. Write warning: 4 bytes requested, -1 written [c97be0c0] link (097be092) element (097bf210) 0: [c97bf210] link (00000001) e0 SPD IOC Stalled CRC/Timeo Length=7ff MaxLen=0 DT1 EndPt=2 Dev=2, PID=69(IN) (buf=097b0000) scanner.c: read_scanner(0): funky result:-84. Consult Documentation/usb/scanner.txt. REG_VERSION returned unexpected value Note: commenting out return(-1); where REG_VERSION is wrong, doesn't help: send_cmd called by reg_r produces many NAKs (whatever they are). ##################### My ct_1240U.c is attached, too. Greetings, Stephan Stephan Rudlof wrote: > > Dear Johann, > > I have problems to get my N1240U scanner working. > > Do you know if it is possible at all with a slightly modified > canonscanner/canon/ct_1220U.c file? > > I have taken this approach to just change the IDs here (as described in the > Readmes); > moreover I have extended the drivers/usb/scanner.h from my 2.4.18 kernel > accordingly. > > Some installed modules are (by lsmod) > > usbcore 48128 1 [uhci scanner usb-storage] > . > > My current errror is > karl:/home/sr# /var/local/sane/canon/canon-test-1240U > Unable to open scanner device: No such device > ; but this makes sense, since no driver creates an /dev/usbscanner or > similar (Should it?). I have tried it with links from > /dev/usbscanner > to > /dev/sg0 > or > /dev/sg1 > , but this hasn't been sufficient. > > The command sane-find-scanners doesn't find anything. > And USB debug messages don't give hints to me. > > My devices: > > karl:~# cat /proc/devices > Character devices: > 1 mem > 2 pty > 3 ttyp > 4 ttyS > 5 cua > 6 lp > 7 vcs > 10 misc > 14 sound > 108 ppp > 128 ptm > 136 pts > 162 raw > 180 usb > 226 drm > 254 pcmcia > > Block devices: > 2 fd > 3 ide0 > 7 loop > > My Linux (Debian): > > karl:~# uname -a > Linux karl 2.4.18 #6 Tue Jul 2 19:32:23 CEST 2002 i686 unknown > > Any hint is appreciated; > > Greetings, > > Stephan > -- > Stephan Rudlof (sr...@ev...) > "Genius doesn't work on an assembly line basis. > You can't simply say, 'Today I will be brilliant.'" > -- Kirk, "The Ultimate Computer", stardate 4731.3 -- Stephan Rudlof (sr...@ev...) "Genius doesn't work on an assembly line basis. You can't simply say, 'Today I will be brilliant.'" -- Kirk, "The Ultimate Computer", stardate 4731.3 canon-test-1240U log.txt |
|
From: Gunther M. <gun...@gm...> - 2002-05-15 16:50:27
|
Johann Philipp wrote: > Hi everyone, > > I haven't been online quite a time, but the last few days I was able to > continue work a bit. Here is what I think: > > a) hardware stuff > > exerpts from Theodore: > > I understand that the N1240U has a National Semiconductor LM9833 in it, as > > does the N670U. But this is hearsay. It _has_ lm9833. > > > > You see, I really want to know how the scanner is working. > Johann > Then try http://home.t-online.de/home/0531896113/merlin670-V0.02.c which shows the scanner working since 3 months. Actually this was much easier than a SANE backend. - Gunther |
|
From: Vladimir D. <vo...@mi...> - 2002-05-15 05:52:46
|
On Sun, 12 May 2002, Johann Philipp wrote:
> Hi everyone,
>
> I haven't been online quite a time, but the last few days I was able to
> continue work a bit. Here is what I think:
>
> a) hardware stuff
>
> exerpts from Theodore:
> > I understand that the N1240U has a National Semiconductor LM9833 in it, as
> > does the N670U. But this is hearsay.
> > ...
> > For those who may not remember my last posting, several months ago, I have
> > a Canon N670U. It comes with a CD and a manual which discuss the Windows
> > setup for both the N670U and the N1240U. Thus I would conjecture that they
> > have the same chip inside, with only some minor differences between the
> > models.
> I agree on all of that, adding, that my win driver CD mentions also the N676U.
> But there are a few open points:
> - In my snoopy logs I found that register 0x70 is filled with the value 0x73
> at some points and register 0x76 with value 0x08. Either Canon is using
> undocumented features, they misuse those registers, or there is not a LM9833
> in my scanner. I tend to ignore all that...
> - when I took my scanner apart the chip was too badly hidden, the sensor's
> only identifying number led to a plastic producer (for sheltering the sensor
> and the LED). The only valuable thing found was a TMTECH-Chip, which most
> probably is the 4MBit DRAM Buffer (TMTECH's coding policy is not really
> strict). Therefore :-(
> - according to the documentation I tried to reverse calculate the FSPI (Full
> Steps per Inch) resulting in 400. If I consider that 4 microsteps are 1
> fullstep, I come up with 1600 dpi vertical hardware resolution. This is quite
> strange for a 2400 dpi scanner, but I tend to ignore that, too...
> - I have no clue what the "default phase difference", regs 0x51 to 0x53 could
> possibly be. Can anyone explain?
>
> b) software stuff
>
> - Vladimir, I took your lm9831.h, added some more register #defines and
> commited as lm9833.h. Apart from two minor changes (reg 0x09, datamode can be
> 16bit for the lm9833 instead of 14bit; reg 0x51, I took two definitions for
> the two functions) I tried to shorten names by removing every "REG_VALUE_"
> Prefix. This is just a suggestion...
This is a bad idea. The reason for REG_VALUE prefix is that the defines
codes the value that can be written or read into corresponding register.
This avoids potential name clash. If you do not want to write it too
often, try using some convenient #define macro for you (one that uses ##
for example).
Also, the lm9831.h file is actually for lm9832 as this is the chip I have.
Additionally I see only minor differences between 9832 and 9833 so it
would probably be better to merge them at a later date - though no hurry
to do this now.
Vladimir Dergachev
> - I'm trying to rewrite my test program, so that it is partitioned in small
> functional blocks (e.g. setup paper_sense and misc_io). I think my snoopy
> logs show a lot of unneccessary repetition, so just reimplementing would give
> a quite messy code. Also I think it will be easier to actually make a SANE
> backend.
>
> You see, I really want to know how the scanner is working.
> Johann
>
> _______________________________________________________________
>
> Have big pipes? SourceForge.net is looking for download mirrors. We supply
> the hardware. You get the recognition. Email Us: ban...@so...
> _______________________________________________
> Canonscanner-devel mailing list
> Can...@li...
> https://lists.sourceforge.net/lists/listinfo/canonscanner-devel
>
|
|
From: Johann P. <jo...@ka...> - 2002-05-11 23:55:26
|
Hi everyone, I haven't been online quite a time, but the last few days I was able to continue work a bit. Here is what I think: a) hardware stuff exerpts from Theodore: > I understand that the N1240U has a National Semiconductor LM9833 in it, as > does the N670U. But this is hearsay. > ... > For those who may not remember my last posting, several months ago, I have > a Canon N670U. It comes with a CD and a manual which discuss the Windows > setup for both the N670U and the N1240U. Thus I would conjecture that they > have the same chip inside, with only some minor differences between the > models. I agree on all of that, adding, that my win driver CD mentions also the N676U. But there are a few open points: - In my snoopy logs I found that register 0x70 is filled with the value 0x73 at some points and register 0x76 with value 0x08. Either Canon is using undocumented features, they misuse those registers, or there is not a LM9833 in my scanner. I tend to ignore all that... - when I took my scanner apart the chip was too badly hidden, the sensor's only identifying number led to a plastic producer (for sheltering the sensor and the LED). The only valuable thing found was a TMTECH-Chip, which most probably is the 4MBit DRAM Buffer (TMTECH's coding policy is not really strict). Therefore :-( - according to the documentation I tried to reverse calculate the FSPI (Full Steps per Inch) resulting in 400. If I consider that 4 microsteps are 1 fullstep, I come up with 1600 dpi vertical hardware resolution. This is quite strange for a 2400 dpi scanner, but I tend to ignore that, too... - I have no clue what the "default phase difference", regs 0x51 to 0x53 could possibly be. Can anyone explain? b) software stuff - Vladimir, I took your lm9831.h, added some more register #defines and commited as lm9833.h. Apart from two minor changes (reg 0x09, datamode can be 16bit for the lm9833 instead of 14bit; reg 0x51, I took two definitions for the two functions) I tried to shorten names by removing every "REG_VALUE_" Prefix. This is just a suggestion... - I'm trying to rewrite my test program, so that it is partitioned in small functional blocks (e.g. setup paper_sense and misc_io). I think my snoopy logs show a lot of unneccessary repetition, so just reimplementing would give a quite messy code. Also I think it will be easier to actually make a SANE backend. You see, I really want to know how the scanner is working. Johann |
|
From: Theodore K. <ki...@ba...> - 2002-05-07 19:20:08
|
Vladimir, Finally I got around to checking this out: For those who may not remember my last posting, several months ago, I have a Canon N670U. It comes with a CD and a manual which discuss the Windows setup for both the N670U and the N1240U. Thus I would conjecture that they have the same chip inside, with only some minor differences between the models. The N670U could be "accessed" by the ct_1220U driver, as of a few months ago, with some minor changes introduced in the source file by my very amateur self. The program would then "run" and would pretend to do a scan, and exit when it "finished." The only problem was that there was no light and no sound (no physical movenent of the scanner head). So in fact nothing really happened. Update is that with one minor change (changing the model number from 2207 to 220d as required by the 670U) and changing the location of the scanner from /dev/usbscanner to /dev/usb/scanner0 (needed because I run devfs) there was essentially no change in status. The program started and pretended to scan. But no light and no sound and no movement of scanner head. There was one minor difference. It wanted to continue to "scan" forever. I had to hit ^C to get it to quit. I can also report as I have done before, that Gerhard Mayer's program (based upon libusb) does already work with this scanner. I continue to hope for the day that this scanner will be incorporated into the SANE package, and I am convinced it is not far off now. Theodore Kilgore On Mon, 29 Apr 2002, Vladimir Dergachev wrote: > > > On Sun, 28 Apr 2002, Theodore Kilgore wrote: > > > I understand that the N1240U has a National Semiconductor LM9833 in it, as > > does the N670U. But this is hearsay. > > > > I've made some cleanups to ct1220U.c file - I believe it would be > interesting for you to take a look at it. Besides everything else, it now > identifies the scanner chip.. > > Vladimir Dergachev > > > Theodore Kilgore > > > > On Sat, 27 Apr 2002, Johann Philipp wrote: > > > > > The referenced document is just great, so I tried to find out, what chip is > > > in my N1240U. Unfortunately the chip is hidden by a metal casing which is > > > soldered to the board. Therefore I cannot recommend to disassemble funktional > > > scanners, that are of the flat type as mine. > > > > > > Johann > > > > > > > I think this might be of use to the more active developers ;) > > > > > > > > Vladimir Dergachev > > > > > > > > > > > > ---------- Forwarded message ---------- > > > > Date: Sun, 21 Apr 2002 16:30:20 +0200 (CEST) > > > > From: Jens Osterkamp <je...@os...> > > > > To: vo...@us..., kz...@us... > > > > Subject: Canon N650U > > > > > > > > > > > > Hi, > > > > > > > > I tried your driver with my Canon N650U some time ago and it worked, at > > > > least I was able to scan a picture in poor quality. > > > > Unfortunately my scanner is broken now so I opened it and saw that there > > > > is a LM9832CCVJD by National Semiconductor on the pcb. Looking this up > > > > i found this page with a datasheet and descriptions of all registers. > > > > Please look here : http://www.national.com/pf/LM/LM9832.html. > > > > > > > > Hope this helps. I would be glad to here from you. > > > > > > > > Jens > > > > > > > > > > > > _______________________________________________ > > > > Canonscanner-devel mailing list > > > > Can...@li... > > > > https://lists.sourceforge.net/lists/listinfo/canonscanner-devel > > > > > > _______________________________________________ > > > Canonscanner-devel mailing list > > > Can...@li... > > > https://lists.sourceforge.net/lists/listinfo/canonscanner-devel > > > > > > > > > _______________________________________________ > > Canonscanner-devel mailing list > > Can...@li... > > https://lists.sourceforge.net/lists/listinfo/canonscanner-devel > > > > > _______________________________________________ > Canonscanner-devel mailing list > Can...@li... > https://lists.sourceforge.net/lists/listinfo/canonscanner-devel > |
|
From: R N d. <ran...@ya...> - 2002-05-05 19:45:05
|
Hi Jason, in a previosu mail i've attached the screenshot of the problems i've seen, i've to try the module version yet, i'll do ASAP, so i can give you this information. thanks for you help and waiting for any news. Angelo --- Jason Millward <mil...@nt...> wrote: > Hi Angelo > > I havent upgraded my kernel to this version yet, > when I do I'll also upgrade gcc and sort this out. I > assume that the module version still works, is this > the case ? Theres alot of problems with the libusb > version that I need to iron out as soon as this is > done I'll post. > > Jason > __________________________________________________ Do You Yahoo!? Yahoo! Health - your guide to health and wellness http://health.yahoo.com |
|
From: Giuseppe S. <epp...@us...> - 2002-05-04 13:53:48
|
Well, while I am fighting with WindowsME, you might have a look at an old log that I made before upgrading Windows :-( It should be available in two formats at http://eppesuigoccas.homedns.org/canonD646U.preview.log 2.3Mb http://eppesuigoccas.homedns.org/canonD646U.preview.log.gz 0.4Mb Bye, Giuseppe |
|
From: Giuseppe S. <epp...@us...> - 2002-05-04 12:53:59
|
Hi Jason, I would have no problem about uploading such a file. My problem is that I never got a snoop program running on my Windows ME :-( BTW, I plugged the scanner on a Linux machine that is accessible via Internet. If your think that this could help, I may let you/other access via SSH. Bye, Giuseppe |
|
From: Jason M. <mil...@nt...> - 2002-05-03 20:53:00
|
Hi Vladimir I know what you mean it is rather large, I will watch your progress = avidly (hope you don't mind, I'm finding Sane rather daunting at the = moment). Just in case you haven't seen it there is a backend for the = scanjet 2200C (another LM983X scanner) which may help and is set out = nicely(at least it was the last time I looked) by Ludo Stellingwerff = available from www.xs4all.nl/~stwerff=20 Sorry if you already know this stuff. Jason |
|
From: Vladimir D. <vo...@mi...> - 2002-05-03 17:54:38
|
On Fri, 3 May 2002, Jason Millward wrote:
> Hi Vladimir
>
> I see your ready to make a Sane Backend, have you looked at the Plustek b=
ackend yet. This is the backend I'm working with to get the N650U working w=
ith Sane. The Backend already covers the LM983X chip and you can easily(sup=
posedly) add other devices that use this chip to the code. There is a test =
version that Gerhard (the writer) did for the N650U (although it dosent wor=
k at the minute, which I'm supposedly sorting out) that should work just as=
well for the N1220U. The backend covers all the calculations from the chip=
specification manual, and also includes calibration. The only problems are=
that the backend is huge, and it dosent follow the procedure that the cano=
n windows driver follows (it follows the national semiconductor windows dri=
ver procedure), another problem is that the test driver is the first (untes=
ted) version to cover CIS scanners. The advatage of this backend though is =
once the CIS side is implemented any Canon USB scanner should be easily imp=
lementable. Hopefully this will help, If you're interested in taking this p=
ath get in touch (I could do with some help after all two heads better than=
one).
>
I looked now.. I am not touching the monster. Actually all I wanted to do
is to make a backend that works with my scanner and lets the user adjust
hardware parameters (like gain, offset and gamma) and has fixed DPI
settings. IMO, rescaling is best done with gimp, why fake DPI that does
not exist ?
Take a look at the CVS - there is a working test program for 1220U,
perhaps you can modify it to work with your scanner. Once you have code
that works everything becomes much simpler.
Vladimir Dergachev
> Jason
>
>
>
|
|
From: Jason M. <mil...@nt...> - 2002-05-03 14:30:52
|
Hi Angelo I havent upgraded my kernel to this version yet, when I do I'll also = upgrade gcc and sort this out. I assume that the module version still = works, is this the case ? Theres alot of problems with the libusb = version that I need to iron out as soon as this is done I'll post. Jason=20 |
|
From: Jason M. <mil...@nt...> - 2002-05-03 14:30:20
|
Hi Giuseppe I'll include that code in the next release, you are correct we need to = go back to usb logs to see how the scanner communicates. The problem is = that I'll need a log of a 150 dpi scan which can be as big as 8 - 10 = Megabytes (should be about 2 megabytes when zipped) This will contain = the USB setup communication data (when you plug in your scanner) the = initial scanner communication (setup and check scanner data) an initial = 75 dpi scan (from the preview in the canon scanner driver software) and = finally a 150 dpi scan (let the log run until it slows down). hence is = there a site you can upload this zipped log to then I'll download it = from there ? I already have a log for the D2400UF thanks to Flo and the = communication is very different to the N650U (although this may be due = to compression, if I can get a log file for the D646U it will help = confirm this speculation). I've probably mentioned this before but there = are subtle differences between the N family of scanners and the D = family, the first is the different light sources which definitely = require a different scan setup function. Jason |
|
From: Jason M. <mil...@nt...> - 2002-05-03 14:29:03
|
Hi Vladimir I see your ready to make a Sane Backend, have you looked at the Plustek = backend yet. This is the backend I'm working with to get the N650U = working with Sane. The Backend already covers the LM983X chip and you = can easily(supposedly) add other devices that use this chip to the code. = There is a test version that Gerhard (the writer) did for the N650U = (although it dosent work at the minute, which I'm supposedly sorting = out) that should work just as well for the N1220U. The backend covers = all the calculations from the chip specification manual, and also = includes calibration. The only problems are that the backend is huge, = and it dosent follow the procedure that the canon windows driver follows = (it follows the national semiconductor windows driver procedure), = another problem is that the test driver is the first (untested) version = to cover CIS scanners. The advatage of this backend though is once the = CIS side is implemented any Canon USB scanner should be easily = implementable. Hopefully this will help, If you're interested in taking = this path get in touch (I could do with some help after all two heads = better than one). =20 Jason |
|
From: Vladimir D. <vo...@mi...> - 2002-05-01 12:20:56
|
Sorry, false alarm. Still take a look at the driver for FB630U that they
have.
Vladimir Dergachev
On Wed, 1 May 2002, Vladimir Dergachev wrote:
>
> You might be happy to learn that http://www.mostang.com/sane seems to have
> a driver for your scanner. Check CVS snapshots for it.
>
> Vladimir Dergachev
>
> On Wed, 1 May 2002, Giuseppe Sacco wrote:
>
> > Il Wed, May 01, 2002 at 07:21:50AM -0400, Vladimir Dergachev ha scritto:
> > >
> > > Btw, could you try ct_1220U.c program with your scanner ? Comment out the
> > > exit() statement and put it right after it prints all statistics. I am
> > > interested in what it says about your scanner chip.
> > >
> > > thanks
> > >
> > > Vladimir Dergachev
> >
> > Voila:
> >
> > canon# ./canon-test-1220U
> > Starting conversation with the scanner
> > Statistics:
> > sizeof(init_data1)=2552
> > Optical resolution: 300x300 dpi
> > Vendor id=0x04a9 Product id=0x220b
> > Unknown USB scanner
> > If you still want to use this program with it edit ct_1220U.c to remove
> > the exit() statement just below this line. Be sure to e-mail
> > vo...@mi... with parameters of your scanner
> > Assuming the following hardware parameters:
> > memory_unit_size=2048 bytes
> > total_memory=524288 bytes
> > max_height=0 inches
> > max_width=0 inches
> > Write warning: 4 bytes requested, -1 written
> >
> > in syslog I found:
> > May 1 13:44:44 localhost kernel: usb_control/bulk_msg: timeout
> > May 1 13:44:44 localhost kernel: scanner.c: write_scanner: NAK received.
> > May 1 13:44:54 localhost kernel: usb_control/bulk_msg: timeout
> > May 1 13:45:34 localhost last message repeated 2 times
> >
> > Bye,
> > Giuseppe
> >
>
>
|
|
From: Vladimir D. <vo...@mi...> - 2002-05-01 12:18:53
|
You might be happy to learn that http://www.mostang.com/sane seems to have a driver for your scanner. Check CVS snapshots for it. Vladimir Dergachev On Wed, 1 May 2002, Giuseppe Sacco wrote: > Il Wed, May 01, 2002 at 07:21:50AM -0400, Vladimir Dergachev ha scritto: > > > > Btw, could you try ct_1220U.c program with your scanner ? Comment out the > > exit() statement and put it right after it prints all statistics. I am > > interested in what it says about your scanner chip. > > > > thanks > > > > Vladimir Dergachev > > Voila: > > canon# ./canon-test-1220U > Starting conversation with the scanner > Statistics: > sizeof(init_data1)=2552 > Optical resolution: 300x300 dpi > Vendor id=0x04a9 Product id=0x220b > Unknown USB scanner > If you still want to use this program with it edit ct_1220U.c to remove > the exit() statement just below this line. Be sure to e-mail > vo...@mi... with parameters of your scanner > Assuming the following hardware parameters: > memory_unit_size=2048 bytes > total_memory=524288 bytes > max_height=0 inches > max_width=0 inches > Write warning: 4 bytes requested, -1 written > > in syslog I found: > May 1 13:44:44 localhost kernel: usb_control/bulk_msg: timeout > May 1 13:44:44 localhost kernel: scanner.c: write_scanner: NAK received. > May 1 13:44:54 localhost kernel: usb_control/bulk_msg: timeout > May 1 13:45:34 localhost last message repeated 2 times > > Bye, > Giuseppe > |
|
From: Giuseppe S. <epp...@us...> - 2002-05-01 10:01:33
|
Il Wed, Apr 03, 2002 at 02:06:25AM +0100, Jason Millward ha scritto: > Hi > > Giuseppe, Try n650u-v1.2.tgz, the two programs are exactly the same except > the latest uses libusb. Unfortunately I've not included alternative > configurations in the libusb versions as I am not that profficient with > libusb yet, yes you do have to be root for libusb. And sorry about that Flo, > the link > to n650u-v1.2.tgz is now fixed Hi, today I made some tests with your code and I posted a detailed message using the wrong From: address. I hope the list moderator will approve it. In the meantime I will try to resume what I did. I downloaded the latest source from your page and I tried it with scanner D646U and a linux laptop using kernel 2.4.18. The program runs and immediatly starts looping writing this text: Write warning: 5 bytes requested, -1 written I fixed the looping (a patch is attached) and ran it again in order to have more informations. The program stop at the first reg_w(0x03,0x00) in intialize_scanner() for a timeout error. I had a look at the kernel function that raises this error and I believe that everything is okay. Now I think that the problem is related to the value that reg_w() is using, i.e. we should start again snooping :-( Bye, Giuseppe |
|
From: Giuseppe S. <epp...@li...> - 2002-05-01 09:25:03
|
Il Wed, Apr 03, 2002 at 02:06:25AM +0100, Jason Millward ha scritto: > Hi > > Giuseppe, Try n650u-v1.2.tgz, the two programs are exactly the same except > the latest uses libusb. Unfortunately I've not included alternative > configurations in the libusb versions as I am not that profficient with > libusb yet, yes you do have to be root for libusb. And sorry about that Flo, > the link > to n650u-v1.2.tgz is now fixed Hi, today I had enought time to make some tests with D646U. Firstly I downloaded the latest n650u-v1.2.tgz from your web page. Then I modprobed my scanner modules and then changed the file name for the scanner /dev/usb/scanner0. The output is: # modprobe scanner vendor=0x04a9 product=0x220b eppesuig1:/home/giuseppe/usb/xx# ./N650U 75 Statistics: Optical resolution: 75 dpi Scan size: A4 Starting conversation with the scanner. Write warning: 5 bytes requested, -1 written Write warning: 5 bytes requested, -1 written Write warning: 5 bytes requested, -1 written Write warning: 5 bytes requested, -1 written .... My syslog was filled with this text (from modprobe to rmmod) hub.c: USB new device connect on bus1/1, assigned device number 2 usb.c: USB device 2 (vend/prod 0x4a9/0x220b) is not claimed by any active driver. usb.c: registered new driver usbscanner scanner.c: probe_scanner: User specified USB scanner -- Vendor:Product - 4a9:220b scanner.c: 0.4.6:USB Scanner Driver usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb_control/bulk_msg: timeout scanner.c: write_scanner: NAK received. usb.c: deregistering driver usbscanner My machine is a Debian GNU/Linux "woody", kernel 2.4.18. The second test I did was to compile your code with debug options and trace it using gdb. The program stops at the first writing to the scanner. In initialize_scanner(), the reg_w(0x03,0x00) fails with the text shown above. I then corrected a very little bug of your program, in order to get the correct error message (the patch is attached) and I got this message: Starting conversation with the scanner. Write warning: 5 bytes requested, -1 written send_cmd failure: Connection timed out I then traced the kernel to see where it stop. It is in function usb_start_wait_urb() in file drivers/usb/usb.c . I believe that there are no problem with this piece of code, so the only reason to get this timeout error is that the reg_w is passing wrong argument for this scanner. Conclusion about the D646U: we should start again from the snoopy logs :-( Bye, Giuseppe |
1 message has been excluded from this view by a project administrator.
