Re: [Hp3300backend-devel] HP 3400Cse scanner not working

[Bertrik S. <be...@zo...>]

Steven A. DuChene wrote:
> I just pulled the latest cvs sources and tried them on a HP 3400Cse (what it says on outside
> of the box) and it does not respond as expected. I am using linux kernel method with kernel
> from Mandrake (2.4.19-16mdk) since it has support for Win4Lin modules I also run. They are
> not loaded BTW when I do these tests.
> 
> Product number from the scanner box is C7720A  ABA  CN1281F06Y
> 
> If I do any commands with testtool I get the following in the dmesg output:
> 
> scanner.c: open_scanner(1): Unable to access minor data
> scanner.c: open_scanner(2): Unable to access minor data
> scanner.c: open_scanner(3): Unable to access minor data
> scanner.c: open_scanner(4): Unable to access minor data
> scanner.c: open_scanner(5): Unable to access minor data
> scanner.c: open_scanner(6): Unable to access minor data
> scanner.c: open_scanner(7): Unable to access minor data
> scanner.c: open_scanner(8): Unable to access minor data
> scanner.c: open_scanner(9): Unable to access minor data
> scanner.c: open_scanner(10): Unable to access minor data
> scanner.c: open_scanner(11): Unable to access minor data
> scanner.c: open_scanner(12): Unable to access minor data
> scanner.c: open_scanner(13): Unable to access minor data
> scanner.c: open_scanner(14): Unable to access minor data
> scanner.c: open_scanner(15): Unable to access minor data
 >
 > and it repeats over and over.
 >

As far as I know, this is just a noisy warning message, not
a real error. I remember seeing this fixed in the latest versions
of the scanner module (perhaps only in kernel 2.5.x yet).
This seems to be caused by attempts to open 
/dev/usb/scanner0../dev/usb/scanner16
Anyway, the message is harmless and is no reason to upgrade the kernel.

> If I run "./testtool -u /dev/usb/scanner0 -ov"
> the output looks like:
> 
> Initializing scanner
> Registering transfer method 0: Linux scanner kernel module
> VendorId 0x03F0, ProductId 0x0405
> Matching USB device 0x03F0-0x0405 ... found Hewlett-Packard ScanJet 3400C
> _ReportDevice '/dev/usb/scanner0'
> Hp3300cXferOpen '/dev/usb/scanner0'
> probing Linux scanner kernel module
> Matching USB device 0x03F0-0x0405 ... found Hewlett-Packard ScanJet 3400C
> Setting params for Hp3400c/Hp4300c
> Probing scanner...
>   Gamma table entries are 16 bit
>   NIASH version > 00014
> Creating calibration file
> _iScaleDown (Dpi,Lpi) = (1,1)
> _iBytesPerLine = 15900
> _iLinesPerCircBuf = 12
> Allocated 190800 bytes for circular buffer
> _iLinesPerXferBuf = 3
> Xfer block size = 47700
> buffer level =  69, <reading 47700 bytes>, buffer level =  72
> Black level (11,6,5), White level (40,27,25)

This looks OK.

> If I do a register dump it appears to have different register contents
> from what is shown for the 3400C at:
> 
> http://home.kabelfoon.nl/~bertrik/hp3300c/devel.html
> 
> (bottom of the page)
> 
> Anyway when I do a register dump I get:
> 
>  ./testtool -u /dev/usb/scanner0 -p
> 
> 0000 0D 0D 0D 4C 07 0D 40 02
> 0008 0D 0D 0D 0D 0D 0D 0D 0D
> 0010 0D 0D 0D 0D 0D 0D 0D 0D
> 0018 0D 0D 0D 0D 0D 0D 0D 0D
> 0020 00 0D 0D 0D 21 00 00 22
> 0028 01 03 79 00 00 00 00 00
> 0030 00 00 00 00 00 00 00 00
> 0038 3B 0C 43 01 01 01 00 00

This also seems to be OK.
Some registers are write-only and return garbage when read.

> I can turn the lamp on and off but when I try "./testtool -u /dev/usb/scanner0 -s 4,5 > stufferty"
> it "looks like it is scanning but I get a zero length file. Also once I have tried the above scan
> command I can no longer access the scanner. When I do I get an error message about:
> 
> ./testtool -u /dev/usb/scanner0 -l
>   No NIASH chipset found!
> _ProbeRegisters failed!
> 
> If I unplug the USB cable and plug it back in then I can do stuff again until I actually try a
> scan again.
> 
> BTW, when I try a scan it move the bar about half way do the track, it pauses, and then it
> moves to the end of the track but does not stop! It grinds a while and then stops. It will
> try to move beyond the end of the track again a few seconds later until I end the above
> command with a Ctrl-C    Also the scan command never returns and I HAVE to Ctrl-C it.
> Is it supposed to scan and then return?

I think the testtool does not behave really well when trying to scan
very narrow strips. I'm not sure why, maybe this is a hardware
limitation.
The grinding noise is probably caused by the transfer buffer mechanism
in the software. This mechanism tries to read data from the scanner in
blocks of size 0xF000 (to optimise USB transfer overhead). When scanning
a narrow strip, it needs to scan too far in order to read the full
0xF000 block and so it goes past the scan area.
This probably needs a fix, or some mechanism to prevent it (like
imposing a minimum to the width).

Basically, it seems that your scanner mostly works.
Please just try it with a bigger width (like 1000 pixels)

Kind regards,
Bertrik