[sane-devel] Information about unsupported UMAX Astra 4600

stef stef.dev at free.fr
Wed Nov 4 05:20:27 UTC 2009

Le mardi 3 novembre 2009 13:38:44 MP, vous avez écrit :
> >         from the chip's label, support for you scanner belongs to the
> > rts8891 backend, even though sane-find-scanner doesn't detect it
> > properly. Do you have a windows box where you can run your scanner ? If
> > so I'd be interested by an
> I have only windows instalation inside virtualbox, but as vbox does
> support bridging USB devices inside virtual machines, I was able to
> scan a 75 dpi color page from the scanner, while capturing
> communication using UsbSnoop.
> Since the log file is 3.6Mb even after bzipping, I uploaded it to
> http://wtx.wz.cz/UsbSnoop.log.bz2
> In the log, actual scanning (when the scanner started warming up)
> starts about 97000 ms
> >  Usbsnoop (http://www.pcausa.com/Utilities/UsbSnoop/) log of a 75 dpi
> > color scan. I'd be able to tell for sure how close it is from supported
> > models. If needed I can provide some explanations on how doing it.
> >         From the known information, your scanner doesn't seem to be far
> > from the Astra 4400 which isn't yet supported. Bringing support for these
> > 2 scanners is a matter of how different they are from supported models.
> > If close, by exchanging logs and test code we'd be able to have them
> > working. But if they are too different, support wouldn't be possible with
> > access for a developer to the real hardware.
> What about some form of remote access?
> >         If you feel like it you could add your scanner to rts8891.conf.in
> > and rts8891_devices.c source files (cloning the HP4470 entry), and have a
> > try. Be aware that when experimenting with partially supported hardware,
> > you can't exclude it could be damaged.
> What type of damage can I expect from that?  Something like the
> scanner lamp driving off the rails, which could be maybe fixed just by
> using a screwdriver or some parts irreversibly burning out?
> Martin Petricek


	I have downloaded and decoded the log. I can confirm that your scanner will 
work with the rts8891. The calibration (offset, gain and shading) uses the 
same algorithm than the one coded in the backend. Also it looks close to the 
Astra 4400 log I already have. 

	But there are some differences that may make support for it difficult. First 
the data in log looks like noise, I'd be inclined to think that it is 
compressed. Did you scanned a white area ? Then it seems the scan is done at 
200 or 100 dpi even if 75 was required. This resolution isn't supported at 
all. So there will be some significant changes. While remote access to the 
scanner would help, I think the best solution is that you start hacking the 
backend. I will provide help for that. Another solution is that I can have the 
same hardware. 

	Based on the backend I worked on, here are the risks I see when adding 
support for new hardware:
	- send wrong commands to motor, having them doing some strange noises and 
	  maybe shortening their lifespan
	- moving scanning head 'past end of device', having it banging on physical
	- send wrong command to scanner's lamp so it make it physically fails
	- wipe out EEPROM, firmware

	On the 10+ scanners I toyed with, I only have my HP4470 damaged. I wiped 
EEPROM. Now it reports itself to be a Astra 4400. The offending code is 
commented out and it shouldn't happen again with the rts8891 backend. None of 
my scanners had a physical failure. If you stand close to the scanner, you can 
unplug it when it is doing some strange noises. 


More information about the sane-devel mailing list