[sane-devel] Epson Perfection1640: problems with TPU

[ael]

A quick provisional update. I hope to explore further later today.

It looks as if scanning with the TPU on the 1640 is failing.

1) I suspect that neither "Positive Slide" nor "Negative Slide" are 
   valid for the 1640SU (with TPU installed).

As yet, I don't know the difference between Positive Slide and Positive
Film. If this is a distinction inside the firmware of some Epson models,
then it looks like the 1640 is not one of them.

If it involves the backend sending different parameters, then I haven't
yet found the relevant code.

2) As far as I can tell from the debug traces, no focus position command
is sent by the backend. I am trying to understand e2_discover_capabilities
and why no focus position command is issued. It appears that the scan is
carried out with focus position = 0x40 which is on the glass, not
above it. In my tests with "Positive Film", it looks as if the focus
may indeed be wrong. I intend to test that more precisely later.

3) As I understand it, the current version intends (but is failing?)
to issue the focus position automatically when the TPU is installed,
and no longer offer an explicit (binary) choice to the frontend.
But that assumes that the TPU is not useful when focused on the glass.
But the TPU can be used to scan transparent objects placed directly
on the glass. I am intending to use a transparent protractor in
my tests for focus later.

So I think that the backend should offer the binary focus choice again.

ael