[sane-devel] Epson Perfection 2400 - conks out on desktop machine but works on laptop

[Olaf Meeuwissen]

Phil Reynolds <phil-sane at tinsleyviaduct.com> writes:

> On Mon, May 05, 2008 at 07:20:44PM +0100, Phil Reynolds wrote:
>> My scanner has been in storage while I had no room for it, but I
>> reinstalled it today after removing a redundant printer.
>> 
>> However, despite both my laptop and my desktop having similar setups,
>> the scanner fails - usually coming to a halt - part way through a scan
>> on my desktop machine. It does not do so on my laptop, but I cannot see
>> any reason why it would fail on the desktop - there are no obvious error
>> messages.
>
> Self-followup - I have enabled debug output for the epson backend and
> had it fail on a preview. This is what happened...
>
> (lots of buf before this of course, various values)
>
> [epson] buf[14332] ff �
> [epson] buf[14333] ff �
> [epson] buf[14334] ff �
> [epson] buf[14335] ff �
> [epson] send buf, size = 1
> [epson] buf[0] 06 .

Here the backend acknowledges that it has received a full block of
image data.  Now it expects some "meta"info for the next block, so
it knows what to expect.

> [epson] w_cmd_count = 44
> [epson] r_cmd_count = 3804
> [epson] sane_read: begin scan2
> [epson] sane_read: end
> [epson] sane_read: begin
> [epson] sane_read: end
> [epson] sane_read: begin
> [epson] sane_read: end
> [epson] sane_read: begin
> [epson] sane_read: end
> [epson] sane_read: begin
> [epson] sane_read: begin scan1
> [epson] w_cmd_count = 44
> [epson] r_cmd_count = 3804
> [epson] receive buf, expected = 6, got = 0

Looks like that "meta"info never got to the backend, for whatever
reason.  This is most likely a timing and/or a hardware problem.

> [epson] scan_finish()
> [epson] send buf, size = 2
> [epson] buf[0] 1b .
> [epson] buf[1] 66 f
> [epson] w_cmd_count = 45
> [epson] r_cmd_count = 3804
> [epson] send buf, size = 2
> [epson] buf[0] 1b .
> [epson] buf[1] 66 f
> [epson] w_cmd_count = 46
> [epson] r_cmd_count = 3804
> [epson] w_cmd_count = 46
> [epson] r_cmd_count = 3804
> [epson] receive buf, expected = 4, got = 0

Ditto here.  The expected reply is not making its way to the backend.

> [snip]
> The preview stopped about 14 cm down the image area.

Have you experienced other problems your desktop's USB port when using
large amounts of data?  Something like not being able to reliably copy
large files to/from a USB stick for example.

The fact that things work okay on your laptop (and assuming that you
use the same version of the same backend there) seem to point the
finger to your desktop's USB hardware (or the kernel support for it).

Hope this helps,
-- 
Olaf Meeuwissen                   FLOSS Engineer -- AVASYS Corporation
FSF Associate Member #1962           sign up at http://member.fsf.org/