> >not critical that it be pretty - just readable. I wouldn't be surprised
> >if you could rig up something like this pretty easily yourself; pick up
> >one of the fiche viewers that they can't give away at most university
> >and government auctions, get a cheap flatbed scanner (even pretty good
> >new ones can be had for under $100), take the mechanism out of the case,
> >and bolt it to the front of the fiche viewer. You'd have to remove or
> >disable the light source, since the bulb in the fiche viewer provides
> >the illumination. I don't think you'd even have to mess with the focal
> >length much; the fiche viewers normally do a rear-projection on frosted
> >glass, and the scanner is set up to focus on a sheet of paper an inch or
> >two away from the sensor, so with a spacer or two it should just work.
> >That sounds like an interesting enough project that I might even build
> >one if I had anything on fiche to scan. (I'm more interested in getting
Yesterday I tried a similar way as described above. I took a sheet
of Butterbrotpapier (sorry, I dont know any translation, its a kind
of semitransparent paper used to cover sandwitches - Butterbrote :)
put it atop the scanner, disabled the lamp (I had to disable the lamp
only during the 'real' scan, since my scanner always add a 'white
adjustment' before scaning, so the the lamp is only temporary
disabeled AND I had to fake the lamp controll signals ... this took
me almost three hours to figure out how they controll it), added one
mirror and used a fiche projector to display one page right on the
paper screen.
<pre>
mirror
/ light --
/+- - - - - - - - - - - - - I projector with fiche
/ : --
/--------------- paper
######## scanner
</pre>
Since the fiche information was only b&w, the
contrast was no problem at all. This method
should work well for low resolution scanners.
> You could do it even without changing the scanner at all. take two
> mirrors and build a 'hat' for the fiche - the light will pass the fice at
> one side, get reflected by the mirrors and be passed back, now thru
> the fiche. This works quite will since in most scanners the light runs
> some one or two cm in front of the scannerline. Of course the 'hat'
> has to be adjusted to this angle. It works quite well for pictures, but
> for a fiche you might need a _real_ high resolution scanner. If the
> fiche document is made from ordinary 'typewriter' papers or drawings
> (like most of my 1960s/70s mainframe fiches) the resolution should
> be better than 1200 dpi to get an idea of the text. If small parts are
> to be viewed, >3000 is needed. One advantage of fiches is still the
> _real_ high resolution :)
Postscript:
<pre>
frontview (in direction of scanner movement)
/\mirrors
/ \
/+--+\ way of light
/ I I \
/ --- I \ fiche
------------ scanner
sideview:
+-------+ mirrors
I I\ I way of light
I I \ I
I I \ I
I ----- * fiche
----------- scanner
>>>>>>> movement
* = this side has to be a bit higher - the 'hat' is
lifted here to fit the lamp angle of the scanner -
this needs some trieal and error. Its also for scanners
with the lamp infront of the scan line - otherwise the
oherer side has to be lifted.
Sorry, but it's hard to do high resolution 3D ASCII
drawinge :)))
Gruss
Hans
--
Ich denke, also bin ich, also gut
HRK
Received on Tue Dec 08 1998 - 07:00:49 GMT