How about optical interrupter switches?  Eg: 
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=H22LOI-ND

The slide would simply occlude each of the optical switches in turn.  

If it interrupts one or two at all times, then it will have a position.  To 
detect n positions, you only need n/2 switches, eg for 3 switches you get 6 
positions sensed: none, 1 only, 1&2, 2 only, 2&3, 3 only.  

David

Quoting Dave Tweed <pic@dtweed.com>:

> ivp wrote:
> > The client has indicated he'd like the slider moved by a brass crank. !!!!
> > This probably means that the slider will be on a lead screw and several
> > turns of the crank will move the pointer from one item to the next
> > 
> > The reason being that this is a museum exhibit of a 1906 Edison cylinder
> > player, which has been motorised but originally was wound up with a
> > crank, so he wants to keep some of that feel to it. The selector is to
> pick
> > a song to play from a bank of WAV files
> 
> Interesting!
> 
> Did you know that there are people who are still making wax cylinders for
> Edison machines? I recently saw a presentation in which they were
> re-creating
> an old-time acoustic recording session. You wouldn't believe the gyrations
> they went through while recording in order to get a decent-sounding
> playback.
> Between the lousy dynamic range (high noise floor) and terribly non-flat
> frequency response, you basically had to do a lot of "pre-compensation" of
> the sound balance just to get something recognizable back out.
> 
> > My idea with a long PCB is that it would have fixed resistors and solid
> > detentes, thus turning the crank is a WYSIWYG. It's for public use and
> > something I'd rather not have to service or field complaints about
> 
> If I were you, I'd seriously consider non-contact solutions, such as
> capacitive sensing.
> 
> For example, the brass crank could move a brass "pointer" along the surface
> of the PCB, almost but not quite touching it. Sensor pads built into the PCB
> artwork (sealed underneath the soldermask) would detect the presence of the
> pointer at particular locations.
> 
> There are many ways to do multi-point capacitive sensing with low parts
> count these days, anything from dedicated chips to ordinary I/O pins on
> general-purpose microcontrollers. Usually, you just need one or two passive
> components per point sensed, which compares favorably with your resistive
> divider.
> 
> -- Dave Tweed
> -- 
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
> 




-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist