Pang wrote:
>
> Hi,
>
> Thanks for suggestions and questions. It shows I have not been thinking
> thoroughly on the method of implementation. First of all, the application I
> will be building is going to be a tilt sensor.
> It is meant to be installed
> in an automotive to detect towing. As such, I think an angle sensor will be
> a more appropriate name.
>
>The potentiometer is
> fixed to a flat board. Whenever the board is raised at either side, the
> weighting will move the shaft.
> I have enquire things like rotary encoder, position sensor, optical sensor
> and angle sensor. But most of them are either costly or that they cannot
> withstand much load to the shaft.


I have a solution that is cheap, small, reliable and
"zero" torque. The only down side is that it will need
a calibration table in software.

Use a IR pair, a transmit/receive reflective pair as
found in VCRs etc. (Some are only about 3mm cube size)
Like this:

                       pin
                        *
                        |
                        |
   ,-------,            |
   |       |            |
   | IR    |==========> | White
   | pair  |            | surface
   |       |<========== |
   |       |            |
   |       |            |
   |       |            |
   '-------'            |
                  ,-----*-------,
                  |   Weight    |
                  |             |
                  '-------------'
              <----- swings ----->

(all under a light proof cover!)
The IR pairs can be attached to a PIC very easily, you
run the led at about 5mA and set the receive diode as
half of a voltage divider, and it will give an analogue
voltage somewhere between 0-5v as the device is tilted.
The voltage output is not linear, but can be calibrated
in software to give an easy 32 positions etc, which is
probably plenty?
:o)
-Roman

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.