Jinx wrote:
>
> > On a timed match, there is no extra time for plugging the pieces :)
>
> Which might rule out some options unless you have the PIC
> wait for a settling time. For instance, as a magnet approaches
> a sensor the sensor will see an increasing field. When does the
> PIC decide that the piece has been placed ? Perhaps when the

0.25 seconds without change?

> timer is struck ? That would also eliminate false positives from
> hovering-but-unplaced pieces

A hovering piece will have a fluctuating measurement due to tremor.
No human could hold a hovering piece in position to 1 part in 256.
A landed piece would be flat line plus or minus A/D noise.

Hall tempco is easily handled by having a 'reference' hall sensor that
is magnetically shielded but at same temperature as the others.
And you use an average of the 'empty' squares to compensate for
any stray fields. A minimum of 50% of the squares
will always be empty after all so no 'tempco' unit is needed in fact.

Halls can also be multiplexed (8x8) using their power leads
and diode coupling their outputs (think backwards wired-or)
with a reference diode to compensate diode tempco effects.
(differential op amp).

Etched coils would be infringing a l980's patent on
spiral PCB inductors. (See Radio Electronic tuner project).

How would MH suggest detecting the resonant circuits in the pieces?
Phase shift using something like a grid dipper? Presumably
at Mhz frequencies to keep the inductances small.

I like the elegance of this but dismissed it as harder
to do. It does have the advantage of not being centering
sensitive.

Maybe a combination of resonant ccts for ID and hall
(or reed switch) for detecting full landing since
orientation on approach and placement will have a
large effect on coupling coefficient.

Robert

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu