Re: [Phys-l] Discrepant amplitude measurement

[John Denker <jsd@av8n.com>]

On 08/03/2011 02:40 PM, chuck britton wrote:
> A 13 track encoder would give a resolution of 0.0439 degrees (if I 
> understand it correctly)

Two tracks is enough (e.g. the finest two of the aforementioned
13).  That's because all you need is relative position, and
therefore all you need is to distinguish count-up from count-down.

And then you curve-fit to the data so you can interpolate and 
extrapolate.

Because you can interpolate and extrapolate, precise reproducibility 
of the count-events is more important than ultimate fineness.  The 
two often go together, but they are not the same thing.