Gecko minimum encoder specifications.

What adverse effect would using an encoder with a smaller resolution
than stated in the Gecko G320 manual have on its operation? Has
anyone tried using a G320 with a servo with less than 200 steps per
rev (800 in quadrature)?

The only thing that I can think of is that the PID loop might become
so severely underdamped due to lack of feedback that it could dither
like crazy between widely spaced lines on the encoder.

There are a couple of reasons I wonder this :
1. I'd like to have a reasonable traverse rate on my current mill
with the windows 8K pulse limit. I could trade slightly less
positional accuracy for faster rapids using say a 100 count encoder.

2. In thinking about a CNC router I've been planning to build I was
figuring strategies for reducing the backlash with the Acme screws
I've got lying around. It occurred to me I could completely elminate
backlash as an issue by the use of linear strips and encoders from
usdigital.com as feedback for the motor controller. These would get
the actual position of the gantry versus the position of the motor.

The problem is that if I were to use the 250 lpi linear strips with a
5-1 lead screw that would be equivalent to a 50 line per rotation
optical encoder. Way below the minimum of 200 for the gecko. Has
anyone tried a gecko drive with such poor rotational feedback?

The obvious advantage would be that my table would be positioned to
within +/-.001 absolute which is pretty good for my purposes. It
would be slightly better if I went up to the 360 lpi strips (+/-
.0007), but not nearly as convenient in even numbers.

Has anyone tried something like this with the inexpensive UsDigital
mylar strips? I think there are commercial machines that use glass
scales in this fashion, but my impression is that these are
prohibitively expensive.

What about magnetic scales? Does anyone know a source for magnetic
encoder linear scales?

Thanks.

-Erik Reikes