On 2010-12-19, Lloyd E. Sponenburgh lloydspinsidemindspring.com wrote:
Ignoramus7337 fired this volley in
:

The trouble that I am concerned with, related to high counts per one
revolution of the encoder, is that encoder or the control may fail to
count them properly and may miss some.

Too few counts == low accuracy due to large vertical move per count
Too many counts == missing counts from time to time

If missing counts is the issue, then you have two avenues of recourse:

You can use a faster processor, so that interrupt service latency is
shorter, or you can use a hardware decoder board for the quadrature
encoder, so that any time you wish to read it, it will be up-to-date.
It's not necessary to resolve every transition in software.

Here's what I know.

1. The US Digital E5 module has a limit given by its 100,000 kHz
frequency. At 4,096 counts per rev, this means no more than 5,859
RPM. So, no problem here, since I cannot run it above 4,600 RPM ever,
and will not tap above 1,000 RPM.

http://usdigital.com/assets/general/...atasheet_1.pdf

2. I use Jon's Pico Systems hardware based encoder counter.

http://pico-systems.com/encd.html

It can count even more counts per second. I have a recollection of
what Jon was saying, but will not repeat since I am afraid I may
misremember.

3. My main concern is, what if due to whatever screwup of my adaptor
manufacturing process, the optical disc is more misaligned than the
specification permits, and thus would not count as well.

i