RE: [CAD_CAM_EDM_DRO] Encoder/Servo question

As for the encoder resolution question...Most controls
do something called a 4X interpolation on encoder lines.
This is a way to use the leading and trailing edges of the
two quadrature channels to effectively multiply the line count by 4.
So a 500 line encoder actually effectively generates 2000 per rev
in a control system. This is done by the device reading it.
Just a bit of extra info.

Now for your example of a 1 turn per degree worm drive...

Generally servo systems for machine tools work well if ten
pulses (not lines) represent about the desired level of repeatability.
This is not a hard and fast rule, but I use it in many applications.
A simple PID servo loop (without feedforward)requires an error for movement.
Very high pulse rates can be used, but this creates some problems.
If the machine structure is not capable of that kind of repeatability
there is no use in creating the extra problems.

One pulse per 0.0001 is commonly used on small mills and generally
gives .001" repeatability or better if the structure is capable.
Note repeatability and accuracy are not the same thing.

On a rotary table I would want about that equivalent on the outer
radii. So lets say the diameter is 10". One degree would have an arc length
of 0.0873 at the outside of the table. At a 10,000 pulse per inch pitch
that would require 872 pulses per worm revolution and would be obtained
with about a 200 line encoder! For things closer to the table center the
resolution would be better. A 100 line encoder would give the same results
for a 5" diameter table.

It might be nice to have the encoder on the table axis rather than the worm,
but that would require about an 80,000 line encoder! (10xPIx10000/4)

As far as the best tooth profile I would have to defer to higher authority
:).
I know the pattern is an involute profile- but only along the worm
centerline.
High pressure angles are often used (25 degrees plus). The worm on
centerline
looks a bit like a rack so the involute shape is a trapezoid in that case.
There is a huge amount of information out there though. Nice thick
engineering
books are written on just low backlash worm gear design. Fun reading I'm
sure!

Ah here's a link...
http://www.hermespenton.com/asp/bookdetails.asp?key=3056

Buy this book then tell me the answers. :)

Les

Leslie M.Watts
L M Watts Furniture
Tiger Georgia

Main page:
http://www.lmwatts.com
Engineering:
http://www.lmwatts.com/shop.html
Cnc surplus for sale:
http://www.lmwatts.com/forsale.html
Carved signs:
http://www.lmwatts.com/signwp.html

It's on the, "todo", list. At the moment, though, I'm just trying to put
together basic info on things. I finally saved up enough pennies to get a
7X14"
Minilathe, and hopefully, will soon be looking at a Minimill. Ya gotta
start
somewhere. I'll be trying to do a manual rotary first, and once the bugs
are
worked through, a servo job. That being said, I'm trying to figure out some
basic ballpark stuff, like what degree of accuracy I can get out of one,
hence
the encoder line count vs % of rotation question. Assuming a 360:1
reduction,
a 1000 line encoder would theoretically give you one thousandth of a degree
accuracy, yes?

On another tack, there's the question of what thread to use. How would a
28/29/29.5/30 degree. (whichever), acme thread compare to the 40 degree gear
type thread for the worm/wormwheel? Seems like it should have better torque
transmission, but whether they would be smoother acting or even more or less
suitable for cnc-ing, one over the other, is something I haven't seen
discussed
in this context.

Bill