Re: [CAD_CAM_EDM_DRO] cnc mill project

John D. Bussema wrote:

>First, thanks for the info regarding Gecko drives, I found it very useful. My friend and I are industrial engineers at U of I and are converting a manual Taig micromill to CNC as part of an independent study in adaptive control systems. We currently in the very beginning stages of the project and are trying to figure out what components to purchase for the initial conversion. I have a pretty broad background in machine tool systems, in addition to being a metalsmith, but I don't know all that much about CNC interfacing. We'd like the machine to be both accurate and cheap. We've decided to use the Gecko drivers for the project, as I've seen them on numerous websites and all have spoken very highly about them. I was wondering if anyone has suggestions for the type of servos we should use, particularly: typical ranges for continuous torque, speed, voltage, and encoder resolution (we plan on using rotary encoders). We're also interested on upper bounds of the same parameters (i.e. the maximum values that are worth paying for). If anyone knows of a cheap surplus servo supplier we'd be interested in that too Thanks a lot.
>
>

Your minimum encoder resoultion should come out so that the system
resolution is
at least 10,000 counts per inch. This gives plenty of resolution to sit
stably at a
.001" point. Too high an encoder resolution gets into trouble when software
generated step pulses can't be produced fast enough. (Hardware can, of
course,
fix that.) Too low an encoder resolution causes problems with step-like
movement at low speeds, and also with finite gain of the servo drives.
So, the encoder resolution is somewhat dependent on the motors you choose,
and what belt reduction ratio is optimum for those. A 1000 Oz-In motor
good to 900 RPM can be used with 1:1 drive, while a 200 Oz-In 5000 RPM
motor will likely require a reduction ratio of 4:1 or more. You can
figure out the
worst-case linear force produced by a cutting tool loading the spindle motor
to its rating, and then work back through the leadscrews to figure out the
torque needed to drive the screw. That's what I attempted to do when
doing my
first retrofit.

Jon