Re: stepper vs. servo - hundreds of microsteps

--- In CAD_CAM_EDM_DRO@yahoogroups.com, Jon Elson <elson@p...> wrote:

>
>
> turbulatordude wrote:
>
> >
> >But that opens the question of how others can get 256 microsteps

or

> >more. And THAT really is a more general motion control questions

as

> >even though we tend to think our interests are 98% of the

universe,

> >CNC machining applications of motion control are probably more in

the

> >2% range. I'll bet Hewlett Packard (one supplier) sells more

Laser

> >Printers in one year using motion control than all the home brew

CNC

> >stuff ever built.
> >
> >
> If you look at the deflection vs. torque curves for stepper motors,

it

> becomes
> obvious that for most motors, the friction in the motors OWN

bearings, only,

> is enough that you can't move the rotor in increments smaller than

about

> 1/10th of a full step. Any moves smaller than this don't produce

any motion

> until you have accumulated some deflection between commanded and
> actual position, then the motor jumps. Thinking that a steper

motor can

> actually
> move in 200 * 256= 51200 microsteps/rev is ludicrous. That comes

out to 25

> arc seconds, I think.
>
> Some of the problem in this stepper/servo debate is caused by the

makers of

> microstepping drives who make drivers that do microstepping down to
> thousandths of steps, without declaring that no motor can be

expected to

> follow such fine steps. There is some maker that has a drive that
> divides down
> to 144,000 steps/rev, or is it 144,000 microsteps per full step?

Either

> way,
> there is no motor that can take advantage of such fine

microstepping, and

> anyone who thinks their machine can be commanded to move in

nanometer

> increments because of this fine microstepping is, I'm sorry to have

to

> say this,
> a fool. Everything is flexible, of course, and the torsional

spring

> rate and
> compressibility of the leadscrew would make this impossible, even

IF the

> motor could actually do it. I get a bit fed up with people saying

their

> steppers
> are more accurate than any servo system, because they have

microsteps equal

> to microinches. But, of course, they have NOT measured the

performance, and

> in fact they may have .010" of backlash, .002" of cyclical

leadscrew error,

> .003"/Foot of leadscrew pitch error, .001" of static friction, and

a

> table that
> is .025"/foot out of square with the supposedly orthogonal axis!

And, they

> have absolutely NO IDEA that any of these huge errors are present

in their

> machine.
>
> I KNOW what these errors in my machine are, because I have measured
> them with the best equipment I have available (no Renishaw ball-bar
> testers or
> laser interferometers, just gauge blocks, precision squares, dial

test

> indicators,
> etc.)
>
> Jon

Hi Jon, Mariss,

The telescope crowd talks about 200 to 500 microsteps in moving a
telescope and talks about 50 microsteps as commonplace.

I hoped I was clear that 'motion control' is not CNC machine control
for mills and lathes.

Yes, their goal is keeping the steps down below 1/4 arc seconds for
the scope, and gearing is needed.

As you might imagine, a stepper motor banging between positions will
induce a frequency at some point that will cause the machine to
vibrate horribly. At least when the harmonics of the machine are in
a similar frequency.

The motion control goal is not step accuracy as we often consider,
but a constant movement so as to prevent vibration from the detent
from the steps and also to prevent a pause between steps that would
make the image oscillate forward and back from the theoretical
alignment.

I have to admit I am impressed at the photographs from some of these
machines. It may be the flywheel effect or the mass of the unit that
keeps the distant stars aligned. Whatever it is, the result can be
awesome.

And, yes a scope trace shows that the actual movement is not linear
to the pulse output.

Dave