Reality bites was Re: Step count / inch problem

Jeff,

As the song goes, "Back to life.. Back to reality..."

Thank you for posting something that will be very important for many
to see/read.
As Jon E has already explained Why, I'm just saying to ALWAYS check
the REAL steps per inch of ANY system you are needing to trust!

But Jon did leave out one possibility; your 2:1 belt drive may not BE
2:1. If the error "stacks up" with longer moves, this is probably the
case. Ballscrew errors, even with the cheapies, do not usually stack
up; but instead are cyclical variations within shorter distances.

He is correct that you will likely need to average your SPI (steps
per inch) after careful measurements all along the screws' length.

Hope this helps.

Ballendo

P.S. This is one reason I smile when someone says they just made a
machine ACCURATE to .001 (It CAN be done, but often these machines
are not nearly as accurate as many of us would like to believe. And
that is OKAY!) And to those who claim .0001... :-)

--- In CAD_CAM_EDM_DRO@y..., "jeffdavis516" <cadman@p...> wrote:
> I have a question about why the math did not work out for my step
> counts per inch. On my servo system I just finished and using the
> calculations I got from help here, see below
>
> servo with 250 cpr, .2 pitch ball screw, 2:1 pulley system.
>
> 250 x 4 x 5 x 2 = pulses per inch or 10,000
>
> When I place this figure into my controller software and test it
with
> a dial indicator. (backlash taken into account) I come up a few
thou
> short. and a little more when I move 2". I had to set the counts
in
> the controller software to 10,010 to make it come out dead nuts on
at
> 1" and very close at 2" (This was the X axis on the mill)
>
> When I checked the Y axis I only needed to alter the value to
> 10,005. The Z was fine at 10,000. What would cause this? Could
it
> be that I did not purchase expensive ground ball screws?
>
> Jeff Davis
> http://www.homecnc.info