Re: Lost steps

Thank-you,
I will try lowering the acceleration and see what happens.

The software is set for 8000 steps (half steps) per inch which was
set
by Tim Goldstein and as far as I can figure is correct.The commands
I'm giving are in 1/2" increments so I wouldn't think that an error
could occur from partial steps being needed.

My proceedure is to run the machine into the indicator, stop
set the indicator to zero
home the EMC to Zero
run the machine away to .5"
run the machine in to Zero.
I figure the indicator reading should be the backlash!
First run I get .002
I repeat the process without changing anything and I figure I should
get .002" again but I get more or less,i.e. no consistency.
Same on another axis.

Can motor damage cause this? These motors are not new.
How about iffy connections? When I first plugged in the Y axis motor
I
could see it stutter. I pressed the plugs together harder and it
smoothed out the motion but maybe the connection is not what it
should
be?

Also, if the accelleration setting is the culpret would the defect be
erratic?

Thinking while writing,
Jerry

--- In CAD_CAM_EDM_DRO@egroups.com, Ray <rehenry@u...> wrote:
>
> Ozzie@h...
>
> > Message: 17 [CAD_CAM_EDM_DRO] Digest Number 621
> > Date: Sun, 13 Aug 2000 21:20:02 EDT
> > From: JanRwl@A...
> > Subject: Re: Lost steps
> >
> > In a message dated 8/13/00 8:01:32 PM Central Daylight Time,
Ozzie@h...
> > writes:
> >
> > << Any ideas, and how should I go about finding the problem???? >>
> > Howz Harriet? I have fiddled with stepper-motors and home-brew
machinery for
> > 25 years, now, and, without seeing it, only reading what you say,
I wonder if
> > you might not be STEPPING at too-fast a rate, or, not
"accelerating" and
> > "decelerating" into/out-of your move? If the "book say" they
will
step OK
> > at, say, 400 steps/sec. without acceleration into that rate, and
you start
> > "cold" at 400 steps/sec., the motors might-just "be surprised"
before they
> > get moving! Try reducing the stepping-rate 25 to 50
steps/second,
and re-try!
>
> You can easily adjust acceleration in the ini file that you are
using.
> The stock emc.ini uses a value of 20 which is pretty steep for some
setups.
> It looks like this:
>
> ; Trajectory planner section
> --------------------------------------------------
> [TRAJ]
> ...
> DEFAULT_ACCELERATION = 20.0
> MAX_ACCELERATION = 20.0
>
> You can play with the numbers there. Try 2.0 and you should hear
your
> motors ramp up. That would take care of any missed steps do to
accel.
>
> You didn't say what your steps per inch are but a small part of
what
you
> may be reading as error could be the difference between the
commanded
> position and the nearest step position. (You may also loose a
little
> accuracy half stepping. I don't remember how much and don't have
the
> url's for those articles at hand.)
>
> The first thing I'd do is set backlash comp to 0.0, command one
axis
to
> move in one direction so that you take up any real backlash. Set
the dial
> indicator to read zero and command a move nearly as large as the
> indicator's travel. Compare the distance on the dial to the
distance you
> commanded to see if the number of steps per unit are what you
expected.
> If not, change the steps per unit in the ini file and try it again.
>
> Now with the steps per unit set properly for each axis, compute the
> real deadband - the distance between steps - and enter those values
in the
> ini file. As you work to minimize the backlash using compensation
you
> need to remember that your indicator readings will be off as much
as
+-
> the deadband.
>
> This explanation is speculation. If you do not set a deadband
equal
to the
> size of your minimum step distance the EMC will oscillate. Since
backlash
> comp is fed in a little each time the motion stuff updates it may
not be
> done compensating when the next oscillation occurrs and it starts
to
feed
> in the comp for the other direction.
>
> IMO Backlash comp is iffy at best. If you try to take up very
much
you
> will get some strange results that show up in surface finish even
on
the
> best production machines with the latest controllers.
>
> Ray