Re: [CAD_CAM_EDM_DRO] Skipping steps, avoiding resonance

Randy Gordon-Gilmore wrote:

Randy,
You have indicated straight line moves cause no problems ??

When step loss issues arise, I find that I am better off finding out
whether all axis are stable by testing with straight line toolpaths and
a fixed number of direction changes. I'll create a file that runs a
single axis 2" out and 2" back at least 400 times. I do this test on one
axis one at a time. If there is no step loss at even the highest speeds
you think you can achieve, wonderful. But, if you do have step loss,
note how much, then run the same file again without re-homing. If your
step loss has effectively doubled, Then you can almost conclude that
your losing steps on each direction change. A third test could really
conclude this. Loss of position because of direction change can really
add up on a file with a lot of short challenging moves.

If all three axis run fine by themselves with the above tests, then
create a toolpath that runs all 3 axis at the same time, 200 straight
line 3 axis moves in each direction. If the above had no loss, but this
does, it could be harmonical, however, Sometimes you need to create a 3
axis move that is at an angle that makes the machine groan to really
check this. Some angles will run smoother than others.

If straight line moves really work well, then I move to a very short
toolpath that can develop the loss of position. This might be thru the
use of a known number of short line segments with a known number of
direction changes, or, in some cases a combination of arcs. I then set
up dial indicators on the machine itself so there is no doubt that the
hardware is not at fault. If a very short toolpath can cause loss of
position, I again look for simularities in numbers. If you run the path
twice, and now your off twice as much... Your losing on Direction change.

Most of my experience with step loss is that one does not have the right
settings for your drivers. Is your Pulse width set correct ? Is it at
the right level ? Does your driver require leading or trailing edge ?

Have you played "systematically" with all of the settings the software
allows ??
I ask this (systematically), because sometimes we get too "jumpy",
adjusting at random, up-down, high-low... without keeping a written
record of what happened each time we made a change. Often times that
written "log" is what leads us to the Fix ! Keep a written log with the
results !

In many cases, some of the PC control software does not even give you
much for options to set all of the above. I've had drivers that just
would not work correctly with some control packages, and I have had
drivers that would run accurate, but only after using completely "Wrong"
settings. Tims suggestion may be a fast way to find out whether software
and settings could be your problem.

Your next test might be replacing a driver or two.
I can't say I have ever had "crosstalk" problems, but then again I
always use shielded cable with the drain wire attached on one end, and
all in the same spot.

Step Loss can be one of the most aggravating problems..... You must rule
out one possibility at a time.

As a reminder for all chasing these problems, remember to compare your
dial indicator results with your software.... Some limit switches could
be throwing off your "home" routine.

Of course, you SEE your position error in your toolpath...... So, that
is different.

Keep us posted....

Chris L