Re: EMC stg2mod.o and problem with homing

Hi Jon, Les, Paul, Ray and all other group members!
Finally i was able to fix this home offset problem. I set
the voltage offset values for each axis to stop motion when machine
is in estop_reset condition
following are ini parameters now:

for x axis
OUTPUT_SCALE = 1.000 0.054
previously it was
OUTPUT_SCALE = 1.000 0.000

for y axis
OUTPUT_SCALE = 1.000 0.080
previously it was
OUTPUT_SCALE = 1.000 0.000

for z axis
OUTPUT_SCALE = 1.000 0.0732
previously it was
OUTPUT_SCALE = 1.000 0.000

now when i home each axis the machine cordinates display
0.000 mm ± 0.001 mm

I am thankfull for everyone who contributed in my post
regards

Asim

--- In CAD_CAM_EDM_DRO@yahoogroups.com, Jon Elson <elson@p...> wrote:
>
>
> Asim Khan wrote:
>
> >Hi Jon, Les, Paul, Ray and all other group members!
> >
> >I had checked the var file which is being loaded by my emc setup.
It
> >has all variables set to zero except 5220 variable which is set to
1
> >and i am still unable to figure it out whats wrong is happenging
with
> >my emc system?
> >
> >Folowing is the condition of GUI setting before i home:
> >Actual position display is selected
> >Machine Cordinates is selected
> >Default Units: mm
> >
> >When i press home button for any axis the axis goes to touch home
> >switch, touches it, then moves in opposite direction to search for
> >index pulse (i know this all is normal), when index signal is
found it
> >stops, turns the position counter of the axis homed to 0.0167 (or
any
> >other value if i have closed the software and re-launched the
program)
> >These offset values are always something under .1 mm.
> >Now when i selected the "Commanded" position display setting then
> >i could see the axis home position to exact 0.0000 mm
> >these offset values getting added in the actual position display
are
> >different for each axis. and are different when program is re-
launched
> >
> >
> What this seems to indicate is slightly low gain on the servo
drives.
> It seems to
> show that the difference between commanded position and actual
position
> is 0.0167 user units. In Inches, this is horrible. Actual
position
> will ALWAYS
> be different than commanded. But, this is a lot of error, as you
have
> the .ini
> file user units are set to inches. The only other
> possibility is that there is an offset in the .var file that is
putting
> in this
> offset from your last use of the system. This is the correct
behavior of
> the system. After homing, you should not look at machine position,
you
> should look a RELATIVE position. (I think you switch between these
> with the @ key.) Relative is the coordinate reference system the
machine
> moves to when axis movement commands are given, such as X 1.234
> If you check the relative actual position, I think you will find it
is
> within a couple encoder counts of zero after the home operation.
>
> (But, once you have set a work coordinate system, it should from
then on
> use that offset from the home position, as recorded in the .var
file, as
> your relative position.) You do this by right clicking on the axis
to
> be set, and entering a particular value, based on measurements from
> a fixture. Many people set the corner of a vise jaw as their
reference
> position. This is done with tools such as an edge finder.
>
> So, some random difference between machine and relative coordinate
> systems is no surprise. Machinists have no use for the machine
position,
> they want to set the zero point at something relative to the part
to be
> cut, not relative to the machine.
>
> >I've already tuned all the three axes. All the three axes jog upto
> >300ipm without any following error and without any much big noise.
> >my min_ferror value was 0.005
> >
> >Do you think i should set min_ferror to some very small value?
> >i know this value is in inches. what should be the value for
> >min_ferror for input_scale value of 250000 (10000CPR x 2.5 pitch)
> >my previous ecperience tells me if we lower min_ferror value then
the
> >tuning gets much difficult, axis speed reduces and we get
following
> >errors.
> >
> >
> No. The ferror parameters set a limit at which a following error
will be
> called. They have nothjing to do with tuning or servo performance,
only
> the difference between commanded and actual that will result in a
following
> error and the machine going into estop-reset mode. The value is
NOT in
> inches, unless you have declared your user units in the ini file to
be
> inches.
>
> >I was having my position counters counting in reverse so i had put
a
> >negative sign (-) for my input_scale values. is it permisible to
do
> >that? otherwise i had to swap my encoder A and B channels, so i
> >preffered the changes in ini.
> >
> >
> Absolutely OK. If the servos run away, then you have to change the
sign
> of the output scale as well. But, you are obviously past that.
>
> >One more thing i observed was this fact that when i commanded in
mdi
> >mode to do G01 X5.000 the same offset that was found in home
position
> >was added and i saw the target position as X5.0167 (relative,
actual
> >display)
> >
> >As i already told I have another machine with the same emc, stg2
> >model -II but with 800 CPR encoders and INPUT_SCALE = 4000, but i
> >never saw the offset value getting added in the actual position
> >display. i have min_ferror (0.005) for this machine same as on
this
> >problematic machine.
> >
> >
> Well, that's because no one ever set the axes on that other
machine.
> When they
> do, it will never have the same offset again, as it is remembered
in the
> .var file. But, the whole point of this is that the machinist
needs to set
> the work coordinates to be relative to the part, not relative to
the
> machine.
>
> Jon