Re: homing and axis errors



Hi Ray,

I guess I should begin by explaining that I never did
manage to get the loop 'closed'.  It seems that the
encoder portion is working because I can eventually
get it to the point where when I turn the shaft, the
display reflects the position change.  The DAC/Amp
portion do not work yet.  The hardware definatly does,
I know this because firstly it works under a small dos
driver I wrote, and secondly because if I plug in my
amp before I start EMC, the motor turns at full speed
due to the uninitialized logic.  As soon as I start
EMC, the shaft stops turning, because in the
initialization dispatcher function I output zero drive
to the motor.  So I'm sure that my initialization
fuction and write DAC function work, but have yet to
figure out why a large position error does not cause
the DAC value to change.  I am using 100 for
porportional gain, and 0 for everything else.

--- Ray <rehenry-at-up.net> wrote:
> 
> On Thu, 14 Sep 2000, you wrote:
> > Hello,
> > 
> > I have been playing around with my little homebrew
> > setup some more, and I learned something curious. 
> I
> > once saw that I could home my single axis, and it
> > would not respond to jog changes, but it would
> show
> > position changes on the EMC display when I turned
> the
> > shaft.  
> 
> I'd guess that the EMC was still trying to home the
> axis and was ignoring
> your jog commands.

I think you are right.  What I don't understand is
that at startup, both commanded and actual position
are zero.  I also uncommented the #define in my hacked
up version of stg.c that specifies no home switch
(NO_HOME_SWITCH or something like that).  Could it be
that the task controller keeps looking for the change
of state of the home switch even though I uncommented
out that line?  I guess that I can force the function
to output a 'homed' state?



> go,
> > it hits 20.000 then drops.  Eventually I get the
> > following error message:
> 
> You can be thankful that you didn't define a 100'
> machine in the ini or
> you'd have been there all night watching the thing
> run to the end.. ;-)

:)  Yup!  


> 
> > minimilltaskintf.cc 938 Error on Axis 0
> > Maximum software limit on axis 0 exceeded
> 
> This means that the EMC never saw the home switch
> inputs as closed.  If
> you define them as closed in the ini file, It should
> always see them as
> closed.  

I'll give that a shot.


> 
> Is this homebrew axis a closed loop?  In other words
> does the EMC drive a
> motor that drives the the encoder.  If it is not,
> you will rapidly
> confuse the EMC when you command a position change
> because it is watching
> the encoder position and expecting to see the change
> that it commanded.

It should be a closed loop :)  
Still need to work on that though.  I guess that what
I was trying to do was to just get this thing working
to the point where I could start EMC, click 'home',
rotate the shaft, and see the display update.  Then I
was thinking I'd look into why my output doesn't work.
After reading your response, I'm starting to realize
why it wouldn't work.  Dont know why, but I thought
that the 'home' button might cause the current
position to become zero (home), but that's stupid,
because what would be the point of having a home
switch?

 
> If this is a closed loop I'd guess that you do not
> have tight enought
> control of the loop.  You need to do some tuning. 
> If it is not a closed
> loop, I'd rig some kind of  connection so that the
> motor drives the
> encoder, set up emc so that it knows distance, and
> start tuning.

The encoder is actually mounted to the motor, so when
I fix the output, it should be closed loop.  Distance
in my case is meaningless, as the shaft does not
actually turn anything (except the encoder).  My plan
is to figure out EMC, then build a better multi-axis
servo controller board.  


Anyhow, it looks like I have some work to do.

Thanks very much Ray,
Garrett.

__________________________________________________
Do You Yahoo!?
Yahoo! Mail - Free email you can access from anywhere!
http://mail.yahoo.com/



Date Index | Thread Index | Back to archive index | Back to Mailing List Page

Problems or questions? Contact