Re: Encoder and P of PID

[Enrique Sanchez Vela <enrique_sanchez_vela-at-yahoo.com>]

Gene, Ray,

 thanks for the all info I'll take a closer look at it.

  regarding the X encoder the color code did not match, but it is
printed on the encoder, thanks anyway. the problem is with the Y
encoder, it gives the info by pin numbers and the guys at TREE or
somewhere else took the time to unplug the Z axis, I took a guess
following the other cables and it seems to count fine, if I move the
table manually.

guys, I really appreciate your help, the table's been moving even the
controller is not tuned at all.

what worries me, is accuracy, although the computer says the table
moved 2" the table really moves a little more, it is ok within 1" but
when I move it 5" it goes way out range. the problem is that it does
not keep constant. also the X encoder keeps failing to keep counting
and the table starts moving on its own.


thanks, 
esv.
I'll keep you posted.


--- Ray <rehenry-at-up.net> wrote:
> 
> Enrique 
> 
> There are a couple of pages in the handbook that you could look at
> here.
> 
> www.linuxcnc.org/handbook/part2/pidtune.html
> www.linuxcnc.org/handbook/part2/rhemcini.html
> 
> (comments added on the ini file)
> 
> On Thursday 18 October 2001 05:00, Gene wrote:
> > Hi Enrique,
> >
> >      I have shuffled my comments in with yours.  I also sent a
> carbon
> > copy to Ray to keep him informed so he can give you any help with
> the
> > EMC part.
> 
> <s>
> >      The PID is set somewhere in the EMC, and I don't know how to
> do it
> > there.  At this point the "I" and "D" of the PID should be set to
> so
> > that it doesn't have any effect on the "P".   Turned off, set to
> zero or
> > whatever.  The "I" and "D" will be adjusted later.
> 
> The tuning values are set for each axis using the variables as shown 
> below.  
> 
> ; Trajectory planner section 
> --------------------------------------------------
> [TRAJ]
> . . .
> DEFAULT_VELOCITY =      0.0167
> MAX_VELOCITY =          1.5
> DEFAULT_ACCELERATION =  5.0
> MAX_ACCELERATION =      5.0
> 
> The traj stuff works for all of the axes.  I'd start with real low
> values 
> here so that you can begin to move without unexpected things
> happening.
> 
> ; First axis
> [AXIS_0]
> . . .
> MAX_VELOCITY =                  1.2
> P = 1000.000
> I = 0.000
> D = 0.000
> FF0 = 0.000
> FF1 = 0.000
> FF2 = 0.000
> . . .
> BIAS = 0.000
> . . .
> FERROR = 1.000
> MIN_FERROR = 0.010
> 
> When the axis is under EMC control, BIAS is the zero setting for the
> PC 
> and the STG board.  You won't see much difference here if the encoder
> 
> position loop is closed because position will override idle. 
> 
> If your axis has a large amount of jitter or shaking when first
> started, 
> you can set Ferror and min_ferror large > 10 so that you do not
> always 
> trip out on following error.
> 
> The EMC will always attempt to accelerate an axis to commanded
> velocity at 
> the acceleration rate.  This ramp takes priority over PID.  What PID
> will 
> do is help the motor move the axis to speed without undue overshoot
> or 
> undershoot or oscillation. 
> 
> I would set velocity and accel first and then start up the EMC.  Once
> it 
> is running and has some control over an axis, I would use the
> calibration 
> popup under the settings menu and change the values there.  These
> values 
> are reset when you press "OK" so you can test jog the axis and see
> how it 
> responds to the change.
> 
> I make a paper list of the existing values.  Make a small change + or
> - 
> and write that value in.  Jog the axis back and forth and make a 
> subjective judgement about better or worse or no difference.  Like
> Gene 
> says start with P.  Set it low.  If it is to low, the axis will run
> away 
> or not respond.  Increase it until you get the expected result when
> you 
> jog.  Keep changing the value until you find the axis to be
> overshooting 
> or ringing.  Now back it off until you reasonably smooth motion.
> 
> Now shut down the EMC and it will write the changes you have made
> with 
> calibration in the ini.  Change the acceleration and max velocity a
> bit 
> and run it again.  Repeat this process until you are reasonable happy
> with 
> the axis' performance.  
> 
> Now you can add it a little I and D and FF1 or vel FF gain.  Your
> mill is 
> not really heavy enough to need much here.  Jon Elson uses about 8.0
> in 
> the vel FF gain for his bridgeport and is very happy with the result.
> 
> Hope this helps.
> 
> Ray
> 
> 
> 


__________________________________________________
Do You Yahoo!?
Make a great connection at Yahoo! Personals.
http://personals.yahoo.com