RE: Using Makefiles / Homing

[Kurtis Martineau <klmartineau-at-shaw.ca>]

> -----Original Message-----
> From: emc-at-nist.gov [emc-at-nist.gov]On Behalf Of Ray
> Sent: Tuesday, January 22, 2002 11:03 PM
> To: Multiple recipients of list
> Subject: Re: Using Makefiles
>
>
>
>
>
> Kurtis wrote:
>
> This includes snips from several posts by Kurtis.  As long as I've got my
> foot in my mouth already, I may as well chew on it.
>
> >I tried motion at 6.0 IPM - Index pulse goes from 0 to 4.5v for about
> >1.0 msec
>
> > I'm using BDI2-09
> ..
> >; Version of this INI file
> >VERSION =               $Revision: 16.Jan.02 $
> ..
> >HOMING_VEL =                    0.1
>
> Since homing velocity is in ips and your test of the pulse
> duration used 6
> ipm, you are dealing with a 1 msec window to see that pulse.  Rather than
> reducing the axis cycle time, I'd reduce the homing velocity to a snail's
> pace/10 so that the index pulse is up there for several loops of
> the cycle
> time.  And since you do not home very often, the wait should not be as
> detrimental to your system as trying to run with real small cycle times.
>
> HTH
>
> Ray
>
>
Ray, the rate of HOMING_VELOCITY that I reduced to is 0.1inch per
second...to me that is pretty slow.  At this velocity the time high for my
index pulse is about 0.001 secs...to me this is a long time for an index
pulse to be on for.
>
>
>>snipped from Steve's email:
   I hate to speculate, but you sound desperate.  Is it possible that the
index pulse from the encoder and the the home switch actuation are aligned
almost perfectly (by chance) causing them happen at almost the same
instant, thereby fouling the logic in EMC?  I'm not familiar enough with
the logic to know if this is possible, but perhaps EMC looks for a state
change of the index pulse input after the home switch input, but by the
time it's reversed direction and starts looking - it's already happened.  I
have seen CNC machines that will occasionally home to a position out of
"true" home by one turn of the lead screw for this same reason.
>>

It is common to set up the encoder so that the index pulse happens at around
180 degrees after the home switch is released.  My problem is happening on
all three axis.  I have since recompiled with #define NO_INDEX_PULSE.  At
least I can home the machine, although homing from a limit switch is not
nearly as accurate as an index pulse. I hope to test Max's diagnostic to
see what is really happening here.


Some hopefully helpful data:

Homing velocity = 0.1 inch per second (could set any - this just seemed slow
enough)
Encoder is really a resolver and a resolver to encoder converter.
Encoder resolution = 4096 pulses per revolution of shaft.
Encoder coupling = 10:1 through a rack and pinion

Therefore: feedback resolution is 40960 pulses per inch.
		Index pulses = 10 per inch or 1 every 0.1 inches
		Index pulse rate = 1 pulse every second


Thanks for all the feedback,

Kurtis