RE: Angular axes

["Tim Goldstein" <timg-at-ktmarketing.com>]

The 4th axis problem is not a new revelation. I mentioned it to Paul
about 4 months ago and he did some digging into it and has found that
you can get better performance with it if you reduce the cycle time on
the angular axis. Even at that it is broke. I have also corresponded
with Art Fenerty about it and how he solved the issue with Master5
(Windows based derivative of EMC). I don't remember his comment exactly
so I will not give you the poor recollection I have. I did comment the
input to Paul and his response was that Art's theory seemed plausible.
To back it up Art's fix for his program solved the problem.

Tim
[Denver, CO]


> -----Original Message-----
> From: emc-at-nist.gov [emc-at-nist.gov] On Behalf Of Ray Henry
> Sent: Wednesday, July 03, 2002 9:55 PM
> To: Multiple recipients of list
> Subject: Re: Angular axes
> 
> 
> 
> 
> John
> 
> It seems to me that one of the problems may be the great 
> difference in the 
> number of steps between a unit on the linear axis and the 
> number of steps in 
> a rotary axis.  If you keep the quantity of steps per unit 
> nearly the same, 
> then you can drive the axes at the same speeds.
> 
> So long as you define a unit as 1 and make the INPUT_SCALE 
> the number of 
> pulses to complete a full revolution you will have the 
> programming problem 
> that you are experiencing.  In order to make a command of a90 
> be a 90 degree 
> move, you must set INPUT_SCALE to the number of steps to 
> complete a 1 degree 
> move rather than a 360 degree move.
> 
> I do remember something about the six axis implementation 
> being incomplete 
> when the new interpreter was integrated into the rest of the 
> EMC.  Your study 
> and reports of the problems that you are having will help.  Thanks 
> 
> Ray 
> 
> 
> 
> On Wednesday 03 July 2002 21:45, you wrote:
> > Hi Chris.
> >
> > I have determined this evening that it does not seem to make any 
> > difference whether the axis is configured as LINEAR or 
> ANGULAR it just 
> > plain does not work at any reasonable speed with the 
> INPUT_SCALE and 
> > OUTPUT_SCALE set to any thing less than 1000 steps per 
> unit.  It also 
> > appears to totally ignore the UNITS parameter on the fourth 
> axis.  You 
> > are correct in that if the fourth axis is configured identically to 
> > the other axis it will work at reasonable speeds but you 
> cannot tell 
> > it to move say 90 degrees other than by decimal numbers such as g0 
> > a0.25 will get you a 1/4 revolution move. This is not how I want to 
> > use my rotary table.  Oh well, I am sure it will get fixed 
> sooner or 
> > later.  Maybe I will have to brush up on my C skills and fix it may 
> > self, which is NOT what I want to do.  I want to build engines, not 
> > fix software.
> >
> > John Guenther
> > Sterling, Virginia
> >
> > > -----Original Message-----
> > > From: emc-at-nist.gov [emc-at-nist.gov]On Behalf Of Chris Wagner
> > > Sent: Wednesday, July 03, 2002 11:29
> > > To: Multiple recipients of list
> > > Subject: Angular axes
> > >
> > >
> > >
> > > Greetings,
> > >
> > > There has been list traffic about various problems with angular 
> > > axes, so I thought I would try some tests.  I will 
> eventually (very 
> > > long term) want to use a rotary table.
> > >
> > > I found a problem similar to J Guenther - very slow motion.  I 
> > > played around with my .ini and found that the problem 
> occurs anytime 
> > > there is an angular axis (any of X R, X Y R, X Y Z R, X Y 
> Z R P W).  
> > > The slow motion also effects the linear axes - they also creep 
> > > along.  I would expect that adding a fourth axis would not effect 
> > > the behavior of the previous three, but it does.
> > >
> > > As far as .ini understanding or parsing issues go, I did 
> get a setup 
> > > that would run all the axes fast, by simply copying my three 
> > > existing linear axes sections (the [AXIS_0] etc) to three new 
> > > sections, changed the new [AXIS_x] tags to sequential 
> numbers, used 
> > > AXES=6, COORDINATES=X Y Z R P W in the [TASK] section.  I 
> did _not_ 
> > > set the three new axis type to angular (or make other 
> changes).  The 
> > > front ends (Tkemc, yemc) then show all 6 axes (XYZRPW), 
> and all axes 
> > > will clock at high speed. Moves work with e.g., 
> G1X10Y20Z30A20B30C50 
> > > ...  (also works with XYZR)
> > >
> > > Of course It would be hard to set steps/revolution or 
> degree since 
> > > the axis section is setup for linear units...
> > >
> > > Tentative conclusions:
> > > 1) Problem is in setup/understanding of ANGULAR .ini 
> configuration.
> > >
> > > 2) Something is wrong with the emc parsing of the ANGULAR ini.
> > >
> > > --
> > > Recreational Calculus - Just For Fun!
> > >
> > > Chris Wagner
> > > clwagner-at-eecs.wsu.edu
> 
>