Dynomotion

Group: DynoMotion Message: 10778 From: Colin Fera Date: 1/5/2015
Subject: kmotion question regarding DRO's
Some background: I am using kmotion CNC (version k from last week).  The machine is a CNC milling machine with 5 turn/inch ball screws and direct drive on all axis's using stepper motors.  The drives are setup to half step (200 step/rot motors).  This gives me a resolution of .0005", everything seems pretty accurate using a .0005 DTI the table will stop in the same place every time. Also I am not using encoders or any kind of feedback.

Here is the question:  After having zeroed a DRO axis if I jog the machine around a bit it will eventually end up displaying something like 2.3452 (etc). 

 Why would the Least significant digit ever be anything besides 0 or 5 ? 

Thanks,
Colin
Group: DynoMotion Message: 10779 From: Tom Kerekes Date: 1/5/2015
Subject: Re: kmotion question regarding DRO's
Hi Colin,

The trajectory motion is always computed and maintained to virtually infinite precision and then as a final operation the closest possible motor step is commanded to the motor.  The DROs report the theoretical desired destination.   The final stopping point is determined by many things such as exactly where and when the Jog key was released and the distance it takes to decelerate.  Neither are likely to be an exact multiple of the stepping distance.

Regards
TK

Group: DynoMotion Message: 10780 From: Colin Fera Date: 1/5/2015
Subject: Re: kmotion question regarding DRO's
Hi Tom,

Your answers are logical as always, but it seems like the DRO's should report the theoretical final stopping point instead of the theoretical desired destination.  

Also wouldn't the actual final destination need to be an exact multiple of the stepping distance (not accounting for backlash in the ball screws or slop in other drive components). I am under the impression that stepper motors are pretty good at repeating a stopping point at least when they are either full stepping or half stepping.

Please don't take these as complaints, this doesn't present any problems for me. The only time I actually notice is when I am getting stock centered in the mill.  When actually cutting parts it doesn't matter since the DRO's of course display the commanded position and the parts come out accurate to within my ability to measure.

Thanks,
Colin 
   

On Mon, Jan 5, 2015 at 3:27 PM, Tom Kerekes tk@... [DynoMotion] <DynoMotion@yahoogroups.com> wrote:
 

Hi Colin,

The trajectory motion is always computed and maintained to virtually infinite precision and then as a final operation the closest possible motor step is commanded to the motor.  The DROs report the theoretical desired destination.   The final stopping point is determined by many things such as exactly where and when the Jog key was released and the distance it takes to decelerate.  Neither are likely to be an exact multiple of the stepping distance.

Regards
TK

Group: DynoMotion Message: 10783 From: Tom Kerekes Date: 1/5/2015
Subject: Re: kmotion question regarding DRO's
Hi Colin,

I can see your point for purposes of measurement I suppose.  But rounding numbers to integer steps could cause a number of other problems.  Especially for large numbers of small moves. 

I think most operators do want to see the theoretical desired position rather than the final stopping point.  For example starting at zero and making 7 relative moves of 1.0001 they expect to end up as close as possible to position 7.0007.   If after each move the position rounded it would be confusing.  The response to a set of commands should be independent of internal things like the resolution of the machine that the Operator normally knows nothing about.

Consider another example with some random offset.  Imagine where two fixtures are known to be physically offset by exactly 10.0001 inches.  To round off DRO positions to 0.0005 would be incorrect.

Also consider a Metric job.

Most systems have a higher level of microstepping like 10X, 16X, 256X so even for measurement it isn't an issue.  Mechanical errors dominate over stepping resolution anyway.

You could probably change your "Center" program to do the rounding if you wish.

Regards
TK