Dynomotion

Group: DynoMotion Message: 3979 From: Albert Płatek Date: 2/20/2012
Subject: Trajectory Planner Bug
Hi Tom,
 
I found bug in TP, when execute small arc 0.005in (radius), motors move very violently, it can cause motor stall on machining complex shapes.
It seem that when arc is very small and smoothed, TP not respect accleleration limit.
In attachement screenshot of my stettings and tp_bug.ngc file.
 
btw. I dont know why but I cant capture this by C program.
 
Tested on v4.28 and 4.29z with same result.
 
Best Regards
Albert Płatek
   @@attachment@@
Group: DynoMotion Message: 4022 From: albertplatek Date: 2/21/2012
Subject: Re: Trajectory Planner Bug
Hi Tom,
Do you have same error as me?
Or its not a bug?

Best Regards
Albert P³atek

--- In DynoMotion@yahoogroups.com, Albert P³atek <albertplatek86@...> wrote:
>
> Hi Tom,
>
> I found bug in TP, when execute small arc 0.005in (radius), motors move
> very violently, it can cause motor stall on machining complex shapes.
> It seem that when arc is very small and smoothed, TP not respect
> accleleration limit.
> In attachement screenshot of my stettings and tp_bug.ngc file.
>
> btw. I dont know why but I cant capture this by C program.
>
> Tested on v4.28 and 4.29z with same result.
>
> Best Regards
> Albert P³atek
>
Group: DynoMotion Message: 4023 From: Tom Kerekes Date: 2/21/2012
Subject: Re: Trajectory Planner Bug
Attachments :
Hi Albert,
 
Yes I can duplicate the problem.  Haven't understood it yet.
 
If you want a laugh check out the attached trajectory that is being generated.
 
Regards
TK
 
 

Group: DynoMotion Message: 4025 From: albertplatek Date: 2/21/2012
Subject: Re: Trajectory Planner Bug
Hi Tom,
Its really weird trajectory, more weird is that it aslo happen
sometimes at 1 degree of "facet angle" on complex shapes.
This error is soo big that cant be smoothed by Low Pass Filter.

At moment I disabled "Corner Rounding" feature,I only use Low Pass Filter smoothing.

Best Regards
Albert P³atek

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Albert,
>  
> Yes I can duplicate the problem.  Haven't understood it yet.
>  
> If you want a laugh check out the attached trajectory that is being generated.
>  
> Regards
> TK
>  
>  
>
> From: albertplatek <albertplatek86@...>
> To: DynoMotion@yahoogroups.com
> Sent: Tuesday, February 21, 2012 11:15 AM
> Subject: [DynoMotion] Re: Trajectory Planner Bug
>
>
>  
> Hi Tom,
> Do you have same error as me?
> Or its not a bug?
>
> Best Regards
> Albert P³atek
>
> --- In DynoMotion@yahoogroups.com, Albert P³atek <albertplatek86@> wrote:
> >
> > Hi Tom,
> >
> > I found bug in TP, when execute small arc 0.005in (radius), motors move
> > very violently, it can cause motor stall on machining complex shapes.
> > It seem that when arc is very small and smoothed, TP not respect
> > accleleration limit.
> > In attachement screenshot of my stettings and tp_bug.ngc file.
> >
> > btw. I dont know why but I cant capture this by C program.
> >
> > Tested on v4.28 and 4.29z with same result.
> >
> > Best Regards
> > Albert P³atek
> >
>
Group: DynoMotion Message: 4028 From: Tom Kerekes Date: 2/21/2012
Subject: Re: Trajectory Planner Bug
Attachments :
Hi Albert,
 
The problem is caused when multiple motion segments are created less than one servo sample time in duration (less than 90us).  That 5 mil radius circle only takes about 0.1sec with your high acceleration rates.  With 1/2 degree facets those are really tiny and fast.  Also each facet needs two segments for a bit of constant speed and a bit of acceleration.  KFLOP's DSP code was changed to handle sub-sample segments.  It now skips over segments to properly get to the next sample time - essentially discarding some segments.  A better solution would be to discard them earlier in the Trajectory Planner.  Or to be honest I'm not sure the "Facet Angle" concept makes sense for all radius sizes (or rather curvature rates).  But for now that appears to be a working solution.  See the attached trajectory is now correct for your settings.  Here is a link for a patched firmware version.  I will report as V4.29Z (capital Z) and is compatible with V4.29z PC code.  Copy it to your KMotion4.29z\DSP_KFLOP directory. 
 
 
Thanks again for the clear example showing the problem.
 
Regards
TK