Dynomotion

Group: DynoMotion Message: 1716 From: albertplatek Date: 8/21/2011
Subject: KmotionCNC Trajectory Planner Corner Rounding
Hi Tom,

When smoothing complex 3d toolpath with 0.5 Facet Angle,
often get message: Unexpected Coordinated Motion Buffer Underflow

What is the reason of this? Is it KFLOP limitation? (I tested same path with 10sec buffer, and it works much longer
but path creation is still slower than needed machining speed)
Should I use 1.0 Facet Angle? or higher?

When KmotionCNC is smoothing toolpaths my pc CPU usage is 2-3%.

Best Ragards
Albert Płatek
Group: DynoMotion Message: 1718 From: Tom Kerekes Date: 8/21/2011
Subject: Re: KmotionCNC Trajectory Planner Corner Rounding
Hi Albert,
 
I assume you have the patch for corner rounding that you tested and is described here:
 
 
 
I'll assume so, then there may be an issue with the USB download rate.  USB throughput on most systems allows about 1000 motion segments per second to KFLOP so the average rate must be below this.
 
Hmmm I guess that places a limit on the average "curvature" rate we can handle.  A circle at 0.5 degree segments would have 720 segments so for example s full circle continuously every ~ 0.72 seconds would be a problem. 
 
Is is repeatable? Can you provide some example gcode and what your velocity and acceleration settings are?  Also how many KFLOP threads are running?
 
Increasing the facet angle should help at the expense of "smoothness".   Increasing the Buffer Time should not change the machine speed at all.  But it will pause before each path begins as more data needs to be downloaded before starting.  Is this what you observe?
 
Thanks
TK
 
 
 

--- On Sun, 8/21/11, albertplatek <albertplatek86@...> wrote:

From: albertplatek <albertplatek86@...>
Subject: [DynoMotion] KmotionCNC Trajectory Planner Corner Rounding
To: DynoMotion@yahoogroups.com
Date: Sunday, August 21, 2011, 6:23 AM

 
Hi Tom,

When smoothing complex 3d toolpath with 0.5 Facet Angle,
often get message: Unexpected Coordinated Motion Buffer Underflow

What is the reason of this? Is it KFLOP limitation? (I tested same path with 10sec buffer, and it works much longer
but path creation is still slower than needed machining speed)
Should I use 1.0 Facet Angle? or higher?

When KmotionCNC is smoothing toolpaths my pc CPU usage is 2-3%.

Best Ragards
Albert P&#322;atek

Group: DynoMotion Message: 1721 From: albertplatek Date: 8/22/2011
Subject: Re: KmotionCNC Trajectory Planner Corner Rounding
Hi Tom,
My settings:
Break Angle 20 degrees
Look ahead 4 seconds
Collinear Tolerance 0.0005 in
Corner Tolerance 0.003 in
Facet Angle 0.5 degrees
Accel X 18
Accel Y 18
Accel Z 15
Feedrate F2000
1 thread is running

I like speed :}

Gcode example (447kB) http://pastebin.com/GidscDD6


If I set set smaller break angle like 15 degrees and Facet Angle 1.0
it works fine.

Best Regards
Albert Płatek




--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Hi Albert,
>  
> I assume you have the patch for corner rounding that you tested and is described here:
>  
> http://tech.groups.yahoo.com/group/DynoMotion/message/1275
>  
>  
> I'll assume so, then there may be an issue with the USB download rate.  USB throughput on most systems allows about 1000 motion segments per second to KFLOP so the average rate must be below this.
>  
> Hmmm I guess that places a limit on the average "curvature" rate we can handle.  A circle at 0.5 degree segments would have 720 segments so for example s full circle continuously every ~ 0.72 seconds would be a problem. 
>  
> Is is repeatable? Can you provide some example gcode and what your velocity and acceleration settings are?  Also how many KFLOP threads are running?
>  
> Increasing the facet angle should help at the expense of "smoothness".   Increasing the Buffer Time should not change the machine speed at all.  But it will pause before each path begins as more data needs to be downloaded before starting.  Is this what you observe?
>  
> Thanks
> TK
>  
>  
>  
>
> --- On Sun, 8/21/11, albertplatek <albertplatek86@...> wrote:
>
>
> From: albertplatek <albertplatek86@...>
> Subject: [DynoMotion] KmotionCNC Trajectory Planner Corner Rounding
> To: DynoMotion@yahoogroups.com
> Date: Sunday, August 21, 2011, 6:23 AM
>
>
>  
>
>
>
> Hi Tom,
>
> When smoothing complex 3d toolpath with 0.5 Facet Angle,
> often get message: Unexpected Coordinated Motion Buffer Underflow
>
> What is the reason of this? Is it KFLOP limitation? (I tested same path with 10sec buffer, and it works much longer
> but path creation is still slower than needed machining speed)
> Should I use 1.0 Facet Angle? or higher?
>
> When KmotionCNC is smoothing toolpaths my pc CPU usage is 2-3%.
>
> Best Ragards
> Albert Płatek
>