Strange Trajectory Planner behaviour

Moderators: TomKerekes, dynomotion

gnrules
Posts: 11
Joined: Fri Jun 14, 2019 7:06 am

Re: Strange Trajectory Planner behaviour

Post by gnrules » Fri Mar 12, 2021 10:40 am

Hi Tom,
I added 4th decimal digit to my code and now it is very smooth without any dicontinuity even in limited velocity segments.
You were right about truncation/rounding error relative to decimal digits in g-code.

It was something that we would experiment because we use 0.5um glass scales, but now it is mandatory!!

Thank you for tour support.

Best regards

Giancarlo

GabrielR922
Posts: 74
Joined: Thu Aug 13, 2020 10:07 pm

Re: Strange Trajectory Planner behaviour

Post by GabrielR922 » Mon Mar 15, 2021 6:10 pm

Hi Tom and Giancarlo.

Let me see if I understood.

Even if I have a precision of ~0.01mm, it's best to post the code with 3 decimal cases? ig: 0.001?

Is there a tutorial to how plot the data like both you are doing?

Thanks,
Gabriel

User avatar
TomKerekes
Posts: 1515
Joined: Mon Dec 04, 2017 1:49 am

Re: Strange Trajectory Planner behaviour

Post by TomKerekes » Mon Mar 15, 2021 11:45 pm

Hi Gabriel,
Even if I have a precision of ~0.01mm, it's best to post the code with 3 decimal cases? ig: 0.001?
Yes it a good idea to have numerical resolution an order of magnitude better or more than the precision you are trying to achieve.

Is there a tutorial to how plot the data like both you are doing?
The charts I posted regarding the GCode segment lengths and angles were simply created by editing the GCode to remove non numeric characters (GXYZ) then pasting into Excel using text to columns. Then I calculated the Deltas. From deltas I calculated lengths and angles. Then deltas of those. I've attached it.


To capture data within KFLOP you might read the wiki here.
Attachments
GCodeNoise.xlsx
(39.9 KiB) Downloaded 7 times
Regards,

Tom Kerekes
Dynomotion, Inc.

Post Reply