FTMotion: Fix K unit and Improve linear advance

[dbuezas]

Description

FTMotion had K defined in a different unit (100 times that of the normal motion system, klipper, etc)

This PR aligns that, and also fixes problems during (de)retraction: these blocks shouldn't use linear advance.

Furthermore, I simplified the math by using the integral of the linear advance equation used in the classic motion planner, which has the advantage of not losing track of the integration constant between blocks due to instantaneous changes in velocity (i.e, linear advance is now jerk aware)

Equivalence of rate base and position based linear advance

1. Typical formulation in terms of rates and acceleration

$$ e_{\text{rate}}(t) \;=\; e_{\text{planned\_rate}}(t) \;+\; k\,e_{\text{acc}}(t) $$

2. Integrate both sides over time

$$ \int e_{\text{rate}}(t)\,dt \;=\; \int e_{\text{planned\_rate}}(t)\,dt \;+\; k\!\int e_{\text{acc}}(t)\,dt $$

3. Using the definitions

$$ \int e_{\text{rate}}\,dt = e_{\text{pos}}, \qquad \int e_{\text{planned\_rate}}\,dt = e_{\text{planned\_pos}}, \qquad \int e_{\text{acc}}\,dt = e_{\text{vel}}, $$

4. we get

$$ e_{\text{pos}}(t) \;=\; e_{\text{planned\_pos}}(t) \;+\; k\,e_{\text{vel}}(t) \;+\; C. $$

5. Where the integration constant $C$ is zero because the print starts with no nozzle pressure. So finally:

$$ \boxed{\,e_{\text{pos}}(t) = e_{\text{planned\_pos}}(t) + k\,e_{\text{vel}}(t)\,} $$

Requirements

FTMotion

Benefits

Same unit, simpler code, no drift, better pressure management

Configurations

Related Issues

[narno2202]

@dbuezas , don't forget to update the motion menu accordingly. I've tested your FT_MOTION code in the last Marlin bugfix, great work, nice improvement in sound.

[dbuezas]

Thanks!
Do you mean the maximum value or what would be missing in the menu?

[narno2202]

I mean the maximum value and the display to be aligned with the new value. float32 instead of float62 should be enough in the menu item.

[dbuezas]

Got it, good catch. Maybe i even do more decimals

[Crazy-Charles]

Any chance of fixing the output from M493?
X base dynamic shaper frequency: 37.00Hz scaling: 0.00Hz/mm
Y base dynamic shaper frequency: 37.00 Hz scaling: 0.00Hz/mm

[dbuezas]

Just realized i can reuse the use_advance_lead from the block

[dbuezas]

@Crazy-Charles Yes, will be fixed in a different PR.
@narno2202 Menues fixed, and also reused the use_advance_lead calculated in the planner.

It's ready for review & merge