Request - Axislabeling, Homing, infinite rotational C, AC-Kinematics, switching between mcs and g54

[NicoDetzler]

Hi everyone!

First of all, thank you for developing grblHAL and making this open source.

Information on my build:
-T41u5xbb Breakoutboard by Phil Barrett
-Teensy 4.1
-Closed loop stepper hybrid
-Ballscrews, 4mm pitch for linear axes
-Harmonic drive 30:1 for rotational axes
-Machine Zero is front,upper left corner

AXIS LABELING: So I built a 5 axis machine with a table-table A-C configuartion, where A rotates around X-Axis and C around the rotated Z'-axis.
My problem is, that the breakoutboard demands for a B axis instead of C. Is there any way to rename the B-axis to C?

HOMING: Further I got some issues with homing. Homing for X,Y and Z is perfectly working and sets the values of the machinecoordinatesystem to X0,Y0 and Z0. My Problem here is the A-Axis triggers the Switch at a value (-45,02) degrees and sets 0. But the real 0 position should be 0 degrees not -45,02 degrees. Is there a possibility to set the homing values for the machine coordinate system (mcs) via $ parameters for each axis ?
x_h_mcs = 0, y_h_mcs = 0, z_h_mcs = 0, a_h_mcs = -45.02, c_h_mcs = 0

INFINITE ROTATIONAL C: In my setup, the C axis is able to turn infinite roatations. Therfore I would like to set the angles between 0 and 360. Every higher or lower angle coming from the gcode, should be converted to that interval.

AC-KINMATICS: For the kinematics I wanted a variable Kinematic_enable. When a M-code (M850/M851 or #TRAFO_ON/#TRAFO_OFF) or a switch in ioSender is triggered, Kinematic_enable should be toggled. For the kinematics definition I would like to have the machinecoordinatesystem (mcs) properly homed, so X0,Y0,Z0,A0,C0 is the real machine position. Because with that I am able to use transformation matrizes to calculate new X,Y and Z values for when Transformation is activated.
What the Transformation should do: When activated and there is no change in angle A or C, it directly calculates the new X,Y and Z values.
When there is an angle change eg A0° to A30° or B34° to B129°, the kinematics should calculate, and interpolate the curve. The segmentlength of interpolation should also be settable through the $ Parameters.

In my Case G54 sits on the Table and rotates with A and C. So my G54 is not directly vectorially added to MCS, it sits on the Table.
The vector chain is:
MCS(0,0,0) -> Sway-axis-center-A(A_X,A_Y,A_Z) -> Turn-axis-center-C(C_X,C_Y,C_Z) -> G54(G54_X,G54_Y,G54_Z)->target

With those vector values, and the given angles A (alpha) and C(theta) and the target vectors added to G54 out of CAM, the transformed values for X Y and Z should be:

X_trans = A_X + (G54_X+target_x)⋅cos(theta) - (G54_Y+target_y)⋅sin(theta)+C_X
Y_trans = A_Y + ((G54_X+target_x)⋅sin(theta) + (G54_Y+target_y)⋅cos(theta)+C_Y)⋅cos(apha) - ((G54_Z+target_z) +C_Z)⋅sin(alpha)
Z_trans = A_Z + ((G54_X+target_x)⋅sin(theta) + (G54_Y+target_y)⋅cos(theta)+C_Y)⋅sin(alpha) + ((G54_Z+target_z) + C_Z)⋅cos(alpha) + Tool_length

I managed to calculate and manipulate GCode for a simultaneous movement with excel.
The gcode with transformation on should have the same effect as in the video:
G54 (sits 3mm above screw)
G0 X0Y0
G0Z0
M851 or #TRAFO_ON (switching on transformation kinematic)
G0 A90C180
G0 A-45C360
G0 A0C540
G0 C0
M850 or #TRAFO_OFF (switching off transformation kinematic)
#MCS_ON
G0 Z0
G0 X0Y0
#MCS_OFF

20260614_173148.mp4

SWITCHING BETWEEN MCS AND G54: Is there a way to switch between machine coordinate system (mcs) and the active coordinate system. So for a easy way to get the z-axis out of the way i woould like to use a gcode like #MCS_ON G0 Z0

I tried to study the grblHAL code. Sadly I'm at the end of my knowledge to manipulate the code myself.

[terjeio]

Is there a possibility to set the homing values for the machine coordinate system (mcs) via $ parameters for each axis ?

Yes, but only via a plugin. I have started one but it is not finished.

Is there any way to rename the B-axis to C?

Yes, in the Web Builder axis letters can be selected by clicking on this button:

Or by adding a symbol/macro to the compilation, e.g:

#define AXIS4_LETTER 'C'

Note that this is a new feature, not extensively tested.

INFINITE ROTATIONAL C ...

Not possible as of now, I have no idea about how to implement this so a detailed spec is required to check if it is possible.

AC-KINMATICS ...

This can be implemented as a kinematics plugin. #-commands are not possible, use custom M-codes or $-commands instead.

SWITCHING BETWEEN MCS AND G54 ...

Use G53 to command motion in MCS.

[NicoDetzler]

Thank you so much for your very fast answer @terjeio !
I could have thought about the G53..shame on me.
I'll try the settings asap :)

[NicoDetzler]

I tried to use the WebBuilder, but I got this error:

I got the same error in Arduino IDE when I tried to put #define axis4_letter 'c' in config.h

[terjeio]

Line 2539 has a typo, it references b instead of c in the limit_fei struct. I'll fix it in the next commit.

I got the same error in Arduino IDE when I tried to put #define axis4_letter 'c' in config.h

You have to put the definition at the start of nuts_bolts.h, or as a compiler command line option (-D) which the Arduino IDE does not allow...

I also had to remove the PROGMEM decoration from set_axis_mask() and get_axis_mask() in settings.c since I got a section type conflict, I do not know why yet.

[terjeio]

Fixes committed.

[NicoDetzler]

Tested with the Web builder - works great! :) My B-axis is now a C-Axis. Thank you! ♥

[spxwh]

Your requirement is actually that of a turn-mill compound lathe. You must solve two issues: the program for homing to the nearest position and recalculating the phase after each use of M5. You need to understand this program in thorough detail.

[spxwh]

Does the AC-KINEMATICS you mentioned refer to the RTCP algorithm?

[NicoDetzler]

Yes! The algorithm mentioned should be called rtcp in some cnc controllers.
Beckhoff TwinCat calles that function #TRAFO ON/OFF

My machine has a typical 5 axis table-table configuration like Hermle machines. On my machine the spindle is not yet installed.
I hope I can calculate the distances from g54 (and other coordinatesystems) to my real turning centers of pivot A and turn axis C.
There are several ways to calculate the transformed x y and z values since the angles A and C do come from the GCode. Either with quarternions or matrices multiplication.
I'll try to implement the kinematics...since i got the flashing via Arduino IDE to work properly.

I'm working my way through the code. Trying to understand what's happening where.

[Bharathk-Gowda]

@NicoDetzler
Dear sir, i noticed above shared video, is it running through GrblHAL firmware? if yes, kindly share $i of your firmware, how to enable RTCP function through web builder.

Thanks & Regards
Bharath

[NicoDetzler]

Hi @Bharathk-Gowda

It is running on GrblHAL. But the movements seen in the video are made with GCodes manipulated by Excel and some Transformation matrizes. I'm working on the code right now to implement RTCP, so the machine can read and process GCode normally exported via CAM. I will give an update when it's working. My programming is not the best so it can take a little time.

My plan so far is:

-M851 to enable RTCP, M850 to disable, M852 to calculate the distances from Center of turning axis C to active workcoordinatesystem with the actual machine angles.(mainly for measuring in a part with tilted table)

-Catch the first incoming GCode, interrupt USB stream, test if there is change in angle, when RTCP is enabled, if there is -> interpolate the curve with parameter s[0,1] and keep the real distance between 0,2 mm... then send the transformed GCode of the many segments to the parser/planner.
If there is no change in angle, directly calculate the transformed target point as GCode. Then start USB stream again.

As I tried using the kinemstics API, at a single rotaion of A or C, or combined, the axis mask locks XYZ movements, because it sees no delta in XYZ. That's why I'm trying to manipulate the incoming GCode. (I'm also maybe a little to stupid for kinematics API and how to use it)

As mentioned I'll give an update

[karoria]

Hi @Bharathk-Gowda

It is running on GrblHAL. But the movements seen in the video are made with GCodes manipulated by Excel and some Transformation matrizes. I'm working on the code right now to implement RTCP, so the machine can read and process GCode normally exported via CAM. I will give an update when it's working. My programming is not the best so it can take a little time.

My plan so far is:

-M851 to enable RTCP, M850 to disable, M852 to calculate the distances from Center of turning axis C to active workcoordinatesystem with the actual machine angles.(mainly for measuring in a part with tilted table)

-Catch the first incoming GCode, interrupt USB stream, test if there is change in angle, when RTCP is enabled, if there is -> interpolate the curve with parameter s[0,1] and keep the real distance between 0,2 mm... then send the transformed GCode of the many segments to the parser/planner.
If there is no change in angle, directly calculate the transformed target point as GCode. Then start USB stream again.

As I tried using the kinemstics API, at a single rotaion of A or C, or combined, the axis mask locks XYZ movements, because it sees no delta in XYZ. That's why I'm trying to manipulate the incoming GCode. (I'm also maybe a little to stupid for kinematics API and how to use it)

As mentioned I'll give an update

@NicoDetzler It is interesting. I am waiting for RTCP. Great initiative by you. Thanks.

[Bharathk-Gowda]

Hi @Bharathk-Gowda

It is running on GrblHAL. But the movements seen in the video are made with GCodes manipulated by Excel and some Transformation matrizes. I'm working on the code right now to implement RTCP, so the machine can read and process GCode normally exported via CAM. I will give an update when it's working. My programming is not the best so it can take a little time.

My plan so far is:

-M851 to enable RTCP, M850 to disable, M852 to calculate the distances from Center of turning axis C to active workcoordinatesystem with the actual machine angles.(mainly for measuring in a part with tilted table)

-Catch the first incoming GCode, interrupt USB stream, test if there is change in angle, when RTCP is enabled, if there is -> interpolate the curve with parameter s[0,1] and keep the real distance between 0,2 mm... then send the transformed GCode of the many segments to the parser/planner. If there is no change in angle, directly calculate the transformed target point as GCode. Then start USB stream again.

As I tried using the kinemstics API, at a single rotaion of A or C, or combined, the axis mask locks XYZ movements, because it sees no delta in XYZ. That's why I'm trying to manipulate the incoming GCode. (I'm also maybe a little to stupid for kinematics API and how to use it)

As mentioned I'll give an update

@NicoDetzler WOW, if your plane executed as you think, Grbl HAL will be 5 axis simultaneous CNC controller, Great initiative by you sir, I will be waiting for it.