Inconsistent height when moving with constant Z values.

[franz6ko]

Hi!

I'm using this library for a project involving a delta robot and computer vision. I'm trying to move the end effector on a horizontal plane by maintaining a constant z and giving different xy commands. However, it descends the more it gets farther from the center.

At first, I thought it was something about my code or the servos calibrations but I did tons of tests and the behavior is the same. So I did a bit more research on the web and I found a repo of @isaac879 with inverse kinematics functions for a delta robot. I tried those out and it works like a charm! Now my robot moves perfectly well on a horizontal plane.

So maybe there's something wrong with the code or the mathematical approach. I wanted to notice this as this is a well-known library and it would be great to have it fixed.

Link to the repo: https://github.com/isaac879/Delta-Robot
(check out the delta_robot/deltaRobot.cpp file from line 69)

[tinkersprojects]

Thank you, can you send the calculation and your code through please and I will review them. Regards, William Bailes https://tinkersprojects.com

…

On Fri, Oct 28, 2022, 12:39 PM Fran Tinelli ***@***.***> wrote: Hi! I'm using this library for a project involving a delta robot and computer vision. I'm trying to move the end effector on a horizontal plane by maintaining a constant z and giving different xy commands. However, it descends the more it gets farther from the center. At first, I thought it was something about my code or the servos calibrations but I did tons of tests and the behavior is the same. So I did a bit more research on the web and I found a repo of @isaac879 <https://github.com/isaac879> with inverse kinematics functions for a delta robot. I tried those out and it works like a charm! Now my robot moves perfectly well on a horizontal plane. So maybe there's something wrong with the code or the mathematical approach. I wanted to notice this as this is a well-known library and it would be great to have it fixed. — Reply to this email directly, view it on GitHub <#2>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEBFHMOEO3QDR3Q2S6OTNLTWFMVERANCNFSM6AAAAAARQTOPLY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[WideAwakeTN]

In case all angles are 0 degress the z coordinate is given by the following simple formula according to the Pythagoras theorem:
z = -sqrt[ (RodLength)² - (PlatformTri + ArmLength - BassTri)² ]
But the code of this repo returns a different value for the FK. Since the IK is consistent with the FK the IK is (equally) wrong too. Seems there is an implementation error hiding somewhere.
The values of x and y should be zero in this example.

[nicolasdetracy]

Hi,

I'm encountering the same problem as explained in this issue.
How can I fix this issue?

Kind Regards,
Nicolas de Tracy

[tinkersprojects]

can you send your code through to me soo i can review it. there is many different ways of calulating it and it may come down to an edge case that the function does not work.
i need to see your numbers and how you got there soo i can fix it

[nicolasdetracy]

This is the inverse kinematics part of my code:

`
#include "DeltaKinematics.h"

DeltaKinematics DK(150, 350, 55, 115);

void setup() {
// move to point 1 (x = -120, y = 0, z = -380)
DK.inverse(-120, 0, -380);
moveMotors(DK.a, DK.b, DK.c);
delay(1000);

// move to point 2 (x = 0, y = 0, z = -380)
DK.inverse(0, 0, -380);
moveMotors(DK.a, DK.b, DK.c);
delay(1000);

// move to point 3 (x = 120, y = 0, z = -380)
DK.inverse(120, 0, -380);
moveMotors(DK.a, DK.b, DK.c);
delay(1000);
}

void loop() {}
`

These 3 moves should move the end effector to 3 different points on the x-axis with the same z-axis value.
But the end effector actually descends the more it gets further from the center in real-life.

Thanks for wanting to take a look at this problem.
Kind regards,
Nicolas de Tracy

[tinkersprojects]

thank you, i will look at this, may take a while, need to dig up my old notes

also please note, this will only calculate the final position. you may get some movement in Z as it is moving

[nicolasdetracy]

Thank you for your willingness to help!
Yes, I understand that this calculates the final position and may get some movement in Z as it's moving.

Nicolas de Tracy

[nicolasdetracy]

Hi, have you had any luck finding the problem to our issue?

Nicolas de Tracy

[nicolasdetracy]

fixed the problems by creating my own library, which works very differently that all the other ones. The end effector moves around with an amazing accuracy and no changes in height at all.

Link to repo: https://github.com/nicolasdetracy/deltaRobot_Inverse_Kinematics

[mohamedwidar]

the repo is private or removed mr nicolas