- Posts: 87
- Thank you received: 0
TCP 5-axis kinematics
- plopes9000
- Offline
- Premium Member
-
Less
More
6 years 10 months ago #151703
by plopes9000
Replied by plopes9000 on topic TCP 5-axis kinematics
I‘ll see if I can manage to post the files tomorrow.
Please Log in to join the conversation.
- plopes9000
- Offline
- Premium Member
-
Less
More
- Posts: 87
- Thank you received: 0
6 years 10 months ago #151771
by plopes9000
Replied by plopes9000 on topic TCP 5-axis kinematics
I've added a file containing the complete changed files, pl-switchkins.tar, to the original post with the patch file on
page 18
.
Please Log in to join the conversation.
6 years 10 months ago #151792
by 3D-Master
Replied by 3D-Master on topic TCP 5-axis kinematics
Thank you,
so i copied and overwrote all the files in the master-branch folder (made a copy before) but linuxcnc said unknown gcode so i fumbled a bit around and then i tried the sim config. The sim config starts correct, knows the G12.1 command and doesnt give me the error i get with my config. that must mean my config is "broken". afaik in Linuxcnc 2.8 and above joints are no longer hardcoded to the axis which means i can for example have axis a using joint 5 right? atm my joint axis is like this:
ini:
x = j0
y = j1
z = j2
a = j3
c = j4
b = j5
hal:
x = j0
y = j1
z = j2
a = j3
c = j4
is this correct?
i noticed that when i jog my c axis in trivkins there is no position change but the velocity shows something. only when i activate tcp i see a value in the c axis
so i copied and overwrote all the files in the master-branch folder (made a copy before) but linuxcnc said unknown gcode so i fumbled a bit around and then i tried the sim config. The sim config starts correct, knows the G12.1 command and doesnt give me the error i get with my config. that must mean my config is "broken". afaik in Linuxcnc 2.8 and above joints are no longer hardcoded to the axis which means i can for example have axis a using joint 5 right? atm my joint axis is like this:
ini:
x = j0
y = j1
z = j2
a = j3
c = j4
b = j5
hal:
x = j0
y = j1
z = j2
a = j3
c = j4
is this correct?
i noticed that when i jog my c axis in trivkins there is no position change but the velocity shows something. only when i activate tcp i see a value in the c axis
Please Log in to join the conversation.
6 years 10 months ago #151853
by dgarrett
In case you missed it, the XYZAC type switchkins
kinematics module (as updated by plopes9000) has been
incorporated into the experimental branch dgarr/s_kins.
The branch uses remapped mcodes (M428,M429) for
switching kinematics instead of gcodes (G12.*)
The branch can be built as a rip from source with
no additional patching or installed as a deb
available from the buildbot.
The branch awaits 1) an offered patch for updating
XYZBC for defining its rotation point and 2) user
testing.
================================================
related:
branch yesterday:
github.com/LinuxCNC/linuxcnc/commit/5719...7dc32838672ea8c473e4
================================================
branch announce:
forum.linuxcnc.org/10-advanced-configura...ics?start=220#150781
XYZBC rotation point:
forum.linuxcnc.org/10-advanced-configura...ics?start=220#150955
forum.linuxcnc.org/10-advanced-configura...ics?start=220#150961
guide for dgarr/s_kins usage:
www.panix.com/~dgarrett/stuff/s_kins_guide.txt
Replied by dgarrett on topic TCP 5-axis kinematics
so i copied and overwrote all the files in the
master-branch folder
In case you missed it, the XYZAC type switchkins
kinematics module (as updated by plopes9000) has been
incorporated into the experimental branch dgarr/s_kins.
The branch uses remapped mcodes (M428,M429) for
switching kinematics instead of gcodes (G12.*)
The branch can be built as a rip from source with
no additional patching or installed as a deb
available from the buildbot.
The branch awaits 1) an offered patch for updating
XYZBC for defining its rotation point and 2) user
testing.
================================================
related:
A bug fix for this error was committed to the masterlinear move on line 0 would exceed B´s negative
limit invalid params in linear command
branch yesterday:
github.com/LinuxCNC/linuxcnc/commit/5719...7dc32838672ea8c473e4
================================================
branch announce:
forum.linuxcnc.org/10-advanced-configura...ics?start=220#150781
XYZBC rotation point:
forum.linuxcnc.org/10-advanced-configura...ics?start=220#150955
forum.linuxcnc.org/10-advanced-configura...ics?start=220#150961
guide for dgarr/s_kins usage:
www.panix.com/~dgarrett/stuff/s_kins_guide.txt
Please Log in to join the conversation.
6 years 10 months ago #151918
by 3D-Master
Replied by 3D-Master on topic TCP 5-axis kinematics
Thank you,
you wont believe what solved (at least atm) my problem with G12.1 TCP.
In my ini file there was "KINEMATICS = switchKins coordinates=XYZAC"
i deleted "coordinates=XYZAC" and now i dont get the error anymore. Also i can see values changing when i jog the c axis in trivkins so i guess everything is working like it should.
Thank you all so much for helping!
This opens up so much more possibilities.
you wont believe what solved (at least atm) my problem with G12.1 TCP.
In my ini file there was "KINEMATICS = switchKins coordinates=XYZAC"
i deleted "coordinates=XYZAC" and now i dont get the error anymore. Also i can see values changing when i jog the c axis in trivkins so i guess everything is working like it should.
Thank you all so much for helping!
This opens up so much more possibilities.
Please Log in to join the conversation.
- plopes9000
- Offline
- Premium Member
-
Less
More
- Posts: 87
- Thank you received: 0
6 years 10 months ago #151963
by plopes9000
Replied by plopes9000 on topic TCP 5-axis kinematics
Glad to hear.
Did want to point out that the way forward to get TCP included in the master branch is to support dgarrett's new branch dgarr/s_kins.
guide for dgarr/s_kins usage:
www.panix.com/~dgarrett/stuff/s_kins_guide.txt
Did want to point out that the way forward to get TCP included in the master branch is to support dgarrett's new branch dgarr/s_kins.
guide for dgarr/s_kins usage:
www.panix.com/~dgarrett/stuff/s_kins_guide.txt
Please Log in to join the conversation.
6 years 10 months ago #152200
by 3D-Master
Replied by 3D-Master on topic TCP 5-axis kinematics
thanks for the hint, i will try dgarr/s_kins someday in the near future.
Please Log in to join the conversation.
6 years 10 months ago #152425
by thang
Replied by thang on topic TCP 5-axis kinematics
To speed up robot, also avoid singularities, change robot config ... i want to control robot in joint mode (this mode really necessary).
I tried branch github.com/rushabhGH/linuxcnc/tree/switchKins for robot to switch kinematic recently , it's worked but there are some problems:
+ When we change kinematic, the trajectory planner will keep using axis limits, max velocity, max acceleration parameters for planning, if these numbers higher than correcsponding joint parameters it can caused "joint following error".
+ Because we just change kinematic, TP still uses old offsets ( G54, G90,G91,...).
I think dgarr/s_kins branch also has this problem so any idea to solve this problem?
Edit: Figured out to change max velocity and acceleration by edit source but didnt find out where and how axis limits are used in auto mode.
I tried branch github.com/rushabhGH/linuxcnc/tree/switchKins for robot to switch kinematic recently , it's worked but there are some problems:
+ When we change kinematic, the trajectory planner will keep using axis limits, max velocity, max acceleration parameters for planning, if these numbers higher than correcsponding joint parameters it can caused "joint following error".
+ Because we just change kinematic, TP still uses old offsets ( G54, G90,G91,...).
I think dgarr/s_kins branch also has this problem so any idea to solve this problem?
Edit: Figured out to change max velocity and acceleration by edit source but didnt find out where and how axis limits are used in auto mode.
Please Log in to join the conversation.
- papagno-source
- Offline
- Elite Member
-
Less
More
- Posts: 170
- Thank you received: 0
6 years 10 months ago #153189
by papagno-source
Replied by papagno-source on topic TCP 5-axis kinematics
Hi.
I'm starting the tests of the Dgarret-skin branch, on a real 5-axis machine.
The measuring systems are incremental and the axes must perform the reference point.
After the reference ok, executed for the XYZBC axes, the machine position for all the axes is 0.
what values should i set for x-offset, z-offset, x-rot-point-y-rot-point-z-rot-point?
I'm starting the tests of the Dgarret-skin branch, on a real 5-axis machine.
The measuring systems are incremental and the axes must perform the reference point.
After the reference ok, executed for the XYZBC axes, the machine position for all the axes is 0.
what values should i set for x-offset, z-offset, x-rot-point-y-rot-point-z-rot-point?
Please Log in to join the conversation.
- plopes9000
- Offline
- Premium Member
-
Less
More
- Posts: 87
- Thank you received: 0
6 years 10 months ago #153224
by plopes9000
Replied by plopes9000 on topic TCP 5-axis kinematics
X and Z offsets are the offsets from the center of rotation of the C axis relative to the center of rotation of the A axis. X, Y and Z rotation offsets are the offsets of the center of rotation of the A axis relative to the machine absolute zero coordinates.
Please Log in to join the conversation.
Time to create page: 0.397 seconds
