robot joint offsets required to overcome internal absolute encoder

More
16 Aug 2022 07:43 - 16 Aug 2022 07:55 #249861 by Aciera

We need a way to add back into the system the "theta" values for following so the robot poses correctly in sim and kinematics work correctly.

Maybe try the "HOME" or "HOME_OFFSET" values in the [JOINT_n] sections or the "HOME" values in the [TRAJ] of the ini file to compensate for those theta rotations.
The simulation model needs to reflect any joint rotations that were made to accommodate the lack of theta values in genserkins dh-parameters.
 
Last edit: 16 Aug 2022 07:55 by Aciera.

Please Log in or Create an account to join the conversation.

More
16 Aug 2022 09:10 #249869 by andypugh
You might need to rotate the parts before linking them to the HAL pins in the Vismach model.

Please Log in or Create an account to join the conversation.

More
16 Aug 2022 09:44 #249870 by ftree
Not sure HOME_OFFSET works with absolute encoders.

Please Log in or Create an account to join the conversation.

More
16 Aug 2022 10:53 #249878 by andypugh

Not sure HOME_OFFSET works with absolute encoders.

It is certainly part of the calculation, though I would have to experiment to see what the result is as the description in the docs is a little unclear:
linuxcnc.org/docs/stable/html/config/ini...t__lt_num_gt_section
 

Please Log in or Create an account to join the conversation.

More
16 Aug 2022 11:23 #249882 by Aciera
The simulation configuration at configs/sim/axis/vismach/melfa-sim uses :
[JOINT_1] ... HOME = -90.001 ...
and
[JOINT_4] ... HOME = 90.001 ...

but I just checked the [TRAJ] HOME setting and using this might actually be preferable as that sets the zero pose (all joints at zero position) to be the same as is usually used in serial manipulator documentation:
[TRAJ]
...
HOME 0 -90 0 0 90 0
...

And again your vismach model's inital joint (zero) positions need to be the same as what you had to rotate the joints to in order to derive the dh-parameters. ie if you had to rotate joint1 by 90° as in the dh-parameter documentation (due to the lack of theta parameter in genserkins) then your model's inital position needs to have its arm pointing to the floor and not straight out.

Please Log in or Create an account to join the conversation.

More
More
16 Aug 2022 13:47 #249888 by ftree

And again your vismach model's inital joint (zero) positions need to be the same as what you had to rotate the joints to in order to derive the dh-parameters. ie if you had to rotate joint1 by 90° as in the dh-parameter documentation (due to the lack of theta
parameter in genserkins) then your model's inital position needs to have its arm pointing to the floor and not straight out.


In hindsight should have used the melfa-sim/melfa.ini from the start.  Would have saved me a lot of time.

Making some progress now.

I could really do with someone that knows exactly what they are doing.  I'm happy to pay for a review and code changes to get me over the line.  I'm struggling with the last 1% of effort that is really hurting my brain and general wellbeing.

Please Log in or Create an account to join the conversation.

More
16 Aug 2022 15:25 #249893 by ftree
I'm trying all kinds of variations of the above advice. My latest is very nearly there. Please review if you have 15 minutes.

github.com/djsftree/linuxcnc_meca500

Please Log in or Create an account to join the conversation.

More
16 Aug 2022 16:02 #249895 by Aciera
Installing your config is not as simple as my machine runs a rip installation so all my paths are different to begin with. Maybe try to describe what you expect and is not working for you.

Please Log in or Create an account to join the conversation.

Time to create page: 0.139 seconds
Powered by Kunena Forum