Interface for the rotation direction

[husch]

Hello,

It is useful to configure zero positions, joint limits, PID gains etc using Scope, since the users do not have the direct access to microcontrolers.

Is it possible to add an interface to configure rotation direction on Scope as well?
It makes sense to have both position offset and sign (rotation direction) configured on the microcontroller side.

Thank you very much. Look forward to hearing your thoughts on that.

[florian_enner]

Can you elaborate on why you would want to change the direction?

The direction is logically consistent with the labels on the actuator, the hrdf model, and other sensors like the IMU. My first thought is that allowing inconsistent negative signs may end up confusing users and may result in unexpected bugs that are difficult to debug. I'm open to adding it if there is a good enough reason for it though.

[husch]

Thanks for the quick reply.

True, if we change the rotation direction, we will need to update joint and motor position and velocity, deflection and deflection velocity, IMU output, etc. accordingly on the microcontroller side.

The main reason to change the direction is to align the actuator rotation direction with the our robot coordinate system.
One example is the left and right shoulder joints of Igor. Using our robot coordinate convention, when Igor lifts both his arms to the same angle, the joint positions of left and right shoulder are the same, instead of having different sign.

It is an option to do the conversion at the higher software layer instead of on the microcontrollers. But it is quite confusing to use Scope to configure and tune the robot, the users have to do this conversion as well.
Look forward to hearing the argument from the other side

Thank you very much.

[dave_rollinson]

One interim option might be to setup some custom plots in Scope that invert feedback.
https://docs.hebi.us/tools.html#custom-charts

-Dave

[husch]

Thanks Dave, that is a good interim option
So what are your thoughts on adding rotation direction in the Scope's 'Danger Zone'?

[florian_enner]

Sorry. This is a reasonable request, but after internal discussion we do not think that it would be a good idea to break assumptions at that level. It offers too much potential for confusion for a limited use case.

[husch]

Thank you Florian for discussing it internally.