Change default PID values using blocks coding

[luanveras3]

Currently, I'm using a solution, but I'm not sure if it's the most efficient one. My workflow is to create the code using blocks, then copy and paste it into a Python file where I have to manually change the default PID values. I would prefer to input these values directly within the block code interface.

Is it possible?

[laurensvalk]

Yes! You can make it work with one Python function that you never have to change.

You can just enter your values in the blocks. Here is an example for a single motor:

The meaning of the values is given here. In this example we only set the first 3 values. The rest remains unchanged.

You can see the original values and the new ones at the bottom.

For a drivebase you can change the controls for the distance and heading:

To make these examples work, create an additional Python file called pid_python with the following contents:

from pybricks.tools import run_task

async def awaitable():
    return None

def set_pid(control, *values):
    print("Original", control.pid())
    control.pid(*values)
    print("New     ", control.pid())

    if run_task():
        return awaitable()

def motor_pid(motor, *values):
    return set_pid(motor.control, *values)

def distance_pid(drivebase, *values):
    return set_pid(drivebase.distance_control, *values)

def heading_pid(drivebase, *values):
    return set_pid(drivebase.heading_control, *values)

So it will look like this:

[luanveras3]

it works perfectly, i really appreciate!

[safarnss]

This is great. But how do I find the range of values for PID parameters. Normally Kp, Ki & Kd is scaled relative to 1, in most PID tuning. But by looking at the default values printed here, see the scale is different. What are the parameter value ranges ? For example, what is the value equivalent to unity gain for Kp ?

[laurensvalk]

Normally Kp, Ki & Kd is scaled relative to 1, in most PID tuning

kp, ki, kd have units.

For example, kp is _the amount of torque (in uNm) to provide for every degree of error (deg)`, so uNm/deg.

But how do I find the range of values for PID parameters.

So you can start from the original values which are printed. Then you could do * 1.5 for 50% more feedback, or * 0.90 for a small reduction for example. This is in essence using relative values as you wanted 😄

[safarnss]

Thank You !