I’m doing some work on flywheel dynamics and I’m hoping I can get some help from crowdsourcing. I need to test my math against real world flywheel shooters, but unfortunately I don’t have any on hand right now. What I need is a graph of speed vs time as the flywheel is starting up. Then a short description of the setup including:
# and type of motors
gear ratio
flywheel moment of inertia (or mass and radius if you don’t know the exact MoI)
I’m hoping some teams may have a record of this from tuning control loops, or be able to easily generate it. And if you have from more than one shooter from various years, that’s great too. Thanks in advance!
I assume you are not looking for flywheel stability or gyroscopic effects and are really just looking for the effective inertia of a geared flywheel system?
After marveling at the wind-down of 118s flywheel at the end of their 2017 reveal video, our team added a geared flywheel to our bot. The effective rotational inertia imputed to the shooter shaft is pretty simple math. I will see if we have any details from the 2017 design.
We used a 775 pro for the flywheel shooter. I will try to find the gear ratio from the motor to the shooter shaft and then from the shooter shaft to the flywheel as well as the flywheel mass, and inertia.
I’m not sure if we have any saved data of speed versus time. But I believe the robot is still functional so we might be able to collect some data.
We did find that we needed to carefully ramp up the motor speed. If we just applied full power to the motor, we would pop the breaker… I think the spool up time was about 3 seconds. It was a very scary sounding 3 seconds and most people would take a step back from the robot as it reached full speed. The final tuning settings should still be in the code. I will ask our current programmers if they can find it (the 2017 programmers have graduated).
Does your math also include the inertia (speed) loss due to shooting the ball? If we can get our 2017 shooter working again, we could shoot a couple of fuel balls (we have plenty laying around) and try to collect some data of the system response. But you will probably need to understand our shooter geometry in order to model it.
Right now I’m trying to model the spin-up time for the flywheel. I have the math done for the time it should take to accelerate between two velocities. But I want to check that the answer I’m getting approximately matches some real world examples before I’m sure that I actually did all the theoretical stuff right.
If you can send data when shooting projectiles, that would be great too. Then I can check the other half of the calculation that deals with the inertia loss from shooting projectiles. And it would also help with measuring recovery time when the flywheel doesn’t start from standstill.
Edit: I didn’t think about having to ramp up motor speed to avoid popping breakers. I might have to add that to the calculation.