We have been running our robot pretty hard the past 4 days at our county fair. We have seen after an aggressive run the motors trip the thermals. Seems reasonable. I have been watching the master vs Slave motors in the Phoenix Tuner Plot. The master spikes at 120amps but the slaves all spike at 30ish. Is this expected behavior? Is there a way to better load Balance?
Just to make sure, are the motors all geared/tied together?
Is the master the one that trips?
They are tied together into a 3Cim ball Sifter. I suspect with the Splined falcons they are still mechanically engaged.
I only just stared to monitor and have not captured the self-test when the event happens. Waiting for batteries currently.
We saw issues like this in 2019. I would not think this is expected behavior – if the motors are mechanically tied together I would expect to see only minor variations in current draw between them (within measurement error probably). Steps to debug:
-
Do you see identical behavior on both sides, or is it only on one side? If it’s only one side it may have something to do with wiring (bad connection on the one that is pulling more current).
-
It’s possible your motors are fighting each other rather than running in the same direction.
a. I would change the code to get rid of the follower stuff and call set for each motor individually
b. Unplug two of the motors for each gearbox and see how it drives (not aggressively, just to see which direction it goes)
c. Repeat for all the motors so you know all of them are trying to spin in the same direction. -
I’m not sure about this, but I think current limiting and some other features may not work on the followers if you only set them on the leader. I know that we tend to set the config stuff for each motor even if there’s a follower. Also, generally, make sure you’re resetting each motor controller to factory settings so that you don’t have some motors retaining old current limits or anything like that.
1 Like
First, be sure all motors are driving in the same direction – you can do this by commanding one at a time, without leader/follower-ing them. Should be able to do via tuner, but this is the kind of thing where some good test mode code is really helpful. This is also a good way to check for mechanical issues and see the current draw when each motor is acting alone.
Try to use the terms “leader” and “follower” instead of “master” and “slave”. Please.
21 Likes
Once you feel good that your measurements, and you did all of the above, you can try changing who’s master and who’s slave ans see if it tracks with who’s who
1 Like
Did run a few test before closing the booth down.
We put the bot on stilts. With out removing the bot code then using Phoenix Tuner
commamded motor 1 (leader) all motors flash green. Wheel spin forward ,
Commanded motor 2 flash green spins forward SLOW. I suspect it’s is because the robot code is also trying to command a zero speed *motor 1,
Commanded motor 3 same as 2
Other side is the same. So phase seems inline and mechanicaly coupled.
Fur the question if both sides appear to have similar results. YES
On stilts M1 peeks 90amp, M2 peeks 20, M3 peeks 30.
I know there is much to try yet. Probably first going to try amd remove all follower code and command each individually with the same Controlmode.Velocity, setpoint.
So there is a question what do the followers follow? The master is running in velocity mode with just a feed forward configured. The followers do not have any pidf. I was to believe the followers follow using percent output of the leader.
@marshall master/slave are the terms I commonly see with my Rockwell Automation Motion Control Systems.
2 Likes
It is in many motion control systems. That’s why it’s even more important to try to not use the terminology. We are attempting to stamp out nomenclature that is based on the enslavement of Africans and African Americans.
8 Likes
Other poster already said it. These terms are outdated and actively harmful to trying to teach concepts to students and create a more diverse and inclusive culture within STEM ecosystems. There are many businesses that have signed on to inclusive language practices, including Rockwell - who is a sponsor for FIRST:
While it might be frustrating to change and update references that were used for decades, it’s a reminder for us all to be open to evolving our language to reflect the changes in our society and culture.
Think of the two terms, “master” and “slave.” These terms permeate the engineering domain, but only recently came into focus across many tech companies because of their association to slavery. Other terms, such as “segregate” and “man hours,” have similar challenges when it comes to inclusive communication.
14 Likes
… and you’re sure none of them have current limits configured? If your code does not set it or turn it off, the falcon will remember it from the last time it got set.
Last night after I got home I did find a find current limit configuration set. The master has just the supply configured but the slaves have both the supply and stator.
I thought these had all been removed or adjusted a few weeks ago when I was here trying to identify the apparent accel/decell ramping.
I’ll report the results later this afternoon when the fair reopen today.
2 Likes
I was not aware of this movement but a quick read of the Rockwell document linked above suppized me with MMI to HMI. I use HMI every day. I started my career in the late 90s. I suspect even this old dog will adopt the new instruction set terms over time.
3 Likes
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.