FRC-Characterization Inversion Issues

Technical Programming
#1

My team is using the FRC-Characterization Suite on NEO mode. (I assume that we should be using NEO mode, but can anybody clarify the difference between NEO and Spark MAX mode?) The issue is that when we run the forward test, it instead turns. The inversion options seem to be odd with this mode, as it says that the inversion is for the follower motors with respect to the master. Our motors are not inverted on the same side with respect to the master, but we need to invert one entire side. We tried to set both the Left side to [True, True], and then the Right side to [True, True] separately, but when we did this, the side that we changed to True simply did not move at all. How do I simply invert one entire side of the drivetrain in the NEO mode?

Thanks!

#2

The DifferentialDrive class used internally automatically inverts one side. You should set all motors to the same inversion. Try all False, and if that goes backwards, set all to True.

#3

It’s not going backwards though, it’s turning. Also, we just tried setting everything to True, and our motors just make a really weird noise. I’m not sure exactly what it is, possibly they are fighting each other or something, but the robot just doesn’t move. It doesn’t sound like fighting, and you can still move the wheels manually. All I know is that I don’t think both should be set to True on the same side, because that doesn’t seem to work. When we run with all false, the robot turns with each side turning a different direction.

#4

What kind of gearbox do you have, do you your motors on the same side of the drivetrain need to rotate in opposite directions?

We struggled with the the tool a little to get the bot moving an rotating in the right direction too, but I think our struggles were different. We ended up just modifying the generated project to invert the value coming from the network table. double autospeed = -autoSpeedEntry.getDouble(0);

#5

We’re using the gearboxes from the KOP. One side of the drivetrain needs to be rotated in the opposite direction to go forward.
Whenever we try to set a side to inverted: [True, True], that side doesn’t move. Only the sides that are not inverted end up moving, so the robot turns. If we set both sides to [True, True], nothing happens, but the SparkMax motor controllers still show that they are moving (green/red blink lights).
When we set both sides to [False, False], then both sides of the robot just end up moving in opposite directions

#6

You’re using the NEO project. As mentioned in the comments in the config file, slave inversion for the NEO project is opposite master. So, [True, True] means the master is inverted, and the slave is also inverted with respect to the master. In a KOP gearbox, that’ll result in the behavior you’ve seen (and is not healthy for the motors/gearbox, to boot).

What you want is for one side to be [True, False].

1 Like
#7

Thank you, we will try this tomorrow! One thing I’m noticing is that if both motor controllers are [False, False], then in the actual code, both the left and right motor controllers will be set to motor.setInverted(false). But, if I set only the left side to [True, False] and leave the right side as [False, False], in the code, it changes both the left and right motor controllers to motor.setInverted(true).
Is this supposed to happen?

#8

This was a bug previously, but has been fixed in recent versions. Be sure to update your version with pip install --upgrade.

#9

We are upgraded to frc-characterization-2020.243 which appears to be the latest, but we are still experiencing this bug. Is there a later version?

#10

Ah, apparently we had only fixed it in the brushed spark max project. It’s fixed now.

2 Likes
#12

Just to confirm, what are your polarities set to when it’s turning? Having no movement, as y’all recognized, most likely means that they are trying to fight each other.

That said, if you can also spin the motors freely while characterization is running, it means that brake mode isn’t being activated—meaning that either your characterization routine is erroring somewhere, or that you aren’t addressing the correct motors.

Could you post the generated project, perhaps? If you are sure the motors aren’t fighting each other, and that the ports for the motors you’re addressing our correct, there may be some more niche error that’s popping up.

#13

Maybe I missed the answer in this thread, but what is the difference between the two options? Brushed/brushless?

#14

Correct.

#15

So just to be sure, “Neo” means brushed, and “SparkMax” means Brushless?

#16

No, that’s backwards, but I assume you meant it the other way around, as well :wink:

#17

Sorry, I’m a bit confused :sweat_smile:. You mean I wrote it the other way around? So “Neo” means brushed, and “SparkMax” means brushless?

#18

“Neo” means brushless, because Neo motors are brushless. You wrote “‘Neo’ means brushed” in your original post, but I assume that was a typo.

1 Like
#19

OK, now I understand (that was not a typo :sweat_smile:). That actually makes a lot more sense.
Sorry for the hassle and thanks!

#20

If we wanted to use the characterization suite with a SparkMax and NEO using an external encoder, do you recommend we simply just edit the code directly to use getAlternateEncoder() instead of getEncoder()? Or is there something else we will need to change/a different process we should go through

#21

Not worth it; just do the characterization with the internal encoder and use the external encoder in your actual code. The empirical constants don’t depend on the encoder.

If you really want to do characterization with an external encoder on a NEO, modifying the generated project is your best bet.

2 Likes