Pick a new brushless motor for FRC

Technical Motors
1

We’re all aware of the issues with current brushless motors. There are other threads that have thoroughly discussed that. This thread is to look at what other COTS brushless motors are on the market that might someday be suitable for FRC. The market is flooded with all kinds of brushless motors. From RC cars to quadcopters to skateboards and more, they’re everywhere. Forgetting for a moment that most of the ESCs don’t do CAN, what motor and controller combo do you think would be a great fit for FRC and why?

5 Likes
2

Although I have no personal experience using it, it would be nice to have Nidec back in the FRC COTS game. As mentioned here

3 Likes
3
4 Likes
4

Honestly I’m pretty happy with the NEO. Compared to the CIM motors we used to use, it’s smaller, lighter, more powerful and almost the same price.

That said, it would be interesting to have a low KV (< 1000 rpm at 12V), high torque, “pancake” style motor like this one: https://store.tmotor.com/goods-957-G100.html. The Low KV means you could get away with only one or zero gear reduction stage for a lot of applications. Plus the thin packaging and hollow bore could allow some really small mechanisms.

26 Likes
5

Tbh, I feel like it wouldn’t hurt frc to basically open up motors at this point. We’ve pretty much hit the limit of size to power with the existing motors. So allowing RC motors in general I don’t think would hurt frc at all.

The harder part is the controllers. People just love their can too much and wouldn’t stand for a pwm only controller.

5 Likes
6

I mostly agree except for this point: Trying to support a thousand motors with a thousand mounting standards will make it harder for vendors to make gearboxes, and impossible for teams to help each other by stocking spares.

8 Likes
7

Personally not too worried about it. I think our suppliers will pick out a couple of the best and most appropriate motors, resell them, and make accessories for them pretty quick.

7 Likes
8

I’ve heard this too. But I just helped a team this weekend who ran their Spark Max’s on PWM control, so maybe there is a market for a pwm-only motor controller? Can the percentage of PWM vs CAN controllers be calculated from FMS statistics?

3 Likes
9

As someone who’s looked into developing motors for FRC fairly frequently over the last 3 years:

REVs prices for sensored brushless motors are actually very reasonable. Frankly, I don’t think they make much at all on the NEO, and the margins are probably mostly in the Spark MAX. A sensored 3548 motor (similar to N550 size) is around $20 in quantities of 1k for any decent quality, though I did get quotes down to $16. This is before shipping and tariffs. The NEO could probably be had for $30 base price, but again you’re going to pay a lot in transport.

So if we can’t reduce the cost of the motor, what about the controller? As it stands, the closest COTS equivalent is the VESC or maybe an Odrive, both of which run about $90 per motor, and without FRC-specific firmware.

Custom controllers is where I think there’s a possibility for improvement. We could potentially have a $20-40 controller that’s just a simple CAN-based sensored motor controller. But the company that sells controllers probably can’t sell motors too and expect to turn a profit - we would need to open up the market for hobby motors, which often have unkeyed shafts, so there would need to be a new pinion gear ecosystem, probably based around set screws.

Ideally, I’d like to see the rules opened up to allow (775-sized) hobby motors with any COTS controller, as long as it’s on a 30A breaker. That would allow companies to more easily develop cheap controllers and allow teams to source motors from the many hobby RC vendors that sell them. The current motor controller guidelines generally require a large number of donations to be approved, but that can be bent in some situations.

10 Likes
10

There’s 2 hard parts with hobby grade controllers (not counting the pwm only part)

  1. Most have a bec, and push voltage down the middle pin. Teams would have to cut that or risk damage to their roborio.

  2. Most have an LVC due to being used with lipos. I know most 2s/3s capable escs can disable the lvc, however pumping an frc battery through one of those would overvolt the controller, since 3s is only 11.4 nominal. So to be safe, you’d need at minimum a 4s controller, which is harder to find one that can disable the lvc.

1 Like
11

We have to balance the needs/wants of a lot of stakeholders here: teams, robot inspectors, manufacturers, FRC suppliers. There could be concerns with, for instance, safety. With no standard, it’s likely that some off-brand will sell inexpensive motors that have not been safety checked, resulting in fires or damage to a team’s robot and control system; obviously not acceptable.

Honestly y’all, we have two BLDC motors that are readily available now (Neo and Neo550), and I’m dead sure that at least one competitor will be available this fall and ready to put onto robots for the 2024 game. We will be okay, no need to open it up to any and all BLDC motors.

12 Likes
12

There is a safety concern with this given basically all non-FRC controllers are PWM controlled–without testing it, we don’t know how a COTS controller will react to PWM control oddities or how long it will take to disable (or bursting to full speed due to seeing a glitchy pulse) after the PWM signal is stopped/cut (e.g. due to a Rio reboot or a signal wire cut). CAN’s not really any better unless there’s FRC-specific firmware on it.

4 Likes
13

I really want to use one of those 2-4s sensored controllers for RC cars, because they are less than $20 in most cases incl. shipping. But the ultra-nonlinear PWM response scares me away.

1 Like
14

Dark Matter.

20 Likes
15

I’ve never seen a controller that keeps going after you stop commanding PWM, but I guess there’s a first for everything. Making them on 30A breakers at least stops most teams from having them on their drivetrain, limiting damage.

I think the likelihood of this happening is really low, as it would probably be something a team discovers during the season, and far greater and more common hazards already exist. At most, I would accept a generic “approval process” where FIRST just checks for safety behavior and then hits the approve button.

2 Likes
16

I’m thinking of motors like those where you can use a generic Hall phase controller with them.

Even something like the ODrive motor controllers, if they could guarantee stock and a good set of tutorials for FRC level students and those would support custom encoders or other methods of control. The ODrive Pro supports can, PWM, serial for input control.

Really any BLDC with good hall feedback and a decent torque in range with the falcons or neos and a similar footprint would work. Anaheim was just the first one I could think of that I’ve used professionally. They weren’t super duper expensive but anytime you go from Educational to Industrial there’s a price jump.

2 Likes
17

I’ve definitely heard about cases happening where the controller sees a short pulse (e.g. the Rio powering off in the middle of a pulse) as a full-scale signal, and only after the “disable” time actually shutting off the motor (and the disable time is seldom specified).

I think that’s mostly what FIRST does to approve controllers already. Usually with some donation involved because there’s a cost to FIRST to do the testing. Note the controllers have to be inspectable as well–if there’s no production control, unlabeled devices, etc, it’s hard to ensure that it is what you think it is.

1 Like
18

Those motors are weaker than a Nidec… not much call for that in frc.

If we want cheaper motors, we should go back to using brushed in lower power applications.

2 Likes
19

I don’t have any specific suggestions, but I do think it would be nice to see a new brushless motor in the 775-size class that falls somewhere between the performance of a Neo550 and a Neo/Falcon.

Sometimes, you need just a bit more power (or stall tolerance) than a Neo550 can provide, but going to a CIM-size class motor would be overkill. Many COTS gearboxes are also already compatible with 775-size motors too, so it would be easy to add them to the ecosystem.

5 Likes
20

That specific one is yes, it was just close to the right footprint size and an example of the right generic 8 wire output. I agree it’s like 2% of the power of a Nidec. But that might be okay for a shooter flywheel motor where you don’t need a massive stall torque, just good control and high speed.

1 Like