Falcon 500 + NEO 550 Swerve Drive

I feel like this could be the ultimate swerve drive. Falcon 500 for the drive, and the new Neo 550 as a steering motor. I definitely don’t have the expertise to design it but I’ll be excited when somebody does.

2 Likes

Real question to those who have designed/built one.

How critical is the steering motor? By that I mean:
Is there a lot of issue with not having enough turning force?
Is there a lot of issues with not having enough turning speed?

I don’t see too many window or bag motors out there and last year teams used Neos which seemed like overkill.

When we did swerve in 2015, we used Banebots 550 motors for the steering and they cooked! With the OP’s proposal, I don’t see the benefit as opposed to just going for another Falcon, especially considering the extra gearing and separate controller needed for the Neo 550.

1 Like

Throttle motor swerve

6 Likes

We are planning to evaluate using the NEO 550 for our steering motor.

The main enhancement that we are excited about over our current system is that the built-in encoder on the NEO motor can serve as a redundant steering encoder to our absolute rotation encoder. That way, if we lose the signal from the primary encoder, we can have a back-up source to be able to keep driving. Losing an encoder is probably our biggest remaining reliability issue with our design and we have several ideas as to how to fix it. This one is one of the least invasive in terms of design changes we would need to make to implement it.

We don’t have the specs on the NEO 550 yet, so we can’t calculate the gear reduction that would be needed for the steering system. We currently run an RS550 with a 40:1 gearbox with a 100:30 final reduction into the module. This gear ratio was mainly selected for PID stability reasons. There is plenty of steering torque to react the torque imparted by the drive motor that tries to spin the module. With the NEO 550, we are wondering whether we can actually use the closed loop control at the motor controller to get a faster response of the system (rather than running through a control loop in the RIO) and maybe be able to reduce the gear ratio to a 10:1 gearbox followed by the 100:30 final ratio. This would allow us to reduce one stage of the VP gearbox which will reduce a little bit of weight and lower the height of the module. Once we have the motor specs, we will run the math and see how it comes out. This lower ratio coupled with what we expect to be higher low-end torque of the NEO 550 should result in quicker turning speed of the module, but that has not really been a limiting item for us. The response of our current implementation is fast enough to any real world driving.

Given how much we have invested in NEOs, we are currently planning to keep those for the drive motor rather than investing in the Falcon 500. But, the Falcon 500 is a drop in replacement so if we ever wanted to switch, it would be very easy to just add that to our module.

1 Like

Falcon all the things.

This is analagous to 2910 running NEOs for steering this year. More power, barely heavier, and less stages of reduction needed.

15 Likes

Fair point. Maybe then weight of the extra gearboxes would offset the weight reduction from using a smaller motor.

I haven’t done the math but I wouldn’t be surprised if the steering NEOs on 2910s modules made it lighter than with a 775/550/BAG with the additional reduction.

2 Likes

We did the math:

The NEO with a single stage VP lite gearbox and the CIM adaptor is 1.408 lbs

An RS550 with a 2 stage VP lite gearbox is 1.061 lbs

So, the NEO steering motor config is about 1/3 lb heavier per module. The RS550 is less than half a pound wheras the NEO is just under a pound. So, the NEO starts off with a 0.5 lb disadavantage. The difference between a single stage and a 2 stage VP light is not 0.5 lbs.

I also recognize that the VP lite is not recommended for the NEO. So, you might need to use a regular VP gearbox with the NEO. The weight a single stage VP plus a CIM adapter (0.61 lb) is heavier than a 2 stage VP lite (0.58 lb). So, in this case, the NEO steering config would be over 0.5 lb per module heavier.

Note: I have not taken into account the weight differences in the motor controllers or the wiring (the RS550 weight does not include wire leads). So the actual differences will be slight different. These comparisons are based on the weights listed on the product websites.

1 Like

I realize that I did not directly answer your questions.

Both the turning torque and the speed need to be considered in picking a motor, gearbox combination. Through trial and error, we have established a steering speed target of 150 RPM. This steering speed is based on experiments we have done where we have been able to produce a stable steering response through PID tuning (control loop in the RoboRIO).

For each candidate steering motor, we select a gear ratio that produces that speed. Then we look at what the stall torque for that motor gearbox combination (current or voltage limited as necessary). This gives us te steer torque available. The holding torque required to hold the module steady is calculated from the drive motor stall torque (current or voltage limited as applicable) times the portion of the gear ratio before the bevel gears. We then compare the torque that could be imparted by the drive motor to the ability of the steering motor to hold the module steady. We like to see a decent safety factor there (2x or more).

In the case of our current steering motor option of an RS550, the optimal gear ratio for controllability gives us more than enough torque to hold the module steady (about 6x more torque than the drive motor produces at the bevel gear). So that steering speed is our limiting factor for our current design (not holding torque).

Higher steering speed is generally considered to be better. But that higher speed is generally limited by your ability to control it, not the speed of your motor. The speed of the motor does factor into it. Our semi-empirical approach to determining our target steering speed does not really quantify how the motor speed factors in. Other teams have been able to achieve higher steering speed. I seem to recall some teams achieving 200 or 250 RPM steering speeds and being able to control them with stable control loops. We have some plans to do more testing to see if we can increase our steering speed (as mentioned above), but if we do, we still need to consider these other factors when selecting the gear ratio and motor for our steering.

While this is true, and I’m assuming you are using this as a clear comparison, I think most teams that are using a neo as steering will make their own reduction, rather than using a vp gearbox. you will likely need two stages, but depending on your ratio, the weight of it could vary.

overall, I think using a neo is better than using a neo 550. the extra added ratio needed to run the swerve drive is not only an issue of size, but also complexity. that extra stage is another stage you need to worry about, and it adds another failure point.

Sorry if I wasn’t clear, I was referring to the Mk2 module 2910 ran last season, which does not use a VP or any COTS gearbox when using a NEO for steering. A 775/550/BAG all require some additional gearbox to achieve the desired reduction. So the comparison is a RS550 with a VP lite to a NEO with no gearbox, and I’m pretty sure the NEO wins out there.

2 Likes

That’s fair. I can only comment on our design. If you use spur gears to get the reduction, one less stage could be substantial weight savings and more than offset the 0.5 weight difference between a RS550 motor and the NEO.

For our design, the weight is in the RS550’s favor compared to the NEO.

I am hoping that the NEO 550 may be even lighter and the torque, speed, etc along with a control loop running in the motor controller with it’s shorter loop time may also allow us to drop to a single stage VP lite which could save a little more weight. Overall, even if we don’t save weight, I am excited to try the NEO 550 for the potential improvements in reliability.

I know why it changed but, it feels weird seeing your name with the poofs logo instead of 973’s.

13 Likes

Adam is right here. The extra stage of gearing makes up whatever weight reduction you would get using a NEO550 or a BAG, etc. The extra stage of gearing also complicated packaging and adds more failure points.

2 Likes

I don’t know who I am anymore.

But I do know I’ll try to only use Falcons next year.

15 Likes

I’m about to use only baby Neos next year

2 Likes

theres going to be some wasted weight tho: we made an offseason 8 NEO defense bot and it can’t go at full speed because it pulls too much current. Falcons are slightly more efficient, but they also go slightly faster so i imagine it’ll get a similar problem, which means ur going to need a permanent power reduction which kinda defeats the point of having such a powerful motor for steering. NEO 550 will probably suffice lol.

1 Like

yeah, but four of those falcons are just for turning the modules, which will draw significantly less current than if it were driving the robot.

that extra power kinda becomes dead weight tho, unless as said before its lighter than the neo 550 + gearbox (which is very possible)