20230510085234996.pdf (121.6 KB)
(not to scale)
My Team (7810 Eagle Robotics) Worked on a 12-motor chassis design recently, but it is in a purely theoretical phase right now.
we are looking at getting twelve REV NEO motors but are majorly concerned about power draw, with my basic calculation - running the maximum amperage of the motors would net us roughly 2 minutes of life, but we would cap the motors pretty well below that power consumption.
do you think this design would work?
Itâll probably work with some current limiting and youâll get the efficiency boost from using multiple motors which was heavily discussed in a few tri motor swerve threads, but you will need to play with current limiting when it comes to spikes browning out your system.
How Limited do you think we would have to bring the motors in order for this to be reasonable with the standard 12v 18AH batteries?
Whatâs the advantage of this over simply having two motors on each swerve module? The coefficient of friction of omni wheels is lower than that of treaded wheels even when moving parallel to the wheel. Any weight taken up by the omnis has its contribution to the robotâs pushing power halved, because two wheels are perpendicular to the direction of movement. Also, making sure all 8 wheels are roughly equally loaded seems rather difficult (youâll probably have to use pneumatic cylinders or springs).
I think this is very likely to be worse at accelerating than a normal swerve drive because you have two weight bearing wheels that arenât driven.
30 amps if you want to smack a limit on everything but Iâd recommend doing some more dynamic limiting to keep overall usage fine but still allowing single modules to output more power when needed assuming others donât need to move very fast at the time.
we are designing two motor swerve modules in-house at the moment - did you mean a three-motor module?
we believe there are two major contributions added via the omni wheels - the first one being acceleration ability, though this is not the primary reason behind this design.
the second, and arguably most important, is a defensive aspect robot. swerve alone is much more maneuverable than tank drive, but lacks the power to escape a push, as you have one drive motor and one rotational motor, where the rotational motor can be âout torquedâ and pushed aside by a well-made tank. these wheels allow for an extremely high torque to fight back or escape, and cannot be turned aside.
we have a design in the works that I personally have been tinkering with on pneumatic load drive boxes, possibly allowing to remove two âdead wheelsâ from the floor when not in use.
and the main reason we want to do this design is that it looks cool, FRC 7810 is a young team that needs experience with design and personal build robots
we have no delusions of grandeur, but we want to have something our own and surprise a few people along the way
plus wouldnât it be such a fun convo in the pits? id be geeking out over the abilities of such a bot, especially if it was doing well.
Thank you, Iâm picking up some electrical engineering and mechanical design books whenever I find the appropriate ones to send to my library, but am currently not knowledgeable in the subjects as much as i wish to be, my idea was just to test in house
From looking at this idea, it looks like it would be mechanically simpler to have for three motor swerve drives, while getting the same benefits.
It looks like it isnât too much work to modify sds mk4is to a 2 motor drive and 1 motor steer design, which both high tide and barker redbacks used. Iâm sure itâs possible to modify REV Maxswerve and WCP swerve to have two motors driving, too.
If youâre designing your own swerve in house you could also design it for two drive motors per module.
I donât think itâs worth it for all but the highest performing teams to build a two motor drive swerve module though, you can do many more improvements elsewhere which would make more of an impact.
Fwiw, I havenât seen well made swerve drives being pushed around by tank drives, mainly the opposite acctually.
Swerve doesnât usually get pushed by tank too much but when a tank drive is horizontal with high traction wheels even two swerve bots fail to push it and it makes defending centre field extremely easy.
Yeah, what I meant was two drive motors, or equivalently three total motors.
What Iâm trying to say is that the proposed design will be worse than an ordinary swerve (with 1 drive motor) in both these categories (not even considering a swerve with two drive motors).
Pushing power and acceleration are determined by the exact same thing: how much force your robot exerts onto the ground, and therefore (by Newtonâs third law) how much force the ground exerts on your robot. For each wheel, the torque generated by the wheelâs motor(s) pushes against the ground, and the frictional force between the ground and your wheel pushes back on your robot.
Now, the frictional force between the ground and your wheel is limited, its maximum is
where F_f is the frictional force, u_f is the coefficient of friction (determined by your wheel and the surface itâs rolling on), and F_N is the normal force, in this case the portion of the robotâs weight that is resting on that wheel.
If you care about acceleration and pushing power, you should make your drivetrain traction limited, which means that it is geared low enough that the motors are able to produce enough torque so that the frictional force is at its maximum. So, Iâll assume that for this analysis.
Now letâs calculate how much force (and therefore pushing power and acceleration each wheel generates). Letâs say your robot weights 120 lbs. In a normal swerve drive, each wheel carries 30 lbs. Iâd estimate treaded wheels have u_f = 1.3, so each wheel generates force equal to 30 * 1.3 = 39 lbs for a total of 156 lbs. In your omni swerve drive, each wheel carries 15 lbs. The four swerve wheels produce force equal to 15 * 1.3 = 19.5, and the force produced by the two omni wheels (Vex says u_f = 1.1, Iâm guessing this is high but weâll go with it) is 15 * 1.1 = 16.5 lbs so 111 lbs total. The normal swerve has 1.4 times more pushing power.
Of course, you have the advantage that you have two extra motors helping you generate torque, so you can gear higher and still be traction limited (or remain traction limited at higher velocities). This decreases your sprint time. However, the benefit is limited because you are limited more by the battery than the motors. So, the main thing you gain is increased efficiency by running each motor at lower current for a given total torque.
Here are sprint time simulations for the two different drivetrains with NEOs at 150 lbs:
Normal swerve:
Your Omni/swerve:
Thereâs a 5% improvement over the full field for the omni swerve at 16.5 FPS maximum speed, and a 11% improvement in minimum sprint time. This advantage goes down for lighter robots, more powerful motors, or a shorter sprint distance.
This is not even comparing it to three motor swerve, which is better in every way.
I would highly recommend running your design by a physics teacher or mechanical engineer if this explanation is unclear.
How are you proposing this should work when the bot is traveling diagonally?
No arguments there! Controlling this well would certainly be a great challenge. I do think that shifting swerve or three motor swerve would probably provide you a greater competitive advantage.
Edit:
I think I misunderstood your concern with regards to defensive ability. To respond to your actual concern, Iâd have to think a little bit about the physics, but in practice does a pushing tank actually spin the swerve modules? Iâve not really seen this happen on the field.
As far as your original question about current draw, itâs hard to provide general advice without knowing the weight of your robot, the size of your wheels, and the proposed gearing.
this is what (more or less) I was looking for when I started on these designs. simple math thatâs easy to understand.
thank you, and also I have apparently misspoken, as I have limited experience, tank apparently rarely pushes swerve around, we were just stupid heavy in our matches and had more power than one of the swerves, and I based it off that.
I have no in-school resources I can turn to, any and everything I do is learned via my own studies. even the swerve the team is making is almost entirely done by me and I cannot find the correct calculations
I will attempt to run simulations to determine if this chassis system is at all effective, but I do like it and already have a good few of the team rallied behind it (not even my fault, blame the build team captain who came up with it).
While I would firmly land in the group of âmodify a SDSSwerve Drive Specialties module to accept a 2nd drive motorâ, if you do go down the more custom path I would suggest checking out The Deceiversâ build thread and their modules theyâve been making for a few seasons.
I will gladly check this out! at the moment the major complication is one of cost, we simply cannot afford COTS swerve, so Iâve taken up the challenge of designing my own swerve, and am currently drawing promising calculations, I donât expect it to be the best, but I do think it is a major step up for our team. With every iteration, my designs get tighter and faster, and Iâve recently been paired with a professional engineer! these designs from everyone look spectacular, and I hope to match these in complexity after this coming year, where I will be attending many engineering courses. thank you everyone for the help!
Programming that is going to be fun.
(not sarcasticâI would enjoy programming that)
Mechanically this will be a mess because you are asking the eight wheels to maintain relatively uniform contact force with the floor. If we add suspension it will be a mechanical mess for different reasons.
FTFY. lbf is a unit of force, not power. I know everyone and their uncle calls it âpushing powerâ but that is simply incorrect.
This seems like a false choice fallacy.
If you spent this design time fundraising instead I bet you could afford the COTS modules.
COTS modules are cheap for all of the components they contain, so unless you somehow eliminate a bunch of bearings/gears/wheels/sensors you are not going to save any real money.
I would work through the above points at least as a thought exercise before you go too far down the rabbit hole of designing your own swerve âto save money.â
OTOH if you want to design your own swerve module because it is fun/interesting/good learning, more power to you.
from the way I see it, as we do have a machine shop, and the only sensor we really need is a REV 1/2 in through bore encoder. we can make a design cost less than $100 a module, because of our privileged position. As the team has agreed we only see two options - purchase a COTS swerve and be a chassis bot, or develop a SWERVE and be an everybot.
plus, I just donât agree with winning by cost, this isnât a team opinion but a personal one.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
