Team 5550 - Fall Project

[rlance]

We decided to try using those omni-directional wheels and make a holonomic drive robot. Here is a video of what we have so far. Great learning experience for the kids. Does anyone have any concerns about the wheels being at the corners like we have them?

https://youtu.be/c_J0yFwFDXM

[Ari423]

Looks good. Two suggestions though:

1. It looks like the wheels are protruding outside of the frame perimeter (assuming there would not be an additional superstructure to increase the frame perimeter). According to 2016 rules, that would be an illegal starting configuration.

2. Right now when driving straight forwards or sideways you only have 2 motors running at a time, which is rather inefficient. Combine this with the fact that you're using omni wheels means you will get pushed around like nobody's business. If you turn each module 45* so they are forming an O around the chassis, when moving forward you will have 4 wheels providing sqrt(2)/2 of force for a total of 2*sqrt(2) of force, effectively increasing your force by sqrt(2). This should make your drivetrain even faster when going straight forward or sideways and make it slightly better in a pushing match (though omnis are notorious for losing pushing matches).

[roboruler]

Originally Posted by Ari423

Looks good. Two suggestions though:

1. It looks like the wheels are protruding outside of the frame perimeter (assuming there would not be an additional superstructure to increase the frame perimeter). According to 2016 rules, that would be an illegal starting configuration.

The allowable "frame perimeter" is based on the size of the entire robot, think of it as a box of a certain size that the robot needs to fit into. It is not dependent on the actual size of the frame/chassis/drivetrain.

If you apply that logic you are basically ruling all cantilevered wheel drivetrains ie WCD illegal( excluding bumper mounts)

This is a fun, valuable learning project rather then something that will be ever used in a competition.

This is a terrible style of drivetrain to use in any season( even in 2015 were it is unnecessary) so I don't believe they are intending for it to be legal.

[Ari423]

Originally Posted by roboruler

The allowable "frame perimeter" is based on the size of the entire robot, think of it as a box of a certain size that the robot needs to fit into. It is not dependent on the actual size of the frame/chassis/drivetrain.

If you apply that logic you are basically ruling all cantilevered wheel drivetrains ie WCD illegal (excluding bumper mounts)

This is a fun, valuable learning project rather then something that will be ever used in a competition.

This is a terrible style of drivetrain to use in any season( even in 2015 were it is unnecessary) so I don't believe they are intending for it to be legal.

Originally Posted by 2016 Game Manual R2

The ROBOT (excluding BUMPERS) must have a FRAME PERIMETER, contained within the
BUMPER ZONE, that is comprised of fixed, non-articulated structural elements of the ROBOT.
Minor protrusions no greater than ¼ in. such as bolt heads, fastener ends, and rivets are not
considered part of the FRAME PERIMETER.

Emphasis mine. I'm not a robot inspector so my opinion is only my opinion, but I would argue that a spinning wheel or axle is not a "fixed, non-articulated structural element" and therefore not part of the frame perimeter. I think the "excluding bumper mounts" exception you made is important, because the bumper mounts are fixed, non-articulated structural elements and therefore count in the frame perimeter, extending it past the wheels of a WCD. That is why I suggested that if they do choose to use this drivetrain or something similar in competition, they should add an additional superstructure (like bumper mounts) to increase the frame perimeter past the wheels.

[rlance]

Thanks for all replies

Originally Posted by Ari423

1. It looks like the wheels are protruding outside of the frame perimeter (assuming there would not be an additional superstructure to increase the frame perimeter). According to 2016 rules, that would be an illegal starting configuration.

If we did try and use something like this next year's game we would definitely add a structure to extend the perimeter if bumpers are needed for the game. Which somewhere it was mentioned that bumper mounts were not included in the KOP drivetrain this year.

Originally Posted by Ari423

2. Right now when driving straight forwards or sideways you only have 2 motors running at a time, which is rather inefficient. Combine this with the fact that you're using omni wheels means you will get pushed around like nobody's business. If you turn each module 45* so they are forming an O around the chassis, when moving forward you will have 4 wheels providing sqrt(2)/2 of force for a total of 2*sqrt(2) of force, effectively increasing your force by sqrt(2). This should make your drivetrain even faster when going straight forward or sideways and make it slightly better in a pushing match (though omnis are notorious for losing pushing matches).

We had a whole discussion on which direction is fastest with this set-up and if speed was an issue we would definitely either make one of the corners the forward direction for the robot or do the 45 degree angle thing.

Originally Posted by roboruler

This is a terrible style of drivetrain to use in any season( even in 2015 were it is unnecessary)

Just curious if you think all holonomic drivetrains are terrible or just omni-wheels ones, or just because of the way we have ours attached or placed on the frame. If you could clarify, that would be helpful for us.

[nickbrickmaster]

Holonomic (4-omni) drivetrains are generally bad because there are better, more efficient, or higher traction options. Mecanum in particular is better than a 4-omni holonomic drive, with the same general setup. It's a bit more expensive, but more efficient (If anyone knows the exact theoretical numbers, please let me know.)

So yeah, in a competition (and you reaally wanted to use omnidrive), I would look into using Mecanum instead of this setup.

[Lil' Lavery]

Originally Posted by nickbrickmaster

Holonomic (4-omni) drivetrains are generally bad because there are better, more efficient, or higher traction options. Mecanum in particular is better than a 4-omni holonomic drive, with the same general setup. It's a bit more expensive, but more efficient (If anyone knows the exact theoretical numbers, please let me know.)

So yeah, in a competition (and you reaally wanted to use omnidrive), I would look into using Mecanum instead of this setup.

I challenge you to show me the math (or at least the reasoning) to why you believe Mecanum drives to be more efficient than Omni drives.

Hint: Consider both drives moving not only in "straight" directions, but also diagonals.

As for traction, Vex quotes a CoF of 1.0 for their Mecanum wheels and a CoF of 1.1 for their Omni wheels.

[nickbrickmaster]

Originally Posted by Lil' Lavery

I challenge you to show me the math (or at least the reasoning) to why you believe Mecanum drives to be more efficient than Omni drives.

Hint: Consider both drives moving not only in "straight" directions, but also diagonals.

As for traction, Vex quotes a CoF of 1.0 for their Mecanum wheels and a CoF of 1.1 for their Omni wheels.

...Yeah :|

Alright, that'll teach me to repeat information without a source and critical thinking. Drawing the vectors makes them look like the same.

Then, why does one never see a 4-omni holonomic drive used in FRC?

[rlance]

Originally Posted by nickbrickmaster

...Yeah :|

Alright, that'll teach me to repeat information without a source and critical thinking. Drawing the vectors makes them look like the same.

Then, why does one never see a 4-omni holonomic drive used in FRC?

In last year's game if your robot's goal was to collect balls and put them in the lower goal as fast as possible, wouldn't the 4-omni holonomic be the fastest option. You could easily avoid shoving matches by zipping around the slower robots. Am I missing something in that thinking, because I think "Nick the Brick" poses a good question.

[Andrew Schreiber]

Originally Posted by nickbrickmaster

...Yeah :|

Alright, that'll teach me to repeat information without a source and critical thinking. Drawing the vectors makes them look like the same.

Then, why does one never see a 4-omni holonomic drive used in FRC?

You do occasionally. But why go through the hassle to build that frame when you could slap mecanums on a more traditional frame.

[rlance]

Originally Posted by Andrew Schreiber

You do occasionally. But why go through the hassle to build that frame when you could slap mecanums on a more traditional frame.

I think that putting this frame together was a whole lot less hassle than putting mecanums on the KOP drivetrains. Maybe there are already holes on that frame for mounting a motor and gearbox on each wheel. Even if it did we didn't have those gearboxes in our stock. We had plenty of planetary gearboxes and mounts for those gearboxes to make it fairly simple to build.

[Kevin Leonard]

Originally Posted by rlance

In last year's game if your robot's goal was to collect balls and put them in the lower goal as fast as possible, wouldn't the 4-omni holonomic be the fastest option. You could easily avoid shoving matches by zipping around the slower robots. Am I missing something in that thinking, because I think "Nick the Brick" poses a good question.

Because "easily avoiding shoving matches" is a lot harder than you're making it sound.

A well driven 4-8 wheel drive is easier to build/program, and just as good as a non-swerve holonomic drivetrain, if not better in 99% of all scenarios in FRC.

Even in 2015, if on Team 20, we could go back and do it again, we would not have gone with a mecanum drive that year, even with the lack of defense. It wasn't necessary, it caused us to spend too much time working on things that weren't scoring mechanisms, and occasionally tethers would get caught in the rollers of the wheels.

[Cothron Theiss]

Originally Posted by roboruler

The allowable "frame perimeter" is based on the size of the entire robot, think of it as a box of a certain size that the robot needs to fit into. It is not dependent on the actual size of the frame/chassis/drivetrain.

If you apply that logic you are basically ruling all cantilevered wheel drivetrains ie WCD illegal( excluding bumper mounts)

If you work with a team in the near future, for their sake, I suggest you watch this video and read this thread.

Many teams had to redo their bumpers and frame perimeter either late in the build season or on the first day of a competition because they interpreted the bumper rules the same way you've described them.

[Lil' Lavery]

Originally Posted by nickbrickmaster

...Yeah :|

Alright, that'll teach me to repeat information without a source and critical thinking. Drawing the vectors makes them look like the same.

Then, why does one never see a 4-omni holonomic drive used in FRC?

Back in FIRST antiquity the early/mid 2000s, you did see more holonomic drives than mecanums. Prior to the 2007 season, there weren't really any COTS options for mecanum wheels, so the few teams that used them had them custom made (namely 357). In the fall of 2006, AndyMark released the first FRC-specific mecanum wheel, and it became a bit of a cultural phenomena within FIRST. At the time, the COTS omni-wheels (also primarily sold by AndyMark) were still several generations behind what they are today, and weren't the greatest option, so when combined with the hype of the mecanum wheels and the relative ease of using a mecanum drive with a "standard" rectangular frame made them the preferred option. Mecanum drives never really looked back from there in terms of being the default choice for teams looking to move sideways.

Additionally, there is one major benefit of mecanum wheels when compared to omni set-ups when it comes to uneven flooring. Unless you're robot has a suspension or is doing some really fancy sensing and drive automation, generally speaking both mecanum and omni wheeled set-ups are reliant on each of the wheels being constantly in contact with the floor AND having a similar load on each wheel. If one wheel has a higher load on it, it will have more traction and exert more force than the other wheels, which can cause some unpredictable driving. Naturally, even worse examples of this occur when a wheel completely loses contact with the floor. Mecnaum wheels have a benefit that they can "climb" obstacles on the floor if approached dead-on such that both leading wheels make contact (more or less) identically. Think of a mecanum wheel rolling up a ramp. That type of maneuver is much more difficult to pull of in an omni-wheel set-up with the wheels spaced every 90 degrees. Thus why you saw mecanum wheeled robots climbing the bumps in 2010 and 2012, for instance.

And to answer your implicit question regarding 3 vs 4 wheeled (or even 5+) omni drives, think back to that previous paragraph. The reason many teams, such as 1114 in 2015, opt for a 3-wheel omni drive ("kiwi drive") is simple. It takes 3 points to make a plane. With only 3 wheels, it guarantees that all 3 of your wheels are in contact with the ground at all times. With 4 or more wheels, even small obstacles in the field can cause a wheel to lose contact with the floor and for erratic driving. I found this out the hard way in 2005, when my team at the time (116) built a 4-wheeled omni drive system, and even the small plastic loading zones (0.5" thick?) were enough for our wheels to sometimes lose contact with the carpet, which caused driving issues. Mecanum drives will also suffer similar consequences if they don't approach obstacles so that all 4 wheels remain in contact with the ground (plenty of teams had issues with the scoring platforms in 2015 if they contacted with only 1 wheel). Conversely, the biggest advantage today of a 4-wheeled omni drive is the simplicity of the frame construction (don't have to worry about a triangular drive base). Depending on the frame perimeter and sizing rules, there have also been points in time where a triangular frame would be space inefficient (or result in significant portions of your robot overhanging beyond your drive base, causing instability issues). Also, there was a time where the limitations on CIM motor quantities were capped at 4, meaning teams would opt to find ways to include all 4 in their drivetrains (difficult with 3 wheels).

[dirtbikerxz]

Not exactly on topic, but if you guys are trying to evaluate drive train options, I would just like to point out this great video starting from 48:38 and continuing for about 5 minutes. https://youtu.be/MW6KqtDR9_U?t=48m38s