Flipping Swerve V1.0

One of the disadvantages of swerve, apart from the fact that it’s apparently rather hard to program, is the fact that it’s not the greatest for pushing and it’s also rather easy to get stuck if the game has random stuff scattered all over the floor, like 2016.

Introducing: Flipping Swerve V1.0

Flipping Swerve is an idea I had at the beginning of September, about 8 months ago. I’ve been working on it on and off for a while now, but it’s finally finished!
Flipping Swerve is a combination of swerve and jump drives, combining the best of both worlds. Currently, it features a modified WCP-SS style swerve drive combined with a custom tank tread drive, while only using 8 motors - 4 Falcon 500’s and 4 RS-775pro’s.
CAD Link: https://grabcad.com/library/flipping-swerve-v1-0-2


Features:

  • Mass: 50.89 lbs (According to SW, but I probably missed some stuff)
  • 15.75:1 overall ratio for the tank tread section, giving it a calculated speed of about 8.59 ft/s and a maximum current draw of about 56.67 amps
  • 5:1 overall ratio for the swerve drive, giving it a calculated speed of about 18.04 ft/s and a maximum current draw of about 117.35 amps
  • 4 Falcon 500’s with Thrifty Bot CIM style shafts - I was using NEO’s until Falcons came out and couldn’t be bothered to modify the mounting to use Falcon splines. That’ll change in the next version I make.
  • The swerve drive is almost identical to WCP’s SS Swerve, with the main differences being the modifications I made to allow the tank tread to use the same drive motor as the swerve, as I removed the encoder connected to the drive motor and replaced it with a bit of gearing to add in the tank tread
  • The “flip” uses the Vexpro VersaDrop
  • The current estimated cost is about $3300, which is somewhat high to use on a robot, unless the year is 2016 and you don’t want to have anything but a drivetrain.

Approximate BOM
Part Link Quantity Price Each Price Total
WCP SS Swerve https://www.wcproducts.com/swervess (Configuration below) 4 $353.94 $1415.76
Falcon 500 motor https://www.vexrobotics.com/217-6515.html 4 $139.99 $559.96
1/4" Aluminum plate, 10.5"x8" https://www.onlinemetals.com/en/buy/aluminum/0-25-aluminum-plate-6061-t651/pid/1248 4 $11.94 $47.76
1/4" Aluminum plate, 6.21875"x6" https://www.onlinemetals.com/en/buy/aluminum/0-25-aluminum-plate-6061-t651/pid/1248 4 $11.94 $47.76
VexPro VersaDrop https://www.vexrobotics.com/217-4824.html 4 $129.99 $519.96
15T Bevel Gear https://www.vexrobotics.com/bevel-gears.html 8 $24.99 $199.92
20DP 20T 1/2" Hex Bore Gear https://www.vexrobotics.com/1-2-hex-bore.html 4 $9.99 $39.96
20DP 30T 1/2" Hex Bore Gear https://www.vexrobotics.com/1-2-hex-bore.html 4 $10.99 $43.96
16T #25 Sprocket https://www.vexrobotics.com/25-sprockets.html 4 $7.99 $31.96
54T #25 Sprocket https://www.vexrobotics.com/25-sprockets.html 4 $14.99 $59.96
Vex VersaHub, 1/2" Hex Bore https://www.vexrobotics.com/versahubs.html 8 $8.99 $35.96
Brecoflex 48T TK10 K13 50mm Pulley https://www.brecoflex.com/products/pulleys/stock-pulleys/tk10-pulleys/#tk10-k13-pulleys-50mm 4 Est. $30 Est. $120
Brecoflex 135T T10/1350 BFX Belt https://www.brecoflex.com/products/timing-belts/t-series/#t10-bfx 2 Est.$20 Est. $40
2x1 0.1" Thick Aluminum Tubing,59" https://www.vexrobotics.com/versaframestock.html 3 (Technically exactly 2) $34.99 $104.97
Various Hardware McMaster/VexPro - Est. $30 Est. $30
Total - - - $3297.89

I hope you enjoy the sheer ridiculousness of this, but I had a lot of fun making this, and I'm probably gonna make another version that's quite a bit cheaper and lighter, as this design is just a bunch of random things thrown together in a quasi random way.

I’m happy to answer any questions you might have, but please keep in mind that this isn’t really meant to be practical, I made it because I can and it’s something I’ve never seen anyone else do before. However, if someone does decide to use this design, please let me know, I’d love to see how it’s implemented and tweaked. (Also please let me know if you play with the CAD and change some stuff up, I’d appreciate seeing what you do with this)

20 Likes

I don’t think either of these things are necessarily true?

19 Likes

Depends on the type of swerve, but I’m of the opinion that tank treads/tank in general is generally better for pushing. I don’t remember seeing a lot of swerve robots getting beached (although I don’t remember a lot of swerve robots in 2016 anyway), but I remember a lot of robots without tank treads getting beached a decent amount in 2016.
But yes, you’re right, neither is 100% true. :slight_smile:

image

33 Likes

I’ll just point out the main disadvantages of swerve are weight, then packaging (Then maybe construction). Pushing power is almost never a “disadvantage” of swerve, in fact, its one of the better DTs for pushing power being its also holonomic.

So basically you took swerve and to fix something its not that bad at, you made everything bad about it so much worse…

The 2016 example is bad since 2016 just wasn’t a good year for swerve… Its not always good to use a holonomic dt, 2016 is just a bad example.

(Sidenote, gearing on that swerve is a bit too high)

4 Likes


You probably want to support both ends of at least one bevel gear, and limit cantilever otherwise. The single bearing on that driven shaft and any flex in the material would make it skip instantly.

12 Likes

I was half expecting a reference to the team from 2010 whose answer to the giant berms was simply “Let’s stick a swerve drive on both sides of our robot.”

1 Like

I think you’re going about this all wrong. Hard mount the tank drive to the chassis and articulate the swerve modules. Much more sensible.

4 Likes

Can confirm that swerve makes a very good pushing robot if you max out the robot weight so that your wheels don’t slip. Being able to push in any direction as well as being able to position yourself between the robot that you want to push and the object you want to push them away from makes swerve a very good drive base for pushing.

Beaching (high centering) is a factor that needs to be considered in any drivetrain. Multi-wheel tank drives have a short enough spacing between the wheels to generally not get beached when driving forward or backward over obstacles, but there is still a decent distance between the wheels laterally across the robot. I have seen plenty of tank drives that high centered due to this lateral spacing. To address this for swerve, you can add idler (non-driven) omni-wheels between the other wheels that are raised up above the floor (they only need to be lower than the bottom of the bumpers to be pretty effective at preventing beaching), or, if the terrain is more challenging than that you can try something a little more crazy…

While this design is interesting, I’m not sure it solves the beaching problem since the wheels at the corners are still very far apart leaving a large span in the middle on which to high center the frame or bumpers of the robot.

i’m also not certain that it improves the pushing power since you still have the same number of motors with the same number of wheels in contact with the floor. So where does the extra traction come from?

3 Likes

Needs more falcons.

And more mecanum

4 Likes

Did you factor the weight of a pneumatics system (mainly, a compressor) into the weight estimate?

5 Likes

I wouldn’t say that necessarily, mainly because it’s much harder to beach now, but as I said, it’s not really practical except POSSIBLY in a year like 2016.

I was surprised when I saw it, I might have mistyped a number but it looks like that’s what WCP uses.

Yeah, that’s a good point. Shoulda thought about that.

I didn’t put this in the CAD, although I now realize I should have: I’m using pulleys with high grip belts (See BOM section for details) as a tank tread system, which solves the beaching problem and increases traction.

No, and I don’t plan on doing so. And that gif is totally me.

If I’m understanding correctly, wouldn’t this make it so pairs of modules were forced to drive at the same speed? This would prevent you from using the swerve as it ideally should be able to be used.

4 Likes

OK, I missed that, but now that you have pointed it out, I can see that the wheels you are using at the corners are, in fact, the tank tread wheels (I see the groove in the middle for the track).

So, I was going to point out that you probably want the swerve wheels as far out toward the corners as possible and move your “pushing” wheels (or now the tank tread) toward the middle since you want to make sure you have the tipping stability when you are going fast and swerving, but you don’t really need that level of stability when you are going slow and pushing straight.

But if you did that, then your drivetrain design would start to look an awful lot like this one

1 Like

Thanks for pointing that out. I’m gonna look at the design and see if there’s something I can do, maybe setting up the system to disengage the wheels would work. I wonder if a ball shifter could do something without adding much weight.

The reason this is somewhat true is as you push someone, it is decently likely that the end of the robot doing the pushing could get lifted off the ground a bit. With skid steer DTs, you would still have all of the motor power on your back wheels. With swerve, you would only have 2 motors of drive power on your back wheels since the DT power on swerve drives are localized to each wheel.

That said, I think this is a minor effect in reality and not much concern. In general the goal of using a swerve drive shouldn’t be to push if you’re playing defense, it should be to obstruct. I believe this is generally the case regardless of FT type. With everyone’s DT getting more performance with the prominence of the brushless motors, it will be less and less common for pushing defense to be the most effective play style.

5 Likes

Needs more Compliant wheels.

11 Likes

Generally swerves are limited to 4 drive motors, and could be out-pushed by something like a west-coast drive with 6 motors that are geared similarly. However, with the brushless drive motors packing so much power, I’m seeing more and more teams use just 4 motors on their west-coast drives. So from that perspective, the pushing power is going to be very similar. And, of course, a swerve can apply that pushing power in whatever direction it wants, which is not always possible with other drive trains.

But I think that a few other posters here hit the nail on the head: with swerve, you are generally trying to out maneuver your opponents, rather than trying to out push them.

My sole reaction to this:

8 Likes

honestly its late and tldr into the details, but WHAT IS A FLIP DRIVE