Revisit: 3d Printing a swerve drive

Technical Technical Discussion
,
#1

OK, bear with me, Ive had this talk about 3d printing a Swerve drive before
(3d printing an entire robot?),
I’ve deemed it (mostly) impracticle… BUT, ive been looking to some stronger 3d filiment like onyx and want to revisit the idea.

things ive already learned:

  1. you CAN print 3d printed gears, as long as you plan to replace them often (or just use metal ones
  2. The frame of the Swerve absolutely can be 3d printed
  3. if its not meant to be a HIGH impact design, it would be great for examples on how one works to newcomers…

NOW, i want to hear what the community thinks, if i got myself a decent printer and printed strong enough materials, how would i go about printing a swerve?

1 Like
3175 Concept Swerve
#2

IIRC, https://www.thebluealliance.com/team/5817/2017 ran at least some 3D printing in their swerve. I can’t recall if they’ve continued doing that since then, though.
@Mitch_Stokes, @steelerborn, care to chime in?

1 Like
#3

We haven’t seen their 2020 robot in action, but they haven’t run swerve since that 2017 robot.

1 Like
#4

https://youtu.be/SJ-yE_yEDnY 3:17 in this video they talk about 2767’s largely 3D Printed swerve.

3 Likes
#6

Just going to casually drop this here:

https://markforged.com/additive-manufacturing-movement/nc-robot-wheel-module/

10 Likes
#7

We ran a mostly wooden (laser cut plywood) and 3D printed swerve for an off season comp. Plywood was baltic birch, and the printed materials were both PETG (okay choice) and SLA resin (terrible choice).

Swerve was functional, but didn’t do well when put on the field with other robots.

I really think it could be done, even without carbon fiber Nylon, but certainly not the way we did it. I think with the large X contact bearings, it becomes more and more reasonable to do.

1 Like
#8

For 2020 team 1640 redesigned our module with many 3d printed parts. The wheel forks, main steering pulley and many other parts were printed from PETG. The one part that PETG was not sufficient was a V belt pulley for our CVT. We had to go to a nylon carbon. In the preseason, the modules were stress tested and the 2020 competition robot had one district. The printed parts held up. We are not afraid of the barriers.

8 Likes
#9

Are the bevel gears printed too? Did you try printing it in a different material (for testing purposes) before printing in Onyx?

#10

We did one (I think?) in PLA to test for some fitment. The gears are from WCP.

2 Likes
#11

2767 swerve is printed.
@Nick_Coussens has a printed swerve.
I believe 910 ran one.
There are others.

When I say printed swerve I mean one that the main structural components are printed AND has been in competition.

We have a Mark Forged. I think the secret sauce is the carbon filled plastic NOT so much the carbon fiber. We use carbon fiber in some areas but Nick has proven that it can be done without carbon fiber if you pay attention to where the stresses are going.

There are a boat load of stuff on a robot that can be printed even out of “cheap” plastic. Pulleys are a good example. The hardest part is attaching them to the shaft.

Another example, in the posted video, the actuator that the trident swings around on is a printed gearbox, gears and all. Even though the trident was ripped off or broke multiple times, the gearbox held up well.

On that robot, just about everything that is black and looks like plastic is printed. Elevator parts, drive motor gear boxes, motor mounts, suction cups, etc.

4 Likes
#12

We printed the large bevel (45 tooth, 1.25 mod) and abuse tested it on the bench with super good results. Put it in a robot and Poof!, plastic debris in 2 minutes.

From looking at the tooth damage, we concluded that the wheel/saddle assembly was flexing from the robot turning. Steel gears are more forgiving when it comes to the tooth contact/backlash opening up but with the plastic gears, the small bevel climbed on top of the large bevel. We didn’t see evidence that the teeth were shearing. They were being topped. We felt that if we could keep the gears from moving, and keep them tight, it might have worked. We abandoned the idea when we considered how much carpet debris gets up in there.

4 Likes
#13

2471 ran a swerve that was mostly 3d printed in 2019, although they used a more traditional aluminum plate design in 2020. The main housing of the module was printed on a Markforged, I believe in Onyx filament.

#14

Were you able to get fiber into the teeth of the bevel gear? Or were they pure onyx?

#15

Onyx only. The teeth aren’t big enough to put fiber in. They have to be really big to do that. In order to take advantage of fiber you also need to take into account how much plastic is surrounding the fiber. Sometimes adding fiber can weaken the part since fiber/plastic bond isn’t as strong as plastic/plastic bond.

Kinda like re bar in cement.

#16

@EricH haven’t been involved in FRC in 3 years or so, but the globoid gear system on the 2017 was 3D printed in ABS and worked really well, 0 failures through the season iirc. We lubricated them by running it against a toothbrush. 5817 hasn’t used swerve since that year.

2 Likes
#17

Interesting. I never thought about it like that. I remember discussing the fiber reinforced 3d printed gears in the arm of your 2019 bot with someone at IRI. Do you know offhand if you added additional onyx floors, walls, or infill to these? I usually just use the standard 2 walls, 4 floors, and 37% triangular infill.

#18

How often did you have to switch the gears out in the trident gearbox throughout the season? We are thinking of printing a couple gears for a gearbox using onyx, we just worry about the structural integrity to the teeth spinning at 200 RPM.

#19

Same worry honestly, ive seen some teams get around this by instead of making “tooth to tooth” contact for their teeth, they make a pully system using timing belts, looks like it works. But i am curious if you do go “tooth to tooth contact” for gears, how well do they hold up?

#20

100% fill. Here’s a screen shot. This is the planet used in the “biscuit” in 2109.
For reference the hole in the center is 1/2" thunderhex. These are big teeth. We haven’t proven that the fiber is needed in the teeth and my gut feel is it isn’t needed for the teeth buuuttt… Big warning here: We’ve experienced poor results driving high loads through thunderhex on Onyx so user beware. The fiber has proven to greatly improve the torque transmission to the hex since it’s all about the hoop strength around the hex.

image
image1494×750 68.7 KB

The ring gear sees a bunch of hoop stress so there is fiber in the circumference of the ring. (not pictured).

1 Like
#21

The only issue we had with this gearbox was the thunderhex interface. We R&Red the planet gear a few times during the season because it did get sloppy after getting beat up when the trident would get caught in the rocket .

Teeth where fine. In an actuator like this one, the last stages of gears are moving slow and their mileage is low.

1 Like