FRC 48 Team E.L.I.T.E. - xtrēmachēn 26 CAD Release

[alex.richards48]

Team E.L.I.T.E. is proud to present the CAD model for our 2023 robot, xtrēmachēn 26!

Note: Our team uses SOLIDWORKS, included is a Pack & Go, as well as a .step file of the main assemby.

This is our 26th season of Encouraging Learning In Technology and Engineering, and this year our team won 2 technical awards in one season, something we haven’t done since 2004. At the Buckeye Regional, we were awarded the Creativity Award sponsored by Rockwell Automation for our 3 jointed arm design which prioritized a low center of gravity, while allowing the robot to score and acquire game pieces on both sides. At the Greater Regional, we were awarded the Quality Award for our use of robust, billet aluminum arm pivot points. This award came to us at a bit of awkward moment. During our last match of GPR, we suffered quite a failure.

The Failure in Question

That’s right, due to some oversight in the manufacturing during build season, and multiple high velocity impacts in the season, our weld failed, and our arm separated. Only moments later were we awarded the Quality Award.

After bouncing back from the struggles and frustrations of the outcome of Match 13 at GPR, we traveled to the Ohio FRC State Championship in New Albany, Ohio. After qualification matches, we managed to seed 3rd overall, and moved into the 2nd alliance captain position. With our 1st pick of 325, and our 2nd pick of 5667, we were able to put up (to my knowledge), the event’s playoff high score of 173. We finished in 3rd place at OSC, and are gearing up to travel to Worcester, MA to compete at Battlecry@WPI23. We look forward to seeing some of the friends we made while attending BC22 last season, and are eager to make new friends this year.

I’d like to thank all of our alliance partners this season for playing with us, and making the most of every elimination run. We’d also like to thank team 111 for their build blog which helped inspire the 2nd iteration of our gripper. Without everyone’s help, we would not have been able to have the season we had, and we’d like to wish everyone good luck in the upcoming 2024 season, Crescendo!

Team E.L.I.T.E. 48 Season Summary

• Midwest Regional - 3rd place finish as 8th alliance 1st pick (Thank you to 379 and 2013)
• Buckeye Regional - Finalist finish as 3rd alliance 1st pick (Thank you to 4611, 3201, and 5413) - Creativity Award sponsored by Rockwell Automation
• Pittsburgh Regional - 3rd place finish as 3rd alliance 1st pick (Thank you to 3171 and 2872) - Quality Award
• Ohio FRC State Championship - 3rd place finish as 2nd alliance captain (Thank you to 325 and 5667)

Here are the flyers our students hand out while presenting to technical judges:

For any questions about our 2023 robot, feel free to reach out to myself or @A_Reed in this thread, or though DM’s. Happy offseason!

[AriMB]

Looks very nice. You might want to clean up the folder structure of the Pack & Go though, it took some time to find where the main assembly is.

I have a few questions about the robot:

• It looks like the intake side plates are polycarb connected by a few churros. Was this stiff enough, or did you have problems with the intake sagging on one side?
• How much did the whole arm weigh? Was one NEO powerful enough to drive the bottom joint?
• It looks like the J1 axle is just 1/2" Hex shaft instead of a tube axle like the J2 and J3 joints. Why did you choose to do it that way?
• Some of the machined parts are really complex! Did you have access to a multi-axis mill or was it just lots of setups and patience?
• The 0.7" blocks above the MAXPlanetaries seem to be some kind of chain tensioner, but I’m not seeing how it works. Can you explain?

[alex.richards48]

Thanks for the questions Ari, let me know if this helps!

@AriMB

• It looks like the intake side plates are polycarb connected by a few churros. Was this stiff enough, or did you have problems with the intake sagging on one side?

Through some iterations, we managed to make the sagging as small as we could, to the point it hardly affected functionality. We never measured the deflection, but if I had to estimate, the first iteration sagged about 1/2"-3/4". We started the initial concept with this gripper, which has cones and cubes coming in from opposite directions. Both of the side plates are 1/8" polycarbonate.

We noticed the sag that this first iteration was experiencing, and waterjet a 1/4" polycarbonate replacement for the plate that interfaced with the J3 pivot. This helped to reduce the sagging significantly. Now, it would sag 1/4"-3/8". You can see the thickness change in this picture of the robot right before we left for the Midwest regional.

When we got back from Midwest, we knew immediately that we wanted to add polybelts to our gripper after seeing the success that 111 and 930 had at the event with it. We even had the next iteration designed before playoffs on Saturday! We stuck with the polycarbonate plates for the final iteration, and still maintained the usage of 1/4" plate on the J3 side, and 1/8" plate on the motor mounting side. We put a heavy emphasis on making the gripper as light as possible, and I think the final version worked out to being less than 7lbs. You can see the final iteration here.

The final version still sags, but the only time this causes somewhat of a problem for us is when loading upright cones on the floor. The rollers being angled from the deflection can cause the cone to tip. Thankfully, it tips right into the belts and is taken in pretty easily. We primarily load cones from the single substation, and cubes from the floor. When floor loading cubes, the deflection is “fixed” by holding the gripper on the carpet. Our arm position for loading cubes from the floor is the same as grabbing tipped cones from the floor, so we see no issues from that either.

@AriMB

• How much did the whole arm weigh? Was one NEO powerful enough to drive the bottom joint?

Solidworks reports the weight of the arm as ~28 lbs. Adding in fasteners (we only model the critical fasteners, as seen in the J1 gearbox), any “as-builts”, etc., we’re likely closer to the 32-35 lb range. We utilized the REV MAXPlanetary’s higher load capacity to our advantage, and ended up with a 607:1 final gear ratio on J1. This gave us an estimated loaded current draw of ~12A. Our final gear ratios for J1, J2, and J3 are as follows.

• J1: 60:1 MaxPlanetary reduction (5:1, 4:1, 3:1), 70:24 gear reduction, 52:15 chain/sprocket reduction - 607:1 overall
• J2: 25:1 MaxPlanetary reduction (5:1, 5:1), 70:24 gear reduction, 44:15 chain/sprocket reduction - 214:1 overall
• J3: 60:12 spur gear reduction (metal), 36:12 reduction (Onyx herringbone internal to gearbox), 40:20 reduction (final drive Onyx herringbone - visible) - 30:1 overall

@AriMB

• It looks like the J1 axle is just 1/2" Hex shaft instead of a tube axle like the J2 and J3 joints. Why did you choose to do it that way?

We chose to go with this method for a few reasons. First, it was simpler and we have a bunch of 1/2" steel hex that worked perfectly. Next, it allowed us to easily use the shaft as a dead axle construction to power J1, while being a live axle construction to power J2 from the main gearbox. Bearing are pressed into the lower billet pivot of J1, and torque is transmitted through the bolted on plate sprockets. We made a plate sprocket stack with some 1/2" hex hubs, and ran #25H chain up to the plate sprocket on the J2 arm to transmit torque. The tube axle construction of J2 and J3 is for our pass through wiring. We didn’t need to have the wires pass through J1 as we could protect the wires just as much as the through-bore method by dropping them through the gearbox.

@AriMB

• Some of the machined parts are really complex! Did you have access to a multi-axis mill or was it just lots of setups and patience?

Yes they are! And they took quite a bit of patience to manufacture. In house, we have a Tormach PCNC 1100 that our CNC manufacturing mentor (also a sponsor of ours) uses with the students. At that same mentors workshop, he uses a Haas VF-2 with a HRT160 4th axis, but we did not need the 4th axis for these parts. The lower J1 joint is a solid block of aluminum, and took multiple operations to complete. All other joints that get inserted into the tubes are welded to a tube insert. This method let us split the work, and made machining in the 5 deg offset of the J2 pivot much easier.

@AriMB

The 0.7" blocks above the MAXPlanetaries seem to be some kind of chain tensioner, but I’m not seeing how it works. Can you explain?

Sure, here is a section view of the chain tensioner. The two outside pieces that are sandwiched together house a leg that gets a threaded coupling pressed into it. A 1/4-20 flat head bolt is captured in the sandwiched pieces, and when the screw is spun, an HDPE roller on the end of the leg pushes into the chain, and applies the tension. In this case, it is spun to the left to tighten the chain, and to the right to loosen the chain. These worked very well for us, and made tensioning the chain very easy. These were 3D printed on our MarkForged printers out of Onyx.