User Tools

Site Tools


builds:invisible_jet

600

Some media:

Playlist of Kart POV

https://www.youtube.com/watch?v=zpoSdUirXHU&list=PLIY2kqPvDRDJVeqkigdMKxDZh4z3fCwVi

Playlist of Gearbox videos

https://www.youtube.com/watch?v=1q5IW1BQX3k&list=PLIY2kqPvDRDIh0ylGUqXeeNi6QQ4EOfW1

Origin Story

The story of the invisible jet kart starts in about 2014, where an old coworker introduced me to Charles Guan's blog and I begun looking into building a racer. I raced with the (then Red shirts) now Big Orange Truck (2018) in 2016, and had a ton of fun. In 2016 I put together a design in Onshape (RIP) of a Spaceballs Winnebago kart, and while it has never gotten built it was an excellent thought experiment that forced me to gain a ton of skill.

The idea with the Spaceballs kart was to have a kart with a short and wide wheelbase. I don't remember what motors I was planning on using, but I believe it was powered with 3-12v FRC spec lead acid batteries and 9×3.5-4 Kenda Tires.

The body was going to be mostly plywood and foam, CNC cut. The wings would only be attached for Moxie rounds, and the model is made closely from actual drawings of the Spaceballs Winnebago, just shortened. I currently don't have any plans to make this vehicle (as much as I would love to) so if someone wants to, please contact me for a CAD model and the documentation I have.

In 2017, I had collected a bunch of PRS-specific components, and was ready to build a kart. I had decided to theme it around the Magic School Bus from the well-known TV series, but the semester started at just the right time to cut that off. However, I had done a bunch of design work on frame, drivetrain, steering, powertrain, etc and had collected a bunch of parts:

  1. Harbor Freight 40V cordless chainsaw motors from a MIT group buy
  2. 1×1 tube from a local metal supplier
  3. 5-cell Nissan Leaf pack from a fellow racer
  4. Tires from a research robot that ended up being solid quality Kenda tires
  5. Steering components from BMI Karts (knuckles, steering wheel)
  6. Brakes from eBay (mountain bike brakes)
  7. A bunch of other steering bits

For 2018, the team consisted of myself, my friends Amanda and Max, and my brother David. We started on the kart right after the FRC and Battlebots seasons wrapped up (RIP Brutus). The only thing that changed for 2018 was the theme: we decided to build Wonder Woman's Invisible Jet plane. I think we're going all in on the airplane theme, expect an airplane themed cart every season from now on! (suggestions always appreciated).

Final Conceptual stage before assembly.

As it sat in the pit at Maker Faire NY

Construction

So for the frame, it's 1×1” 1/16” wall steel box tubing. Super standard stuff, dirt cheap from suppliers and super easy to work with. It's all MIG flux core wire welded with an Eastwood MIG (love that thing), and the angle design is heavily inspired from RUGreasy and FUBAR (2 if I remember correctly).

Initial frame construction:

We originally planned to use this as the frame, however with some highly precise “standing on the frame and bouncing up and down” testing [inset slo-mo], we determined there to be way too much flex in the long 1×1 runs. That necessitated a bit of a fix:

With bars on the top and bottom, this now meant that the frame was solid to take any team member we would be throwing at it.

The steering was a part of the kart that I put a bunch of thought into, but having never designed any sort of steering system, there's only so much one can learn without taking a college course in it. I ended up using the Red Shirts/Big Orange Truck design for the steering, which uses a couple of ball joints, thrust washers, shoulder bolts, and a milled plate to greatly simplify the steering process. Definitely contact Matt Hagan if you'd like the design, it saved me a lot of headache. Here is a video/animation of the steering set up.

We used a BMI bowtie steering wheel, a piece of conduit as the column, running through two Markforged bushings to a BMI double arm Pittman arm. Using some tie rods that I has saved from a dumpster, it made a crude, albeit effective steering setup. See the upgrades section for recent changes.

Powertrain was something I had to do on my own. Charles had done a teardown of the motor but it still hasn't been too widely used. The gearbox you see in the photo is a bench test setup that got kind of…used…on the kart. We weren't sure if two of those motors would cooperate chained together in a gearbox, so we needed to do some testing.. It ended up working fairly well with cheap e-bike controllers, at least at lower speeds. These motors top out at around 6000rpm, and the path right now is 10t sprocket → 16t and then a centrifugal clutch with a 12t → 60t output. It works fairly well, the clutch is there to try to maximize off the line power seeing that sensorless brushless motors have that issue.

The clutch has its internal spring removed to help facilitate it kicking in at a lower RPM. It makes a hell of a racket but works ok.

The rear axle is just a piece of 1” steel round stock, with a keyway cut on each end and the ends drilled and tapped for retaining bolts. It runs on a pair of Dayton 1” bearing blocks, which are bolted to the frame. This setup seems to work fairly well, and I don't see changing it unless it breaks.

The electrical aspect of the powertrain was quite simple. We had to modify the e-bike controllers a bit to the low voltage cutoff of the leaf cells, but that was somewhat simple. For our first event, we wanted to keep the electrical as simple as possible, and it held together quite well through the entire event.

The body work is SUPER COOL and I'd love for Amanda to write about it

Event Reports

New York Maker Faire, 2018!

NYMF '18 being the first race of this kart's history, I wasn't hoping for much, but really was hoping for some reliability. Some initial thoughts:

  • Have a list of stuff that's being brought up, prepare the week before
  • Corner pits are bad, I hate having a $100 welding jacket walk. Get there earlier on Friday
  • Always bring a welder
  • Battlebots are hard to move
  • Better bumpers on the front (not .5” plywood)
  • HF tires suck badly enough that they pop in the second test lap
  • Body work that's sturdy enough to be used to get in or out of the kart is nice
  • No 3d printed parts in critical systems
  • Steering is a bit loose, though it works well
  • Ground clearance: ugh
  • Going 15mph in something you built is both fun and terrifying
  • Worry about all the things breaking, not just the steering, or the frame

The biggest issue we ran in to was that the gearbox was still the bench test prototype: the gearbox plates were 1/8” 6061, and the jack shaft bearing blocks were Markforged plastic.This led to multiple chain issues and ended with a bearing block melting into a bearing.

Otherwise, the components and systems I was most concerned about performed spectacularly. The electronics were rock solid, though the cheap e-bike controller seemed to have trouble getting out of the “1st” electrical gear

The steering also kept together quite well, and after the event when I did a teardown, there was a lot less wear than I originally thought there would be.

Upgrades, changes, and future stuff

September 29-30

Gearbox work. Replaced the Markforged plastic bearing blocks with CNC'd aluminum blocks, and a longer output shaft. Immediate improvement, much less flex in output. With properly spaced output shaft, no longer have issues with bolts coming out.Noticeable difference in testing.

October 6-7

Steering work. Replaced column with solid shaft. Replaced Markforged bushings with oil-infused brass bushings. Took apart and inspected/greased/oiled all the components. New shoulder bolts for the tie rod ends that go into the knuckles, to limit steering slop/play. Additionally, 10” slicks installed. Better ground clearance, though skids a lot in the school hallways. Will have to test on asphalt.Noticeable improvement in steering.

October 13-14

New VESC's came in. Replaced the e-bike controllers, additional noticeable difference. Regenerative braking is a nice plus to have: I like being able to let off the accelerator pedal and get quickly to a desired speed. Blew the fuse after turning regeneration on though, will have to install a diode to protect the fuse from the back current. The acceleration is brutal though. Will need better brakes as well. Neat-ified the wiring. More power poles, but much better run: less chance of catching fire or coming apart. Noticeable improvement in testing.

October 20-21

No work

October 27-28

Replaced the centrifugal clutch with a straight sprocket: wanted to get some testing in on difference between a centrifugal clutch and a direct connection. There's a lot more cogging than we'd like, but this might be due to not having the controller parameters right.

In testing, we also came across another issue: mainly that we forgot to put a ground lead in the CAN cable between controllers, and the back EMF from one managed to blow the CAN chips on both up when trying to manually push it backwards. New controller chips have been ordered, and a ground lead has been installed. This is also a good excuse for me to just make the heatsink I've been planning for the VESC's for 3 weeks now.

November 3-4

No Work

builds/invisible_jet.txt · Last modified: 2018/11/05 09:20 by prysk