Skip to content

Home

Jump to bottom
Nolan Hergert edited this page Aug 8, 2024 · 87 revisions

Slow motion roller coaster with hot wheels tracks

Single rail hot wheels tracks is the cool new thing! Jersey Devil in Six Flags Great Adventure image, image

Make a jump rotation vertical and land upside down vertically using magnets, would be a cool demo. Hard to get consistent maybe, but seems cool. Boost the jump hot_wheels_vertical_jump

Allows even more creativity and theming image

Trying to decide between knex and these straw and connector things. Both are going to be cheap, prefer knex though as its easier to place things at points. Need them longer!

MVP

  • Speed up / slow down section by using 3d printed bump section that you place on the track sides, just above tires and accelerometer detects and interprets. If you really want to push it, something like parameters? Variables: left/right side of track, intensity/smoothness, spacing. But needs to be distinguishable from noise or normal tracsk motion.

Nice to have

  • Design for LEGO characters for riders? Attachments to riders such as hair or feathers or other features that show g forces motion even at scaled speed like Line Rider. Maybe "hanging" like deja vu?
    • Hard to do as real gravity isn't scaled :-/ Feathers might work though.
Software Things
  • Turn off power after motion with no bumps or changes for ... 10 seconds? Can tweak, don't need physical switch I think. Pushbutton turns it back on

auto_power_off

* Need to have diodes to protect GPIO from pushbutton voltage while cpu is in power off state. It can handle external interrupts when ready for it, but not when power off.
  • Can make music by compressing normal STEP toggles to arbitrary frequencies (maybe two freqs at once?) and changing M0/M1 to slow down frequency if needed.
  • Control stepper motor current from microcontroller in order to conserve power. Should scale to gravity direction, but might need to detect skips or something and adjust.
Track bump scripts functions

Program / adjust using RGB LEDs as light sensors, using the phone blasting colors on the screen. Might need a hole / light pipes, plus testing! Otherwise use USB serial when car is docked. Probably fine, 4 pin.

  • Save to memory slot, restore from memory
  • Go set speed until top of hill, then halt for a little? Various things to program

Unsure

  • Direct remote control? Might be fun for "racing cars", but draws some power too. Maybe for another revision.

Weights

  • 2g motor
  • supercap? Need 8v w/ 100ma per coil, although might be able to get a lower voltage driver.

Ideas

  • Detect pick up through capacitive touch sensors on side of vehicle? Or just no motion for a few seconds SLIPpery
  • Predone "tape" roll that you just peel off VHB and apply to hot wheels track
  • Support structure, made out of cheap and cuttable straws? you make supports out of pens/pencils and join with clips? How much are they at dollar tree? Do I really need cuttable? Probably not, it's easy to adjust angle
  • what is another killer app?
  • You can theoretically turn by providing different power to the motor on each side.
  • How cheap are optical flow on bottom of mice?
  • How to flash/communicate with CPU. Use a third/4th pogo pin? Can use multiple to gang up on power too

Can't jump, but you can defy gravity. How to make even more cool slow mo effect?

How to do a curved continuous sideways turn with hot wheels track? Who does that?!

CPU

ESP32 C3 is awesome and cheap! https://www.espressif.com/sites/default/files/documentation/esp32-c3_technical_reference_manual_en.pdf#ledpwm. Limited on pins, but I think I can work around it.

Overall use FastAccelStepper library as it handles PWM libraries well out of the box: https://github.com/gin66/FastAccelStepper

Swivel / ball joint

Need a swivel / ball joint somewhere, since a wide 4-wheel car can't keep up with sharp turns (one wheel is off). Can do a basic magnet on either side + metallic bearing in the middle. Or 3d print something with really wide tolerances.

Realistic model

3D model of cars somewhere...Would like several of them connected eventually.

For prototyping, matchbox and hot wheels are not the way to go actually! Hard to remove rivets, and they don't have flat "insides" of the car I think. https://www.youtube.com/watch?v=UqW1rNJGJYM

Better are these simple dollar tree car with launcher, just screws to remove and flat bottom.

Realistic motion

Use accelerometer, maybe gyroscope and use gravity at 1/16 or whatever. Also put down colored sticky things that the camera or color sensor pick up on to slow down, speed up, etc.

All pins on ESP32 support i2c, great.

Mounting track

Off the shelf pipe? Ideally cuttable with scissors (cardboard), but really just want cheap. Want tightenable ball and socket so it's orientable, strong, and lasts a while hopefully.

  • Would PEX pipe work, or is it too bendy? It's cheaper than PVC I think.
  • What about straws, specifically unbendable or boba straws? It's just not that permanent. It would be if you found a solid mounting mechanism.
  • https://www.thingiverse.com/thing:5161432, or original. https://www.thingiverse.com/thing:2069128
  • Want to modify so it works for PVC pipe. Final parts with 3/4" PVC pipe is way too thick with enough strength. Even with 1/2" (inner dimension) pipe. Final version should use thinner pipe, like TAP plastics clear acrylic or polycarbonate. Needs to be strong though, and not that flexible. Or, attach them to end of rulers, but need to add more stuff for 3d print (no threading allowed).
  • Duplos and legos break too easily with strain at top.
  • Knex works, but seems too flexible and too long to set up? Gives angles for free too.
    • Don't use small k'nex, more easy to kick and break.

Joints

Are triangles really the strongest shape? Great video... image

  • The joints needed for a triangle are just pin / revolute joints!
    • However if you allow swivel in 3D, then you need a pyramid. I don't think anyone wants that. Can make pseudo swivel by allowing magnetic coupling on straw and that holds fast at whatever angle you place....hopefully.
    • My internal jury is still out here...
  • Too many degrees of freedom with ball and socket. Going with pin joint for arm and pivot joint for base. Want them to be sorta hard to move though so it's easier to place. Have a tightening system (nut?) printed in place as the pin?? Might not help that much in the long run, we'll see.
    • Other option is to have the user add nail polish or superglue to increase friction, like scale model figurines do. LEGO works pretty well though.
    • Pretty sure pin joint for arm and pivot for base is the same in the end as ball joint. But maybe not, worth a shot?
    • Also could investigate a fixed angle, but then you need to trim the straws a lot more often and have wasted straws. Allow people to determine optimal angles more naturally. Needs to go down more than the 30 degrees that it is currently. Might be best with a fixed angle in the end though.

Base

Needs to be easily movable by human, yet difficult for weight pushing down or at off angles to move it. Can be bumped and won't fall over or wear out over time. Easy to move/remove truss without getting underneath it. Packable into small size. Hooks together like Dollar Tree pegboard design. Ideally want something one can buy from dollar tree or equivalent.

  • CAD a design that is either thin (with wide flange on adapter) or tall
    • Adapter for straw should be either tall or flange out wide (probably both) so it has less deflection.
  • Magnet with grippy overmold / sticker (or kitchen magnet is similar) is probably best. Suction cups would work too.
Not as great

  • Dollar tree peg board might be good, but is thin-ish plastic and narrow hole and doesn't fit in itself (so takes up lots of space in box)

  • Thread board (https://www.kickstarter.com/projects/3dprinteracademy/thread-boards-3d-printable-peg-boards-with-a-twist). Don't want to have to twist to remove, right?

  • A type of "Tooth" adapter might work on normal pegboard, but needs some redoing. image

  • Suction cups on window? Problem is, can't very easily drive car with gravity in a different direction, due to gyro drift. Could do it, but need magnetometer, etc, slightly more expensive part.

Can be tipped
  • Square patch of drawer liner? Works on clear floor, not sure about carpet.
  • Spiky carpet support things, but doesn't work on smooth floor.

Camera streaming

Probably want to have people order their own module since doing wifi is annoying and takes up too much space. https://shop.runcam.com/runcam-5-orange/ is pretty sweet, uses QR code and camera to set exposure. No wifi, but tested for FCC. Nice.

  • SQ11 is a cheaper option. $5 or something and 18g.

OV2640 is cheap $(2-3) and plentiful. OV5640 has i2c support for shutter (ov2640 uses dedicated FREX pin but isn't exposed, workaround: https://github.com/ArduCAM/Arduino/issues/108#issuecomment-267542587, more: https://www.esp32.com/viewtopic.php?t=11126), which is useful for iPatch project. But thankfully they are pin-compatible and ESP32 cam supports both and more! https://github.com/espressif/esp32-camera#supported-sensor

There are some good ESP32 projects, and the latency is great! (<100ms). I've tested with an action cam, and it's more like 200ms.

https://www.reddit.com/r/esp32/comments/mz6bnx/esp32_camera_fpv/. References this apparently good talk: https://media.ccc.de/v/36c3-10630-wifibroadcast. Project (https://befinitiv.wordpress.com/wifibroadcast-analog-like-transmission-of-live-video-data/); https://github.com/jeanlemotan/esp32-cam-fpv

Antenna

Can use a monopole antenna since I have the required wavelength/4 of 3.125cm vertically for the seat!!

  • However in ideal state it doesn't transmit vertically. :-(
  • Still want matching network just in case though

https://www.researchgate.net/publication/229726376_Compact_paper-clip-shaped_wire_antenna_for_24_and_52_GHz_WLAN_operation

PCB antenna designs: https://github.com/sad-electronics/wch-kicad-lbr/tree/main/footprints/wch-antenna.pretty image

Power

https://docs.google.com/spreadsheets/d/19r14_vL9nFiVYM3p7StDWISgVfOAb2pPXsbqhz31ikA/edit?usp=sharing

https://docs.google.com/spreadsheets/d/19r14_vL9nFiVYM3p7StDWISgVfOAb2pPXsbqhz31ikA/edit#gid=0

Charging

Pogo pins can take 3A, supposedly. That should be good enough assuming the super caps are in series.

Need to make sure voltages on supercaps are balanced. Easiest (I think) to ensure that no supercap exceeds 2.7v, which you can use a voltage reference for and some NPN/PNP. https://www.sensorsiot.org/simple-and-cheap-way-to-protect-your-super-caps-video-139/. EEVBlog video uses another one, but cheap basic part on JLCPCB is CJ431. $.03, fine, done. Make sure it works the same though!

NFET and switcher solution

Kinda want to do a bi-directional synchronous converter, as I need to buck down from power supply but boost for the motor driver (to overdrive it). But it can be in opposite directions and I think I can share the same parts. However commercial synchronous chips don't generally support bi-directional out of the box apparently. Getting decently far (need a half bridge driver for the fets, since I can only switch an NFET from a uC at 8mA in less than 10Khz), but need a solution for safe low quiescent current and shutoff (driver is 200 uA).

Others

From this video: https://youtu.be/trEnwNaEgJ0?t=185. Explained well here too: https://youtu.be/trEnwNaEgJ0?t=1824

The point is the shunt resistor/nfet needs to be appropriately sized so that the capacitor can't continue to charge (pulls all the charging current away from capacitor).

Maybe can not blow out resistors by dynamically lowering the charging current. But requires measurement and communication and monitoring and timering and stuff.

Why can't nfet physically remove connection to capacitor? Cut off flow instead of shunt it? Need to open it again...how? :D

Voltage reference/comparator: https://jlcpcb.com/partdetail/HX_hengjiaxing-HX61CN2702MR/C296257 is a compatible part, whereas the datasheet could be XC61C: https://product.torexsemi.com/system/files/series/xc61c.pdf

Boost converter

TD8228 or ME2188 (on JLCPCB), maybe MT3608L but just try switching regulator from USB power banks things first. See how long it lasts! Reduce voltage too to 3.3 using resistor, I think.

I'm intrigued enough to go further with these devices. But just as I was searching for the datasheet link for the ME2108, my eye was caught by the ME2188 and this seems to be able to drop the schottky diode and switch at a higher frequency (smaller inductor), for a small tradeoff in max current. I now want to try this! Here's the mainly chinese datasheet for those playing along at home!. From https://hackaday.io/project/164674-coin-cell-challenge/log/164162-joules-to-photons-finally

Motor

I am liking a cheap stepper motor, potentially removing gears and microstepping it when needed, but still able to speed it up to fast speed (no gear reduction). Don't need a lot of torque, and would like it cheap so lots of people can implement it.

Not sure how powerful the cheap and small ones will be. Just gotta try it out.

Need braking at low/zero speed, so it rules most direct drive dc motors out unless they are geared and have position feedback.

Do I need gears to increase holding torque? Want it cheap and easy to buy

  • Worm gear is one option
  • Increase torque with big gear.

Stepper Driver

  • A4988 is cheap but not spread spectrum
  • DRV8834 doesn't go to voltages above 10V. Might be ok (3X LIC), but can go higher too.
  • TMC2208 is classic great driver, $2. Lots of data and control over UART (can skip DIR/STEP for example and just provide velocity updates). Saves some wires.
  • DRV8428 is a TI option too. $1.25
  • Both of the above need tuning to get really silent.

Might be nice to have would be the ability to start the car by dragging it and detecting speed using back-emf pulses. Might need to do it on the microcontroller, since it's not a typical thing for a stepper driver to measure and report back on.

Anki Drive

https://www.microcontrollertips.com/teardown-inside-anki-overdrive-racecar-set/

  • They are opting for brushed dc motors (cheaper I think) and hall effect gear sensors. Interesting.

Tires

Need traction, and have it last in the face of long term dust. The regular plastic tires are not sufficient.

Aquarium tubing (glossy) worked well. Heat it up so it's flexible, then work over the tire. There is glossy heat shrink too, so that might be easier.

Try rubber cement, or something like draw grip but less thick and no holes.

  • Increase traction using magnets, especially when upside down.
  • Heat shrink on tires? Might work, but could use improvement
  • Tacky grip tape (like for tennis rackets) on track? Might lose tackiness over time though, get dusty. Would need black version.

Clone this wiki locally