Shop OBEX P1 Docs P2 Docs Learn Events
Measuring Skateboard Tilt - Page 2 — Parallax Forums

Measuring Skateboard Tilt

2

Comments

  • It's been a while since I've worked on cars.
    My wife has a '71 WV Super Beetle which she still uses just about daily. That's not too bad to work on but modern cars are a bit of a mystery. Gone are the days when I would feel comfortable changing our a starter or water pump.

    I might feel more comfortable working on a modern electric vehicle. If you guys haven't watched any of WeberAuto's videos, you're missing out. One of my favorite is when he took apart a Bolt's motor.

    My favorite part of this hour long video is at 39:19 when he pulls the core out of the motor. This guy is in a wheelchair.



    Some of these modern EVs seem more electric robot than car.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2020-08-18 17:59
    JRoark wrote: »
    EDIT2: You could make it so the "shim" could be quickly removed.

    Thanks for taking time to both sketch out your idea and upload the sketch to the forum.

    I just purchased a board with 35 degrees built into the nose and tail. I had to purchase several (OK, I had to buy ten) to make up the minimum order from Alibaba. If anyone wants to buy a Russian Maple mountain board let me know.

    Ha26f2f976d4f4280ac57d1437c2095efm.png

    One of reasons I decided not to use my existing board is the big hole cut through the board to allow the original trucks the board used.

    WedgeWithBoard2.jpg

    Here's a side view showing the inadequacies of my original wedge attempt.

    WedgeWithBoard1.jpg

    I don't care about cutting an elongated hole in the wedge but neither the board nor the trucks have enough material around the hole to allow the holes to be elongated.

    I'm not all that concerned about damaging the original deck. My main concern is the damage a broken deck would do to me if it broke while I was riding it.

    I'll likely put this project on hold until I receive the new deck. If I can figure out how to design a part in CAD to 3D print, I might get that a go as well. Right now my CAD design skills are very limited.

    What do you think of the strain gauges I linked to? Do you think those would work for this sort of application?
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2020-08-18 19:16
    Raw strain gauges are tricky to apply (in both senses of the word). If you could find a ready-made load cell that fit the desired space, you'd be miles ahead.

    -Phil
  • Raw strain gauges are tricky to apply

    I watched several videos on this topic and I agree. I'm hoping the application isn't as critical if one is willing to calibrate the gauge after being applied.

    Still, if anyone has a suggesting for an appropriate load cell, I'd like to hear about it. None of the load cells I've seen seem appropriate for this application.

    I'm pretty clueless about how I'll apply the strain gauges but at least they're inexpensive so it shouldn't hurt to experiment a bit. I'll also have extra decks I can use in these experiments.

  • Been thinking more about the tilt sensor. This might be the perfect app for a differential capacitor:

    differential_cap.png

    The individual plates (I and Q) are excited with high-frequency signals that are 90° out of phase. They are picked up and summed by the common plate (Y) in proportion to the amount of coverage each of the I and Q plates has with the common plate. As the board tilts, the the I/Q coverage ratio changes in proportion to the angle, so the relative signal strengths comprising the sum also change.

    The Y signal is fed into a sigma-delta ADC on a P1, where the I and Q signals are separated, as shown in my thread about receiving radio signals:

    https://forums.parallax.com/discussion/105674/x/p1

    The capacitor plates can be made from copper-clad FR4, with the Y plate being covered by clear packaging tape to prevent shorts.

    I know this is a very brief introduction to the concept. Given time, I might be able to experiment with the idea (sans skateboard, of course).

    -Phil

    800 x 494 - 113K
  • This might be the perfect app for a differential capacitor:

    Very cool!

    After seeing Chip's touchpad demo today, I started wondering if I could have sheets of metal embedded in the skateboard. If the metal pieces were in the right places, it might be possible to figure out the rider's stance. It would be pretty cool to be able to use one's feet to provide input. I'm pretty sure there electric skateboards advertised as "AI" which uses the rider's foot pressure for input. Here's one.

    I was thinking I could experiment with some of the extra decks I ended up purchasing since I couldn't purchase fewer than ten.
  • Duane DegnDuane Degn Posts: 10,588
    edited 2020-08-26 19:26
    Here's a photo of one of the VESC boards I'm using.

    Vesc200826a.jpg

    A Parallax LaserPING can be seen on the right.

    I have some additional cheap sensors from China.

    Sensor200826a.jpg

    I may use four of the cheap sensors.

    In an earlier post I mentioned I was willing to purchase 10 decks through Alibaba in order to get a deck with the correct angle on the nose and tail. I received an email stating they didn't have these decks so I'm back to trying to figure out how make an adapter for my old board.

    I'm currently printing the bottom portion of the adapter.

    BottomPlates200826a.PNG

    BottomPlates200826b.PNG

    The above part takes about 6 hours to print at 5% infill and a single shell thickness. Hopefully this will let me know if it will fit properly with my board and trucks.

    There are a couple 3D printable adapters on Thingaverse but I wanted access to use a socket wrench on most of the bolts. I also wanted to strengthen the cutout portion of the board. Besides the bottom adapter, I'm also making a top plate to sandwich the portion of the deck with a big hole in it. The top plate took three attempts to get the size right.

    TopPlate200826c.jpg
    TopPlate200826a.jpg
    TopPlate200826b.jpg

    It was very satisfying when the part finally snapped into place.

    Once I get the size of these parts corret, I'll print the parts in PETG with a high infill percentage. I'm hoping the parts will be strong enough for my timid riding style. I'm guessing they will work fine.

    It is a lot of fun to have a 3D printer. Being able to produce functional parts relatively easily has been great.
    1363 x 1022 - 555K
    1165 x 647 - 346K
    1252 x 1261 - 765K
    1404 x 1536 - 1M
    1363 x 1022 - 368K
    698 x 1002 - 127K
    1134 x 953 - 240K
  • jmgjmg Posts: 15,140
    Duane Degn wrote: »
    Supposedly this is an off road skateboard. I'm not crazy enough to try riding this off road, but someone might want to try riding the board off road. It would be nice if the tilt sensor weren't vulnerable to rocks and logs.
    Everything will be vulnerable to some extent, but mounting a sensor away from the leading edges would certainly help.

    You can also use magnets and simple inductors as sensors, as the magnetic field moves the BH curve, and the inductance changes. It's not linear, but it is monotonic.
    If you make a LR oscillator, you can measure the frequency to high precision.

    The differential capacitor Phil suggested above is also low cost, and it has an advantage it has a natural mid-zero
    - you can change the PCB planes in this, to vary the response shape.

    You could also look at a coupled vane differential capacitor design, so that no wires need to exit the base plate. ie the rotating vane has 3 overlap-pad zones, the mid larger & stable zone is the signal-take-off and the partial outer ones are the sensing zones, as in Phil's sketch. The PCB material can be used as the insulator to prevent shorts.



  • W9GFOW9GFO Posts: 4,010
    Another option for 3d printing large items that must be durable is to print them with a gyroid infill. Afterwards drill some holes through the shell and fill with epoxy. The gyroid fill allows the epoxy to completely fill the interior.
  • Related fun in our office a couple of years ago:

    Ken Gracey
  • jmg wrote: »
    You can also use magnets and simple inductors as sensor

    That's another interesting idea. I have a feeling I'll be trying out all sorts of different sensor ideas with this board.
    W9GFO wrote: »
    The gyroid fill allows the epoxy to completely fill the interior.

    Thanks for the suggestion. I don't think I have any epoxy in the right consistency for this sort of application. I'm thinking PETG with a high percentage infill will likely be strong enough for my needs.
    Ken Gracey wrote: »
    Related fun in our office a couple of years ago:

    "Don't ride the skateboard and let someone else have the hand held controller." Words I intend to live by. That and "What would JonnyMac do."

    Are those Arlo wheels? The wheels which came with the motors and trucks I'm using are relatively nice pneumatic wheels. I shouldn't need to cannibalize the set of Arlo wheels you sent me for this skateboard.

    As I mentioned elsewhere in this thread, I'm planning on ultimately making a set of Mecanum wheels for this thing. One of the many things I still need to figure out is a way to add a third axis with my Wii Nunchuck controller.

  • The part I'm calling "Bottom Plates" finished printing and it mostly works as hoped. I forgot to make the modifications I had made to the top plate so the part to strengthen the hole had to be reshaped a bit. The rest of the design appears to work as expected.

    BottomPlates200826b.jpg
    BottomPlates200826a.jpgBottomPlates200826c.jpg

    This extremely light version of the print cracked as I tightened bolts. I expect a PETG with high percentage infill to be much stronger.

    There are plenty of things I see which could be improved but my current skill with CAD had convinced me not to pursue the improvements which could be made. I think it will work well enough as it is.

    The truck must be bolted to the adapter before the adapter is bolted to the deck. The long round holes allow my socket wrench to fit down inside the adapter just as I had hoped. The trickiest bolt to secure is the one which requires a nut to be held in place through one of the side slots.

    A 3D printer is pretty amazing. I have an idea for a part and I just need to model it in CAD and have it printed out. I feel kind of silly for waiting so long to get a 3D printer.
    842 x 791 - 144K
    936 x 829 - 191K
    984 x 982 - 190K
  • I knew I wanted to be able to run wires from the top of the board down to the trucks so I crudely added some cutouts for wires.

    Here's a view from the slicer software showing the path of the bottom channels.
    BottomPlates200827a.PNG
    The wires will be routed through the openings used by the socket wrench. Here's a few screen grabs from the CAC software to give you all a better idea how this wedge is designed.
    BottomPlates200827b.PNG
    The view below is inline with the bolt holes used to mount the wedge to the trucks.
    BottomPlates200827c.PNG
    The wedge has to be bolted to the trucks before being bolted to the deck.
    Here's a view inline with the bolt holes used to attach the wedge/truck assembly to the deck.
    BottomPlates200827d.PNG
    Sensor wires will be routed down a single hole which is a continuation of a hole through the top plate (not shown in earlier top plate prints). This single hole splits into two channels which eventually route the wires down near the trucks. Here's my attempt to sketch the wire path.
    BottomPlates200827f.PNG
    I'm currently printing this wedge (it should take about 24 hours) in black PETG. I'm using four perimeters and a 50% gyroid infill. I chose gyroid infill just in case I wanted to use W9GFO's idea of adding epoxy resin to the empty area. After seeing how dense the gyroid pattern is at 50% infill, I don't think the epoxy resin idea would work with these slicer settings. I think the part will be plenty strong without adding epoxy.
    I added the yellow highlighter lines (only one line survived for some reason) to the last image to show where a long hole for socket wrench assess would end up. The access hole for the top nut, would intersect the bolt hole used to attach the wedge to the truck. Adding nuts through the side cutouts is awkward compared to having a long hole for a socket wrench but I didn't see a good alternative to using these side cutouts.
    My knowledge of how to use Fusion 360 just barely allowed me to complete this model. There are plenty of things which could be improved but I'm at the limit of what I can presently do with CAD.
    I have a feeling I'm going to need to improve my CAD skills if I hope to design Mecanum wheels for this contraption.
    2141 x 1716 - 2M
    1248 x 953 - 552K
    696 x 903 - 351K
    962 x 1115 - 461K
    1000 x 855 - 288K
  • Duane DegnDuane Degn Posts: 10,588
    edited 2020-08-28 03:41
    Have I mentioned owning a 3D printer is really fun?

    I just went down to check on the print at it looked a lot like the render I used to show the wire channels.
    BottomPlatesPartial200827a.jpg
    Here's the render I posted earlier.
    BottomPlates200827a.PNG

    This is likely a common sight to those of you who have a 3D printer. The printer is still new enough to me for this sight to make me unreasonably excited. It still amazes me something which had just recently been an image on the computer screen is now in the process of becoming a physical object.

    Learning to use a 3D printer reminds me of first learning to use a microcontroller. I had always thought being able to program a microcontroller would be fun. The enjoyment I gained from learning to program microcontrollers was much more than my high expectations had anticipated. Likewise, I always thought it would be fun to be able to 3D print robot (and other) parts. As with microcontrollers, my high expectations have been exceeded.
    1317 x 1025 - 440K
  • I agree. I'd like to have one here, but there's no room for one. I saw one a while ago at the Micro Center in Brooklyn NY a while ago, and it was trying to convince me to take it home. (He's still waiting.)

    One of my sisters (Older) saw the same one later that month, she was, ah, moved.
  • W9GFOW9GFO Posts: 4,010
    I agree that with such a high infill percentage that flooding with epoxy would not be practical.

    It is going to be a fairly highly loaded part, I think I would use double the number of shells and half the infill percentage. The number of shells (walls) being the most important.
  • W9GFOW9GFO Posts: 4,010
    Duane Degn wrote: »
    Have I mentioned owning a 3D printer is really fun?

    I just went down to check on the print at it looked a lot like the render I used to show the wire channels.

    Always a good sign. ;-)

    As far as your CAD skills, they seem good to me.

  • W9GFO wrote: »
    I would use double the number of shells and half the infill percentage.

    I was debating this myself. Four perimeters is double the usual number. I take you're suggesting eight perimeters? I'll likely do this on the second wedge I make.

    I discovered the front and back holes are not in the same location relative to the bolt holes. I'll need to shift the part which fits inside the cutout by about 2.5mm. The front and back wedges will not be identical.
    W9GFO wrote: »
    The number of shells (walls) being the most important.
    Yes, I want CNC Kitchen. I just rewatched his video about infill vs perimeters and I must have subconsciously remembered his "4 perimeters/50% infill" suggestion (found at 8:49 in the previously linked video).

    I rewatched the video after starting this reply. Now I'm not sure if I should use more than 4 perimeters or not. As you suggest, I think this is one application which would benefit from more than 4 perimeters.

    If any of you have a 3D printer and aren't subscribed to CNC Kitchen, I suggest you do so. I think CNC Kitchen is one of the best 3D printing YouTube channels.
  • W9GFOW9GFO Posts: 4,010
    Yes, I am suggesting eight perimeters. I quickly skimmed the video but did not see where he compared additional perimeters vs less infill. I don't disagree with what I saw but I don't think his idea of strong enough and my idea of strong enough are the same - especially considering this is something that a person will be putting their weight (and other stresses) on while in motion.

    It is more than just the shell strength. There is also the material underneath the bolt heads. You want that to be completely solid, also it is good to have the shells around the holes meet the sides of the part for best strength.
  • W9GFO wrote: »
    There is also the material underneath the bolt heads. You want that to be completely solid

    I think you're correct about this. I don't know how to selectively change the infill. I know it's possible, I'll need to learn to do this if I want use your suggestion.
    W9GFO wrote: »
    also it is good to have the shells around the holes meet the sides of the part for best strength.

    I'm not so sure about this. I think there are advantages to not having a solid mass. I think it's related to the way tubes are stronger than rods.

    Hopefully, I'll be printing the wedge for the nose of the skateboard today. I'll likely include some of your suggestions. Thanks.
  • Go with the load cell sensor, something that can get dirty, wet, etc. without affecting the operation
  • DigitalBob wrote: »
    Go with the load cell sensor, something that can get dirty, wet, etc. without affecting the operation

    Any suggestion on which sort of load cell I should use? I ordered some of the cheap strain gauges from Aliexpress. They were so inexpensive I didn't mind trying them even if they don't turn out to be useful. Load cells tend to cost a bit more. I want to make sure I'm ordering the right kind for this project.

    As can be seen from the CAD drawings, the thing I'm calling a "wedge" is pretty big. There's likely room to incorporate some sort of load cell if I knew what kind to use.
  • They have a load cell kit on Amazon, 4 ea. cells and control $9.00. Called 4pcs. 50KG half strain gauge load cell body scale.
  • If you want to go with the pivot sensor you could use a car TPS sensor (Throttle position). These have a magnetic inside so no parts to wear out like a pot. Or use an accelerometer on the deck and another on the trucks. Then figure the difference from normal attituded.
  • DigitalBob wrote: »
    These have a magnetic inside so no parts to wear out like a pot. Or use an accelerometer on the deck and another on the trucks. Then figure the difference from normal attituded.

    I'll likely try a couple of different methods of determining tilt. I have a bunch of magnetic encoders left over from an earlier project. Even if these don't wear out like pots, they're still a bit of a challenge to mount.

    There's some debate on how many IMU sensors would be needed to determine tilt. I'm not sure if the tilt value can be determine with just one IMU but I'm not convinced the data from a second IMU would be very useful. I have four BNO055 sensors on the way. If I decide to use more than one for this project, I'll be ready.

    I haven't given up on using the little time of flight range finders. I figure I'll have fun trying out a variety of sensors.
  • I'm going to flip back to load cells. Like a measurement specialties FX29KO series, you could mount one under each spring seat. Program the spring tare. About $25 each but not an issue for a high board.
  • High end board
  • W9GFOW9GFO Posts: 4,010
    Duane Degn wrote: »
    W9GFO wrote: »
    There is also the material underneath the bolt heads. You want that to be completely solid

    I think you're correct about this. I don't know how to selectively change the infill. I know it's possible, I'll need to learn to do this if I want use your suggestion.
    W9GFO wrote: »
    also it is good to have the shells around the holes meet the sides of the part for best strength.

    I'm not so sure about this. I think there are advantages to not having a solid mass. I think it's related to the way tubes are stronger than rods.

    Hopefully, I'll be printing the wedge for the nose of the skateboard today. I'll likely include some of your suggestions. Thanks.

    Tubes are not stronger than solid rods, but they are much stronger for their weight and usually a better choice (rebar being a good example where they are not a better choice). The very center portion of a solid rod would not be a benefit to any bending forces but it would add shear strength. In the case of this 3d printed object, weight is of no concern and material usage is of no concern so I would err on the side of thicker walls rather than thinner. 100% solid would be the strongest but that is overkill. IMO, of course.

    To selectively make the walls thicker on a hole you can extrude-cut a thin annular ring around the hole nearly all the way through the part. As if you were using a circular cookie cutter to remove the area surrounding the hole. When printed, both sides of this groove will be printed with walls effectively tripling the number of walls around it.

    In the example below the width of the annular ring is 0.1mm and the number of walls is set to three. If I had made the annular cut a little further from the hole there would have been room for it to put another two walls.


    Screen%20Shot%202020-08-29%20at%209.26.12%20PM.png

    Screen%20Shot%202020-08-29%20at%209.28.07%20PM.png

    Screen%20Shot%202020-08-29%20at%209.27.38%20PM.png
  • W9GFO wrote: »
    To selectively make the walls thicker on a hole you can extrude-cut a thin annular ring around the hole nearly all the way through the part.

    That's a great trick. I even think that's within my 3D modeling skill set.

  • DigitalBob wrote: »
    High end board

    This certainly isn't a high end board but it's also not a going to be a traditional electric skateboard.

    As I've mentioned elsewhere in this thread, it's my intention of making this board 4WD with Mecanum wheels. Of course this is easier said than done and I'll be taking this a step at a time. The first step is an attempt to get the skateboard to have an electronic differential with 2WD using "normal" wheels. The normal wheels in this case include 8" (200mm) pneumatic tires.

Sign In or Register to comment.