Weight Sensing Longboard with Inline Wheel Motors

[justin_le]

Bit of Background:

This project is about 9 years coming to fruition. Back at the early days of the UBC electric bike club we go ahold of an Exkate with badly sulfated lead acid batteries, so took it as a project to upgrade with a state-of-the-art (in its day) 'F' cell NiMH pack. After some bruises and a short learning curve it soon became apparent that powered skateboards are about the most thrilling way to get to class and back:


However, I liked to use it to get groceries and hold the grocery bags in my hands while riding home, but the hand-held wireless throttle on the ExKate made that a bit cumbersome. So one of the ideas we wanted to try was embedding strain sensors inside the board to create a hands-free control interface, where you can shift weight forwards and backwards to control the board's acceleration and braking. This idea was high on my mind when at the 2006 Vancouver Electric Vehicle Show show a guy Frank Schmidt from Bellingham showed up with a homemade 3-wheeled powerboard running on this exact same principle! You can't see it in the photos, but the deck itself is connected to the frame through a pressure sensitive material whose electrical resistance changes with the amount it is squeezed.

http://3wdm.blogspot.ca/2006/01/faq.html

So we immediately schemed on working together to build a more refined version of this concept. At the time I had some geared ebike motors from "ethinkar", which when disassembled contained an outrunner that was just the right size to direct couple to the mountain board wheels.

The mountain board trucks were machined to fit the hub motor bearings:
View attachment 1

Here is an exploded view showing how all the motor drive parts stack together. There are 5 bolts that go right through the motor bell and the plastic hub to couple the rotational force between motor and wheel


The way these trucks are made from folded plate metal made it fairly easy to attache strain gauge sensors:


We decided to use industrial servo motors controllers for driving the motors, (from http://www.elmomc.com/products/piccolo-main.htm) since we need smooth bidirectional torque control and the ebike controllers of the day didn't offer this. A large circuitboard was built to tamplify the strain gauge signals, run an LCD module to display what's going on, and control the servo controllers based on the weight distribution:


And it all seemed to come together pretty well



We got the software just finished in time for the 2007 Maker Faire. It definitely needed some dialing in but on the whole seemed to have hit the mark and looked set to be a rugged platform for testing and refining the idea.
http://zedomax.com/blog/tech/zedomax-maker-faire-2007-coverage-electric-skateboard/

Unfortuantely though, on our way back from Maker Faire the skateboard got stolen from a warehouse that we were crashing at in Portland. That was a bit of a bummer, and derailed the plans somewhat for quite a few years.

 

[liveforphysics]

I'm excited for the next update on this project my friend!

 

[justin_le]

I'm excited for the next update on this project my friend!

OK, well the next update didn't happen until 6 years later, when we finally decided to retackle this project. BUT, this time around we wanted to really get it right. The 2007 version was built around an off-road mountain board which is appropriate for a motorized skate, but it's not very stealth and attracts a bit of attention on the road. What massively appealed to me was making a slick build that looked like any non-electric longboard. What also appealed is making something that could conceivably be a retrofit 'conversion kit' for longboards just like we do for bicycles, and that similarly embraces the idea of using standard longboard hardware throughout.

Almost all the powerboards (both commercial and homebrew) that people make these days are based on the RC sized motors bolted on a flange on the truck, with a chain or belt drive to the wheel. It works great but the transmission makes audible noise and it also adds noticeable drag when you want to just kick the the board without power. The direct drive approach requires a lot more custom motor machining, and requires larger size motors for the same power output because there are no benefits of gear reduction, but it would run smooth and silent and result in less drag when always enabled for regen. Pretty much the same pre vs con arguments as geared vs DD hub motors in ebikes.



Although it is tempting to try and cast a wheel right around the motor, skateboard wheels themselves wear down pretty fast in use and that would result in a short lived motor. It would also mean that all the harsh impact and vibrations would directly hit the motor ball bearings, which would need frequent replacement. So we wanted to make it in a way that skateboard wheel is easily replaceable and still rests on its own ball bearings, and that only the spinning torque is coupled to the motor, not the lateral impacts.

Here's a section view of the motor concept and an exploded view showing how all the parts would fit together on a truck. The truck could be any longboard truck but it needs to be machined on a lathe to have a concentric shaft for the motor itself to slide onto, leaving the normal 8mm steel axle protruding for the skateboard wheel:

 

[EBJ]

Awesome Justin.
Would love to see some more of this thing.
Perhaps you will have it at the Bay Area Maker Faire?

 

[justin_le]

Would love to see some more of this thing.
Perhaps you will have it at the Bay Area Maker Faire?

Perhaps! That's all assuming I can get it up and running by this weekend.

The machining involves a lot of tight fitting parts, both interference fits for some things and snug fits with no play for the bearings. The concept design supports the motor rotor from two large diameter ball bearings on the opposite side as the wheel, and the stock RC motor shell did not have enough material on the end for dual bearings. So the first step was to machine an insert from 2.5" round stock to press into one end of the rotor and give an extra protrusion:



The finished machined part is on the left, and it should be a tight interference fit to the rotor on the right:


With loctite and a serious bench vise to get the two pieces together, they hopefully should never see fit to com apart again:


Similarly, the stator requires some tight fitting parts too. After disassmebling the RC motor we found that the iron lamination stack had a 20mm ID, while the skateboard truck could easily be machined to 16mm diameter without much loss in strength, allowing for a 2mm wall thickness of the stator suport.
View attachment 3

Here's what the finished stator support came out like, with the CAD drawing on top:


Now the RC motor we got was delta wound, while for the inline wheel we need to slow the RPM down so the plan was to convert this to a 'Y' winding. The connections were undone and separated into 6 strands, and these fit through the 6 holes in the stator support:



And, once again some loctite and a bench vise enables us to press the the stator and stator support into a single unit:

 

[justin_le]

With both the rotor and the stator finished, in theory the two halve should just press together and produce a an open ended motor with a 16mm hollow bore that can slide over a skateboard truck.


And presto! I was a bit nervous about how well the dual ballbearings on one end would be able to support the rotor concentric to the stator over such a span, but the airgap seemed pretty uniform around:


And compare this to the CAD diagram in the previous post:


What required figuring out next was just how the skateboard wheel would couple to the rotor. Frank had the clever idea of making a coupling plate that has splines on it that would fit between the rotor magnets. So this was turned with a rotary table on the milling machine and made so that each protrusion gits snugly between the magnet gap.


Then, to link the motor to the wheel we put 6 bolts through the plate that protrude on the wheel side, and then machined a cavity in the longboard wheel with a set of holes in it to match each of these screws:

 

[justin_le]

Everything done in the previous steps had to be repeated all over again to make a 2nd matching motor for the other side of the truck. With a belt or chain linking the RC motor to the wheel with a gear reduction, just a single drive would have been sufficient power. But in the direct drive configuration the reduced torque output had us concerned that a single motor may not be sufficient.

So just the other weekend Frank came over to machine a new set of stator and rotor supports:


while I worked on machining a new coupling plate to link the 2nd motor to the 2nd wheel


Everything on this one seemed to go together fine, except that one of the press fits was a little too tight,:

and that ended up pinching one one of the phase leads which resulted in a short to the winding. That had us kind of bummed out, since there was no easy way to unpress the stator off the stator support and fix the wire. However, when we separated out all the strands, it turned out that it was only a single strand of a single phase that had the short, and so we left this out of the bundle and then the remaining wires remained isolated. Loosing one strand meant just 8/9th of the copper fill in that one phase. I wasn't sure if we should similarly remove a strand from each of the other windings to get a balanced load, or leave them there to reduce the copper losses. You can see the stray shorted wire here:


One thing that wasn't apparent to me before is that when you convert from a Delta to a Wye winding, the polarity of the phases which connect a neutral is really important. If you get it wrong, it spins super rough (if at all) with a controller and the back-emf shows the 3 sine waves being just 60 degrees out of phase.


This meant was that the middle phase of the waveform (blue in this case) was connected backwards, and needed to have the opposite end connected to neutral. When the motor is spun by hand the back-emf waveforms then look like so:


And runs nice and smooth with a sensorless controller.

 

[justin_le]

I ended up switching to a different and more conventional truck to facilitate the attachment of strain sensors. This is a 180mm Randall truck that had the axles turned down to 16mm diameter for the motors to slide on, and is shown having a flat surface machined on the bottom for the bottom strain gauges.


Installing the strain gauges was easier than I was expecting, using scotch tape as a carrier to maneuver them and hold in place while the then film of epoxy adhesive cured.


For the best sensitivity with strain gauges, you want to build a full bridge with two gauges in compression and two in tension. So I also machined a small flat on the under side of the trick and affixed two strain sensors there as well, and after soldering all the lead wires the gauges were covered in a thick film of RTV silicone to protect against exposure:
View attachment 1

After that, we just slide the motors on, slide the wheels on aligning the holes with the screws on the coupling plate, then secure everything in place with the axle nut, and Voila!


So that's where things are at as of this evening. Next up is wiring up and configuring some 2-quadrant motor controllers, soldering some instrumentation amplifier circuits for the strain gauges, and then writing a new set of Cycle Analyst firmware designed to take as two inputs the weight on the front and rear trucks and then drive the motors with either forwards or regen torque depending on how that weight is distributed. Lean or shift your weight forwards to accellerate, and backwards for regen.

 

[flathill]

Bad ass

Make it fwd so you can still ollie
Or ride in reverse and drift

Is the torque from to motor only coupled through the magnets or is the coupler also a press fit
Seems the magnets may crack or come loose
Neo mags are only strong in compression

Scorpion motors? If so nice choice

 

[sk8norcal]

wow, that's pretty crazy.

I am not a believer yet in weight sensing e-boards, seems like the holy grail for sure.
Had a very brief try on the Zboard.
there is also Goboard, and a couple of other attemps here,
http://wackyboards.blogspot.com/2010/06/z-board.html

seen this other two eboards with hub motors?
http://wackyboards.blogspot.com/2011/03/hardbord.html

 

[Chalo]

That's some pretty fancy work there, Justin.

I gotta say, this is the first e-skateboard related post here that didn't only make me want to go watch some more skateboard fail videos on Youtube. Weight-shift based control of a powered board is also the first thing I've heard of that might make skateboarding a slightly less terrible idea.

 

[justin_le]

Make it fwd so you can still ollie
Or ride in reverse and drift

This particular deck has enough of a tail that it could be possible to run it in electric 'manual' mode once you lift the front up! I don't think I've done an ollie on a skateboard since I was like 14, but will give it a shot and see if I or it survives. It's pretty easy to swap the trucks around between the front and back.

Is the torque from to motor only coupled through the magnets or is the coupler also a press fit
Seems the magnets may crack or come loose
Neo mags are only strong in compression

That's a valid concern. It's a pretty snug press fit as well. However, these are cheap turnigy hobbyking motors and that spacing between the magnets is anything but uniform, so I'm sure that there are a only couple that are taking the bulk of the load. However, if those start to chip a bit then the load will get quickly distributed amongst all the others too.

Scorpion motors? If so nice choice

No, that would be nice. Next time I will definitely use higher grade motors as the starting point, since the hongkong hobbycrap is always a disappointment once the novelty of being super cheap wears off. They've also got fairly poor no-load losses, just running the motors unloaded for 5 minutes and they are hot to the touch.

 

[justin_le]

wow, that's pretty crazy.

I am not a believer yet in weight sensing e-boards, seems like the holy grail for sure.
Had a very brief try on the Zboard.
there is also Goboard, and a couple of other attemps here,
http://wackyboards.blogspot.com/2010/06/z-board.html

Awesome link, I wasn't aware that someone had been compiling these builds, and nice to see that our original attempt from '07 made the list!

seen this other two eboards with hub motors?
http://wackyboards.blogspot.com/2011/03/hardbord.html

That looks super nicely done too, though I'd love to hear how well the RC motors hold up being the actual wheels and dealing with all the impact forces to which a skateboard wheel is subject. But if they hold up well and if the rubber is easily replaceable as it wears down, then that may be the better approach. Here are the relevant skate hub motor images from the above links:

 

[sk8norcal]

i forgot this one, one of my favorites...

the yamaha ivy, very little info on the board,
self balancing manuals ??
http://wackyboards.blogspot.com/2009/11/yamaha-ivy-sonic-skateboard.html

 

[bowlofsalad]

I definitely prefer the wheels on 'MetroMotive'. Maybe I have very little experience on skateboards, but I know on roller skates, a rock on pavement definitely tossed me to the ground more than once.

 

[sk8norcal]

of course on small rollerskate wheels,
ride 75 mm longboard wheels, you don't worry about small rocks.
big air tires will give nice smooth ride, they are also there to trip you when you run off the board.

I gotta say, this is the first e-skateboard related post here that didn't only make me want to go watch some more skateboard fail videos on Youtube. Weight-shift based control of a powered board is also the first thing I've heard of that might make skateboarding a slightly less terrible idea.

actually, i would be more nervous to try a high powered weight shifting eboard than a hand control one.

a low powered weight shifting eboard can definitely be easier for a beginner vs a regular skateboard,
dont have to learn how to push or footbrake. (this is true for hand control eboards also)

but guess what, so is a self balancing bike...
so is autoshifting and cvt for cycling.

i said before, if skateboarding doesn't sound like a good idea to you,
dont try it, its not for the uncoordinated and its not for someone who is scared or timid of falling.

btw, you should take a look at some BMX fail videos and tell me who slams harder, bmxers or skaters.
bicycle is the real danger toy :wink:

 

[justin_le]

actually, i would be more nervous to try a high powered weight shifting eboard than a hand control one.

a low powered weight shifting eboard can definitely be easier for a beginner vs a regular skateboard,
dont have to learn how to push or footbrake. (this is true for hand control eboards also)

I pretty much agree with this assessment too. The idea for this board is to have a ~500 watt power level, not a multi-kW machine, and for that the natural kinetic interface of leaning forwards and backwards for acceleration should meld perfectly with the way you lean left and right to steer on a skatebaord. But at higher power levels I don't know if I'd be so comfortable with it. There is a bit of built in negative feedback though which you don't get on a throttle control. As the board takes off, the acceleration puts your weight over to the rear again, causing it to slow down, and so this build in damping could be beneficial with the higher powers.

You also touched on one of the other things that I want to explore, and that is the idea of still using the motors as electric "assist", while you still push the board by kicking, but have the motors there to level out the terrain. I found with throttle controlled electric skateboards that it's awkward to kick them like a regular board while under power. Once your weight shifts off the deck and onto the ground, then the board tends to accelerate away from you. But with the weight sensing on the trucks, the device would in principle be able to detect when you are actively kicking, and it could swap over to a 'constant speed' mode where it it wouldn't accelerate out of the way as you push with the other leg, and it wouldn't be so sensitive to your forwards and backwards weight distribution. Then, once both feet are on the deck again it would revert back to leaning mode.

That's the idea at least. Today we just got the PCB's in for the strain gauge amplifier circuits, so some quick soldering ahead and we can start testing that out!

 

[xenodius]

Despite the cool-factor of a weight-sensing throttle, I think I'd prefer a tiny little remote hand-throttle like these Kickstarter folks use: http://www.boostedboards.com/

I think they're using a little outrunner at 2kw. Pretty hefty. But the cool part is the throttle... Roll it one way with your thumb to accelerate, the other way to decelerate. I think it's impossibly cool.

PS: Haven't read the thread, if they were mentioned... my bad.

 

[Chalo]

Dutch moms carry multiple babies on bikes sometimes. With bikes, it's all about what you do with them. With boards, the concussions begin right as folks step on, at zero mph. I was a kid in the '70s; I learned all I needed to know about using mechanic's creepers, furniture skates, and other foolish contrivances as personal transportation back when my little brother had his first Free Former.

But my hunch about weight-shift control is based on riding Segways. (I used to keep a small fleet of Segways in my machine shop as part of my job.) On the face of it, Segway is even more daft than a board, and it's got more power and torque than all but the most badass illegal e-bikes. But it's docile and easy to learn because it has an intuitive and somewhat self-stabilizing control system.

If a weight-shift board had a fast and robust control system like a Segway, it could begin to be safe and stable enough for sane people to use for pragmatic reasons. I'm not saying that the issues would certainly be solved that way, but unpowered boards or throttled boards are inherently the opposite of stable, sane, or pragmatic. Weight shift control offers something better, even if it isn't quite good.

Without some sort of better-than-human control, these things simply provide an awesome, bottomless supply of raw stupid that would solve all our problems if it could be converted into energy at an efficient rate.

EDIT: Skateboard ass-bustage videos removed to tidy up the thread a bit, out of respect for Justin's noteworthy and valuable build documented here.

 

[justin_le]

Despite the cool-factor of a weight-sensing throttle, I think I'd prefer a tiny little remote hand-throttle like these Kickstarter folks use:

The analogy is probably pretty similar to ebikes too. I used to believe I preferred throttle control over a torque sensing pedalec style ebike. But then, once I started riding more ebikes that had nicely integrated pedalec I realized that for the most part this was a nicer and more natural control interface between the human and the bike, since it's using the same input (your leg power) that you already use. Similarly, since leaning is such a core part of how you maneuver a skateboard (or surfboard, snowboard etc) I think that it would be a more natural control extention for motor power than moving one of your digits across a throttle. If you've ridden a throttle controlled board, you learn quickly that you need to lean forwards first before you engage it or the board will shoot out from under you. So the leaning already has to take place, and the brain has to learn to do that first in anticipation of you hitting the throttle.

In any case, our brains have a great way of figuring out control schemes and turning them into an extension of ourselves regardless of how they're implemented, so pretty much anything can work. Like gamers who achieve all kinds of delicate feats from pushing buttons on a keyboard or control pad, even though those button presses have nothing to do with the action going on the screen. We are used to pushing foot pedals to accelerate and brake a car, but why not instead have little throttles and buttons on the steering wheel for that? Or why have a steering wheel at all when you could use your feet to turn and then use your hands on a joystick to accelerate? I don't know if anyone's actually studying what is the 'best' way for a human to control a 2000lb vehicle, but given that we're no longer constrained by having to do everything with mechanical linkages like in the old days it's a question worth asking.

-Justin

 

[sk8norcal]

that's the thing, leaning to turn is natural. Leaning forward is unnatural to a skater.
leaning forward to accelerate is something completely new.
weight sensing forces you to do it.
with hand controller, i have to train my mind to lean before 'squeezing'.

Also having a 'soft start' helps alot with this issue,
There is only so hard u can accelerate an eboard anyways, (nothing to grab on to, unless u nosegrab)

You also touched on one of the other things that I want to explore, and that is the idea of still using the motors as electric "assist", while you still push the board by kicking, but have the motors there to level out the terrain.

yeah, that would be holy grail #2, kick-assist.

you are spot on regarding the weight shifting with high powered setup,
my back leg gets sore when riding the boosted.

here's holy grail #3,
carving and pumping.
there is a lot of fore/aft weight shifting going on,
experienced riders would not like to be lock into a 50/50 weight distribution at all times.
pump-assist?
carving under acceleration? deceleration?
this area is where i have major doubts about whether the board will work smoothly with the rider.

GoBoard did have a video showing him carving, but this is with your feet are lock into two fixed spots.

 

[sk8norcal]

Chalo, i knew you were gonna post irrelevant videos of street skaters slamming, :lol:
so predictable...
i could also post videos of bmx street guys slamming twice as hard,
but what's the point.

Go find me a video of someone taking a slam pushing a longboard on the flats.

riding on a longboard with big wheels on the flats at cruising speed ranks pretty low in terms of risk.
that's why thousands of college students using them daily to get around their campus.
and most of them are not even skilled skaters.

You want to see danger,
check out this push race through the streets of NY,
dangerous, YES! so are AlleyCat races on fixies....

https://www.youtube.com/watch?v=pVUO5zuy3jc
I would love to see an eboarder try to beat these top guys 8)

 

[t3sla]

yeah, that would be holy grail #2, kick-assist.

you are spot on regarding the weight shifting with high powered setup,
my back leg gets sore when riding the boosted.

here's holy grail #3,
carving and pumping.
there is a lot of fore/aft weight shifting going on,
experienced riders would not like to be lock into a 50/50 weight distribution at all times.
pump-assist?
carving under acceleration? deceleration?
this area is where i have major doubts about whether the board will work smoothly with the rider.
.

Gyroscopic sensors and a FPGA could achieve this

 

[montyp]

Awesome project Justin!!!

that's the thing, leaning to turn is natural. Leaning forward is unnatural to a skater.
leaning forward to accelerate is something completely new.

With respect, I disagree for longboarding. In the case of longboarding its pretty typical to shift your weight forward when accelerating downhill, especially when fighting a wind resistance.

In fact I'm curious how this system would work when getting speed wobbles. Since leaning forward is a way to kill wobbles, might want to shut the power off at high speeds. Or maybe this isn't an issue at all because the wobbles occur at a speed faster than your motors will provide power.

Another app of this design would be to monitor the frequency of the wobbles and dampen the system using differential steering on the two motors.

 

[EBJ]

Despite the cool-factor of a weight-sensing throttle, I think I'd prefer a tiny little remote hand-throttle like these Kickstarter folks use:

I would agree with you on this one, but if anyone is going to make a "lean to accelerate" skateboard to change my mind, it's going to be Justin.

Also having a 'soft start' helps alot with this issue,

I'm not a fan of software-controlled "soft-starts" of "soft-stops". I rode a popular board that boasts such features, and it was one of the scariest electric-skateboard I have ever ridden. I prefer a direct-drive "instantaneous" throttle-response. It seems like a "traction control" vs "no traction control" argument, and I guess I'm a "no traction control" guy.

There is only so hard u can accelerate an eboard anyways, (nothing to grab on to, unless u nosegrab)

I have not yet hit the physical-limit of acceleration. With decent grip-tape and the willingness to lean forward, I imagine it would take a lot of force to shear a rubber-on-grip-tape connection.