It’s indeed a little bit of try and error. You have to find the right point where the chuck of the lathe clamps hard enough to turn the wheel evenly during processing and at the same time having only the minimum affordable deformation of the wheel. Another possibility would be do turn yourself a slotted can which fits tightly on the wheel, applying even pressure. I only thought of that option but didn’t use it by myself. Instead of turning with a lathe tool you can use a hole saw if you find one with the appropriate diameter.
I think making a mold and pouring pu is a sure thing compared to cutting up expensive wheels where it’s very hard to get a good cut or it centered. Have you tried increasing the motor amps instead of a sensor. If the motor only has access to tiny amperage at the slowest speeds/ duty cycle, especially relevant on a hub motor, it cogs a lot in my experience, but with the amps available I can go from 0 without almost any cogging or sound. Motoramps.com was put up by Devin which shows how to chose for consistent wattage at all rpm.
@The_Dude thanks for the tips, I will see what turns out better
@Hummie I agree, the problem is that here in Brazil I didn’t manage to find the urethane, my idea was to 3D print a mold and pour
could you do a tutorial on that? how is the pu quality compared to kegel/flywheels?
I was thinking about making my own urethane rings for the trampa hubs but not sure if ill accept the pu quality. 
1 Like
Compared to kegel wheels the room temp pu I got from bjb enterprises grips just as good, can be just as coshy soft, but it doesn’t have the rebound and won’t roll as far on a “push”. Hot pour stuff is harder to get and has better rebound but it’s largely unnoticeable Making a mold I’ll make a video next time maybe. I like to use coconut oil’s ability to melt and do two part silicone 50duro molds . Prop the wheel with anything in a big enough can an inch off the bottom, half cover in liquid coconut oil, harden, pour silicone over second half, melt off the oil, flip, and then pour the second half, with a wide pour hole and breather out of wax. That’s pretty much it. Bjb or jbj enterprises stuff is what I used. Get the long pot life. A vacuum chamber and pump you might be able to get away with if you mix it slowly at the bottom and use a solid color.
2 Likes
Video tutorial would be great.
Thanks for your answer. I was also thinking about moulding on my own but I really have no clue how to do it. Any help is appreciated. My motivation is not the money I have to spend for some wheels, I was more thinking of a better and durable connection between rotor can and pu-wheel. Do you think it is possible to use the rotor can as the inner boundary of the mould? Best case there will be no urethane slipping any more. Btw. Another question - has anybody tried PU-Adhesive to bond the wheel to a rotor?
Hall sensors vs. motor amps. The startup stuttering I get using FOC does not depend on motor amps/torque at standstill. I even get it on the bench where there is really no load on the wheel. Sometimes you have this stuttering sometimes not (and I already tried different motors). It only depends on the rotor position when starting. The only thing what helps with the current firmware is using hall sensors when using FOC. BLDC is a completely different issue - clean startup every time even without halls.
My motor amps are “almost” unlimited (2 x 80 A) - so I had the issue that when testing my board the first time it went WOHHHHHHH, while leaving me on the asphalt 
I know the Devin-theory, I understand and did the math but doing so I think you will get deficiencies either in the low or high ppm range. It is always a compromise and for the sake of even acceleration you cut down the abilities of the drive somewhere. What helped was defining a suitable throttle curve (thanks @ackmaniac) and also introduce more PT1 delay in the BLDC-tool advanced tab.
O I see there’s no grooves on the rotor to adhere to. Does it slip, have you got it going yet? Maybe you could epoxy some symmetrical grooves on and then get a 3D print made of what tire will match make a mold like that. But that’s problematic and time and think u should skip it and just glue it on Using the rotor I think is more trouble then pvc then Maybe you missed Jacobs hub motors that slip and people are gluing them on or adding attachment ribs to the rotor. Gluing on might be an easy solution as long as it’s not too hard to get off. You could make a mold of a simple inside diameter then and could make the tire plug/dummy with just pvc or something. Making the mold will be easy and with the pvc dummy especially cheap.
I don’t get you with the ppm efficiency. I though simply low duty is a bit less efficient but that’s it. What’s the pt1 delay do ?
Nice! I used tiny hall sensors to fit it between the stator slots.
3 Likes
A hole saw in a drill press works too.
1 Like
Bummer - why didn’t I read your thread before? Great solution with these tiny halls - would have saved me several hours and headaches. Thanks!
Sweet build, dude. Incredible documentation.
As for the wheels, I have an idea that might be worth a try. Back when I was a golfer, I used to build my own clubs. They have a tried-and-true way a keep the rubber grips securely fastened to the steel shaft of the club. Essentially they use double-sided tape that adheres securely to both the shaft and the rubber, and a solvent that temporarily turns the glue into a lubricant.
So you pour or lathe your urethane to 1-2mm smaller than the rotor can and apply the double-sided tape to the can. Heat up the urethane with a blow dryer so that it expands just enough to be pushed on. Spray that tape down with the solvent and jam that wheel on while the tape is nice and slick. The solvent will evaporate, leaving the glue which creates a very strong bond.
This method prevents slippage against the torque forces of swinging a golf club; it may work for wheels.
https://www.amazon.com/dp/B00SRGJ53W/ref=cm_sw_r_cp_apa_uCjazbV9RXVWT
3 Likes
Until now I have no issue with slipping urethan - I even think I know why, in the post above I described this in detail.
Regarding pwm (sorry not ppm, parts per million) I meant not efficiency but rather deficiencies/shortcomings. Approximately you have the low pwm / high motor current range at startup and the high pwm / high battery current range at end speed. If you go for constant power you will not be able to get the full available power/performance what your system is able to provide, esp. if you have high power LiPos. PT1 is a term from control theory. Practically this is a measure how fast the motor reacts or how fast he will adhere to the controller input. Benjamin Vedder showed this in one of his earlier videos. The corresponding parameter is “Positive ramping time constant” in the Nunchuk tab.
I don’t know what u mean by going for “constant power”. What u wrote above makes sense abou motor and bat amps. The way I have the vesc it puts out the same 900watts regardless of rpm/duty. I could increase that constant or unbalance it with more power at high speeds or low.
The pt1… Does that change the ramping of power so while the motor/bat amps may be set to something it determines the speed it’s applied after the throttle is applied?
Last chapter:
VESC and Mounting and Housing
This is not really closely related to the hubs, but it think it’s a good conclusion to the project. Dual hubs, so dual VESCs are needed. I decided to mount the VESCs directly onto the board, for convenience. Since the board has some reasonable flex I put them onto small rubber dampers. The assembly is shown in the next pictures - no words.
I’m a little proud to say that I successfully soldered both VESCs on my own 
The master VESC is connected to the receiver for my Nunchuk RF or Kama, the slave to a bluetooth LE module for the connection with the @Ackmaniac app.
The controllers are covered by an 3D printed enclose, contour crafted to the boards shape.
And this is how the board looks right at this very moment.
Currently I use some LiPos which I reshaped to smaller height, fitted with velcro and three of my puristic Dude clamps. No switches, just an XT90 anti-sparc plug and balance plug. From an aesthetic viewpoint I would love to go for the beautiful lower surface covers and housings of @whitepony. We’ll see … I already looked for some fancy vacuum pumps and epoxy resin tools 
. LiIon batteries are also on my list. My spot welder is already made and ready for use - but that’s another story.
The link to the 2D-drawings is now added to my first post. Thanks for reading!
15 Likes
that spot welder looks really familiar 
True, true 
[quote=“Hummie, post:54, topic:21519, full:true”] I don’t know what u mean by going for “constant power”. What u wrote above makes sense abou motor and bat amps. The way I have the vesc it puts out the same 900watts regardless of rpm/duty. I could increase that constant or unbalance it with more power at high speeds or low. [/quote]
With constant power I mean exactly what you said, e.g. 900 watts regardless of rpm/duty. I prefer to go for the maximum (to be honest, almost) for motor, batt no matter if this reveals different watts for different rpm/duty and adjust the behaviour by max_watt and the shape of the throttle curve.
Suppose according to your joystick deflection (watt or current mode) you are traveling in a steady state. Then, you suddenly pull the joystick further for some amount/degree. The parameter now determines how fast the VESC is aiming to reach the new state you are demanding through your input. You can experience this very easy, while putting the board on the bench and observe what happens when you quickly move the joystick, depending on the value you’ve set. When I remember it correctly the default is 0.9. If you set it e.g. to 5 the response of the motor to changes is more smooth.
1 Like
I was contemplating soldering my own vesc as well. Did you start from a semi assembled board? Any pointers you can give (such as where to buy a quality pcb board)?. Is it doable with a soldering iron (soldering station with adjustable temperature) or do i need a reflow station?
Man, upload more videos, your work is just amazing