Search This Blog

Monday, August 15, 2011

Win!

The ELB successfully ran for the first time yesterday, after a year and a half of work. WOHOO!!!!

Let me apologize in advance for the shitty pictures...still using my cellphone camera.

But back up a day:

I modified the pistol grip AM r/c car controller I bought off of ebay to be more useful for controlling a longboard. In otherwords, I took it to a bandsaw:

Unscrewed.

Random PCB that didn't have anything connected to it...

The control board.

post-bandsaw action

Post dremel action. That bit is amazing for carving plastic. The white shiny stuff on the right is just bad lighting.

It fits perfectly!


Wired up.

I only needed 5 AA's to power the board, so I cut the other 3 off.

I also cut out small delrin plates to cover up all of the holes in the controller, but didn't put them on yet in case we needed access to the transmitter board (which we did).

I needed to finish pressing on the scooter tires next. I used the exact same process as last time.


I made a boo-boo milling this one, and a approximately 20 degree arc ended up too large in diameter. I just pressed a piece of aluminum in the gap after I put the tire on.


It looks awesome.

Re-wired.
After a few hours of shane being awesome and tinkering with software...



IT RUNS!!!!!!!!!!

All four motors spinning.
The thing is a beast. It sounds almost like a jet turbine spinning up. Unfortunately, there is a lot of mechanical friction, especially in one of the wheels (too tight of a press fit on the large bearing), which results in about 300W total at full speed no load. So the mechanical/electrical/magnetic losses are eating about 2/5 hp at full throttle, which sucks.Probably a combination of machining errors, press-fit errors, and really cruddy bearings (VXB's cheapest bearings).

Lessons learned up to this point:
  • Scooter wheels make awesome looking tires.
  • Mechanical accuracy is essential for efficiency. 
  • Shitty bearings are shitty.
  • Two motors is probably more efficient despite increased current/windings necessary to maintain torque.
  • Find a better way to have charge leads and battery disconnects outside. The current setup is wayy to sketchy. *This is the part that I am totally at a loss on and could use advice.*
Here is a video of it running:
 
Things to listen for: the awesome spin-up, the horrible sounds of the shitty bearings, the slipping of the hubs on the axles (the "cogging") , the battery box rattling like a shopping cart.
Things to watch for: The magnets taped to our feet. The really cruddy turning. You can see attempts to turn; the board is even all the way rolled at points, but you can see that the risers are giving way and the trucks aren't torquing.



We definitely got the breaking right, haha. Thanks to Shane for taking video!


Post run.

You can tell by the wear patterns on the tires that the axles are slightly swept.

Results:

The steering sucks....bad. Like 15 ft turning radius bad, which was really disappointing because I'm so used to the awesome trucks on my current longboard/mountainboard thing. Part of it is the trucks, I think. But the majority of the problem is coming from the giant stack of risers you can see in the above picture. The risers are there for 3 reasons.
1. Vibration/shock absorbtion. Mountainboards usually rely on the deck springyness to absorb shocks, but I bolted a 1/4" aluminum battery box to the bottom, which makes it super stiff.
2. 2 of the risers on each side are angled, which gives me an extra 15ish degrees of truck tilt, which should aid steering.
3. They raise the ground clearance under the battery box enough in order to turn. Without them, the battery box will scrape the ground. And I can't mount larger tires without significant ($ and time) overhaul.

The last reason is the most detrimental...it means that I really need that much riser.  But why are they a bad thing? Well, when I lean out to turn, the deck rolls, causing a torque on the trucks, which causes them to pivot. The rubber (which makes for a very flimsy/flexible mount to the deck), just flexes in the opposite direction, instead of forcing the trucks around into the turn. I totally should have (but didn't) foreseen that.

Another problem came up: Even with the axle set screws as tight as I could make them, the hubs were still slipping around the axles.

There's also a weird control issue. While the transmitter is transmitting a slight breaking command at the neutral point, and 3 of the 4 motors are listening, one of the motors actually accelerates at neutral. We have no idea why. The same motor also oscillates while it's doing this.

It's also blatantly obvious that it's wired too much for torque and not enough for speed.

Possible Solutions:
1. Not care and move on to V2 - HeavyBoard. Leave ELB 90% complete
2. Fix/finish ELB. This involves machining a stiff plastic riser to replace the rubber ones (which might not even fix the turning issue), grinding flat spots in the axles for the set screws to set, fixing the weird control issues (not sure if possible), implementing the field oriented control, adding a fan circulation system, adding lights, waterproofing...probably another 60 hours worth of work for a board that won't be very comfortable to ride.

While it would be nice to finish ELB, after spending thousands of dollars and so many hundreds of hours working on it, it's probably a better move practically to leave it alone and start V2. On this note, I'll part with ELB for $3000 (controllers and battery charger not included) if anyone is interested, haha.

Lessons learned:
  • Folded sheet aluminum battery box. Lighter and more flexible.
  • Don't use skateboard style trucks on a mountainboard.
  • Avoid massive risers if possible.
  • Figure out a better hub-to-axle interface.
  • Follow all lessons learned from ELB on V2.

4 comments:

  1. considered replacing a couple of your current risers with metal or hard plastic ones, this would reduce your flexing to a single riser on front and rear

    ReplyDelete
  2. There are ~2" tapered risers you can buy that would fix your turning radius issue! You can see them here (http://sfbay.craigslist.org/pen/spo/2550941565.html). I've spent the past hour searching for somewhere to buy them without luck, but they're out there! It would make for a high ride, but killer carving. :)

    ReplyDelete
  3. Just some advice for your steering issue: I think you're running all tires at the exactly same speed, that's what the hall sensors are for, right? When you're driving a curve, the outer tires spin faster than the inner tires. that's no problem when you have four independently spinning wheels. On cars this issue is solved with a differential in the axle, but on your board the controller adjusts the speed of each tire to a fixed value, right? You should check the influence on your steering. You could compensate that with a gyro and some expanded software for your speed controllers.

    ReplyDelete
  4. Thanks for the risers tips.

    Philipp- The tires are not being run at exactly the same speed. No, that's not what the hall sensors are for. Each motor/controller combination can be thought of as an independent entity. The controllers do not implement a fixed voltage (fixed speed) control algorithm. Instead, the controllers are current (torque) controlled, so the motors will automatically adjust to the speeds they need to be running at. Think of it like an electronic differential.

    ReplyDelete