11 May 2012

Crüscooter Megaupdate 3: Build Part II and IT'S ALIVE!

Mmm, sleep. Now moar build. 

I dremeled out the spaces needed for the fork to fit, and got the fork mounted. 

From there, I mounted the handlebar assembly to the rest of the chassis, officially making Crüscooter a rideable kickscooter! Let's ride it around the labspace, ride it around and around, it's wonderful, right?

Wrong. I didn't constrain the front of the top plate because I forgot take into account how it would deflect under load, effectively throwing 2.001 out the window. Oops. 

After clamping down the deflected plate for a day, I drilled out a couple holes and tapped them. Two button cap 6-32 screws later, it was effectively fixed. 

The brake ended up working better than I expected. Two aluminum standoffs, one on each side, served as hinges for the two sideplates for the fender. A few tapped holes and some bent steel sheet later, and I've got a nice looking fender!

Here you can see the lever (left) that actuates the band brake (right). At the Cambridge Bike Shop, I purchased a brake cable and an old handmedown bike (that had the perpendicular adjustable adapter I needed to mount the cable to the lever) and got the brake working. 
After braking, the brake wouldn't reset, meaning the band brake's internal spring force was not enough to return the brake to its previous state. An additional spring pulling the band brake pulley back to its unactuated position solved that problem. 

Here's a bottom view of my assembled drivetrain.  


Now it's time to start entering the realm of the electronics. I made longer handlebars and mounted the throttle. I didn't want the throttle's cable just hanging out, however. I strayed away from the ugly but tried-and-true zipties in favor of snaking the cable inside the Razor's handlebars. 

I drilled a hole at the top where the throttle cable can immediately enter the Razor handlebar, and a hole at the bottom where it can come out. The result is elegant and cleaner looking than the zipties that dominate MIT. It also helps that I don't need to rout a brake cable, either. 

At this point, I went home for Spring Break for a week, anxious to get back, wire everything in one night 

Wiring. And why it sucks when you're sleep deprived.

Why is Blogger uploading this upside down!? Whatever. I'm tired. Here's a basic wiring diagram from my lab notebook. Time to start wiring things. 

Look at all them wires. Logic power (doubles as charger port), precharge circuit, motor power, throttle, indicator lights, 3-phase DC motor leads, hall sensor board... yup. All handled by that nifty Kelly Motor Controller

We interrupt this blog for another episode of What I Learned In 2.007: 

When in doubt, just cut the unknown connector off and solder on your own. Don't crimp it, crimps do not hold well.

Kay, got that out of my system. 

Here's my hall sensor board, provided to me by my instructor, necessary for our motor controller to be able to drive the brushless motor. I was really tired at this point, having pulled 2 straight allnighters trying to get everything wired up so I could ride the damn already, so I made a really simple mount out of a chunk of plastic available at lab. A couple minutes at a bandsaw later and I had the sensor board mounted. 

At this point, I was nearly ready to have my motor driving. After a bit of time of determining which order to plug the motor wires into the contorller in order to get it spinning forward, it was still being really noisy and generally frustrating. 
A couple of hours of debugging later and Charles and I determined it was a bad solder connection -_-. ('nother 2.007 lesson: MAKE SURE YOUR SOLDER CONNECTIONS ARE SOLID!).

And... IT'S ALIVE! 

Here's some video of me riding up a nearby parking garage. 
(Note, it's been alive for over a month now, I'm just bad at updating my blog.)

Yes, the cameraman is chasing me in an insanely fast gokart
And my fellow EV classmate David, who built this awesome thing with Jacqueline, is sprinting after me on foot.

It has great acceleration, and a decent top speed. I love that I can just hold down the throttle from a standstill and just go, thanks to properly tuned controller settings and the fact that it's a sensored system. It has nice ground clearance, easily going over railroad tracks and small curbs with its 8-inch pneumatic tires. I've also driven it for miles in the rain, and it hasn't skipped a beat yet. 

As a vehicle I've been riding every day for about a month and a half, I can say this thing is pretty robust. 

I can't pick just one favorite feature. I love how clean the entire thing looks and feels, the way the accelerator wire is routed through the handlebars, the awesome brake mechanism, the sensored acceleration performance from a standstill, the day-to-day range. I built the entire thing in no time at all, so the mechanical design is a cool feature on its own. I'll just go ahead and say it: Crüscooter freaking rocks!

Video from the actual challenge run: a timed drag race and a timed and watt-metered garage hill climb.
 I have the green shirt and black hoodie. Yes, Ass-Scooter Mode (c) at 0:40 is incredibly fun.

I think the one feature that really set my scooter's performance a step beyond the performance of most of my classmates' vehicles (I had the fastest scooter drag race time and the fastest AND most energy efficient garage climb of any of the class's vehicles) was the SK3 6364 motor that powered it. It is simply the perfect motor for a mid-sized scooter like mine. My friend Ian used the same motor as I, with similarly impressive results.

And it still fits inside Melonscooter

I've learned a ton this semester. I've gotten deadly with Solidworks and just designing things correctly and precisely. I carry my calipers EVERYWHERE now, and my experience has helped me a ton in 6.141 (Robotics: Science and Systems I, more on that later) for the final challenge, as well as in my general engineering experience. 

Thanks in large part to this guy,

I'm an exponentially better engineer for having taking this class, as it's shown me how to apply all the more theoretical principles I've learned so far as a Mechanical Engineer. It's allowed me to flex muscles I never knew I had like CAD design and machining parts. I've had to come up with many different designs, iterate through them to find the best one that fit my criteria. I've had to make compromises and last-minute design changes when things happened. 

I've learned a ton, and I really hope this class can continue to touch young aspiring engineers for years to come. I really want to be an Undergrad Assistant for the section next year, if at all possible. 

But first...

Yes, this is a thing. And it's freakin' heavy. Let's call it MegaArm for now, and wait til I pass my 2.004 (Control Systems Engineering) final for me to talk about it, shall we?

10 May 2012

Crüscooter Megaupdate 2: Build Part I


I got my motor and 14-tooth pulley first, so I started my machining there. I threw the pulley on a lathe and drilled the bore hole from 6mm to 8mm so I could fit it onto the motor shaft. I also drilled and tapped my own setscrew hole between a couple teeth to make sure it stayed on. 

Crüscooter Megaupdate 1: Final Design

Alright, the semester's nearly over and I now have a good excuse to update my blog: my grade depends on it. In order to not spam one semester's worth of blogging into a single post, I'll split it up into a few. Without further ado, Crüscooter's final design! 

I want a scooter that's portable (lightweight and foldable), has room for both my feet to be at rest, is reasonably weatherproof, can brake similarly to Razor scooter, can drive me around MIT's campus reasonably fast (20MPH), has decent acceleration and can go uphill, is robust against bumps and railroad tracks and such, and can generally put up with day-to-day use. Given these criteria, here's the design I came up with. 

Approach and Initial Decisions

My approach to designing the scooter was to stand on the shoulders of giants and learn from the successes and failures(read: lessons learned) of the many others who've built electric vehicles before me. 

I chose to use a timing belt-drive instead of chain-and-sproket or other transmissions because belts are smooth and silent. I also chose to go with a sensored motor controller because it can get me going from a standstill. Also, I need 16 Inches of free space on top to fit both my feet. Also, because the back wheel pulley had 72 teeth, I chose a 14-tooth motor pulley for a gear ratio of Kg = 72 / 14 = 5.153. 

Brake Design

The only real innovation for this entire project, my brake design combines the simple and intuitive Razor step brake with a band brake that doesn't wear the tire down. The fender also serves to prevent water from being kicked up all over my back when riding on wet surfaces. 

The result is a brake you step down on that pulls on a short brake cable, which is attached to the brake. The cable actuates a pulley on the brake which tightens the band brake to slow down the wheel. When tuned properly, the brake maintains full effectiveness while preventing the fender from ever contacting the wheel. To quote my instructor, "this mechanism is legit."

Choosing a Motor and Drivetrain

I need a motor and drivetrain that can get me to 20MPH as quick as possible. Hobbyking.com seems to have the best selection of Cheap-ass Motors, or "ICBM"s (Iridescent Canadian Banana Monsters?) so that's where I started looking for the perfect outrunner motor for me. 

First of all, I had the option of choosing 1 to 3 batteries. The batteries provided to us are A123 12V 4.5Ah batteries with 40-amp fuses. Because of the space each battery took up and the fact that I didn't want a 4-foot long scooter, I decided to run on 24 Volts (2 batteries). 

Motors are characterized by a constant that is manipulated in various ways to show off a different type of spec. On Hobbyking, that value is its Kv which has units of RPM/Volt. I converted these to Newton*meters/Amp with a little but of unit magic. I narrowed down my motor selection to the SK3 class, which has a 63mm can diameter (59 mm stator) and comes in different sizes. After doing some math, most of which is detailed in the Instructables link above, I picked a motor and determined my scooter would have the following theoretical specifications: 

Turnigy Aerodrive SK3 - 6364-190kv Brushless Outrunner Motor
Voltage: 24V nominal, 26.4V charged
Current limit: 40Amps

SK3 6364
Kv = 190 RPM/V 
Kt = 0.05026 Nm/A

Gear Ratio: = 72 / 14 = 5.153
Top Speed = 23.21 MPH
Max Acceleration = 1.1199 m/s^2

Additional Design Features: 


The chassis is 1/4" Aluminum plate because it is light, strong, and easily waterjettable. Using T-slots and tabs, the chassis can be assembled easily I chose 6-32 button cap screws and square nuts because they would sit flush against the chassis's surface. The front fork is also designed to be waterjetted and t-slot fastened, essentially copying Charles Guan's design (see above "failures" link, which is Charles's answer to a broken plastic front fork). 


In order to tension the belt, I chose to put slots that will allow the motor mounts to rotate about a pivot point, giving the belt a few centimeters of play. 


There's a hole in the side that mayOrMayNot be covered, but it allows me to access some wires on the internals, and the serial port. Yes, my scooter will have a serial port. Also, there's a hole on top that was originally for a Hella Switch, but the CAD model that website provided was like half as large as an actual switch or something. I ended up putting the status lights, precharge switch (A main motor power switch with a resistor, so you don't run infinite current into the controller when you turn it on) and a "Dean's key" for main power switching (That I learned from these guys) slot there instead.