Be careful what you ask for!

Zamboni 6 photo

Several months ago, a humorous request went out for a Zamboni that could be used on the Nerdy Derby track.

Last year the Milwaukee Makerspace held a Maker Fest and a Nerdy Derby track was made for the occasion. The design allowed the track to be disassembled in 4 foot long sections.

When the track was reassembled, earlier this year, for the South Side Chicago Maker Faire, it was found that the joints did not match up as well as when it was first put together. Small ledges, that went up and down, would cause the cars to bounce off the track or hit the bottom of the car. Both of these scenarios prevented the cars from traveling freely down the track.

As many of you know, we just had a GREAT Maker Faire here in Milwaukee last month and the Nerdy Derby track was needed again!

We produced, and ran, over 1000 Nerdy Derby cars over the 2 day event. Wow!

Zamboni 10 photo

A month or so before the event I started working on an idea for a Zamboni type of device. My first thought was of a custom contoured planer that could be used at each joint to smooth them out. This idea seemed like too much work so I proceeded forward with my second design. This consisted of a simple sled hat used a drum sander, which smoothed out the high spots. Wood putty was then used to fill in any low spots.


More Banned Nerdy Derby Cars!

I made the car “Sling Shot” to enter in the recent 2012 Milwaukee “Nerdy Derby.”  During a few initial test runs, my car proved to be more than 10 times faster than the super-clever winning car made by HaveBlue.  Unfortunately, my  car was banned from competition because it was considered a threat to the spectators’ safety!  The consensus was that it had “too much momentum, or energy,” and would hurt someone if it went off the track and hit them.  What does too much energy mean, you ask?  Well, kinetic energy is ½ X Mass X Velocity^2.  Really then, the car was banned because the velocity is too high: It is just too fast!  I’ve already minimized the mass by making the car out of pine, although I could have made it from Balsa Wood, or even entertaining alternate materials such as these.

Anyway, its no fun to think of how to slow Sling Shot down so that its slow enough to safely race, but still fast enough to win.  Instead, I made some new car prototypes that amp up the speed and danger.  If I’m going to be banned in the future, I may as well get banned with style!

Below is a photo of Sling Shot, which traveled the 40′ track length is 0.1 or 0.2 seconds, for an approximate average speed of 300 feet per second, or 200 mph!  Note that the block is anchored to the finish line, thereby stretching the surgical tubing which acts as a spring to propel the car.

I realized that the dominant energy loss mechanism is air resistance – largely because Sling Shot’s wheels don’t even touch the track.  You see, the car doesn’t follow the contour of the track, it just heads directly to the finish line, through mid-air.  I spent some time engineering a more aerodynamic shape to further boost Sling Shot’s speed, searching for a shape that would really slice through the air.  I even consulted a team of highly trained German aeronautical engineer friends, who all approved of my slingshot propelled Henckel Car.  With the improved aerodynamic design, it should easily be faster than the 200mph Sling Shot car shown above.

The other car I built this weekend is also based on Sling Shot, but incorporates some classy chandelier bulbs.  The numerous ‘safety’ lights alert the time keeper of the imminent arrival of the derby car – for safety.

Unexpected Detour

When I arrived at the space Sunday, I had planned to work on a circuit board design in DipTrace.  After I left, I had spent six hours rewiring a golf cart.  Allow me to explain…

It all started when I went to take the trash out.  I used the golf cart with the flatbed to ferry the cans out to the dumpster.  After emptying the cans, I rode back and decided to charge the cart’s batteries.  Tom and Rich had just returned from lunch and Tom suggested we swap out batteries instead.  While swapping them out, we decided to also rewire them.  While rewiring them, part of the cart broke.  There’s a small white plate under the driver’s seat.  It’s about 4″ x 6″, likely made of asbestos, and holds a series of copper contacts that a lever attached to the gas pedal slides over to select the speed of the cart.  And it broke in two when we tried to tighten fix a wire on it.

We had a few options: try to mend the old, brittle plate, replace it with something new, rewire the whole thing, or scrap everything out for a solid state motor controller.  Not wanting to adopt a new project or sacrifice a motor controller that could be better used elsewhere, I volunteered to try and fabricate a replacement for the broken part.

First I documented everything just the way it was.  I labeled wires, took photos, scribbled down notes, etc.  Next I went about removing the broken plate.  There was probably more rust than metal on those bolts.  Then I took a pair of digital calipers and a ruler and measured the locations and sizes of holes for each component.  I considered using the CNC router or drilling a plate by hand, but the laser cutter seemed to be a much faster and precise approach.  I drew up my replacement plate in CorelDraw and found a scrap of 1/4″ acrylic that matched the size and thickness of the old plate.  After some tinkering with the printer driver and a dozen passes with the laser, I had a copy of the original in plastic form.

The next few hours were spent migrating the old parts over to the new one and wiring it back in.  Right around 7:00 PM, I tied some batteries together and the thing leaped forward.  A few more tests and it should be as good as new.  Someone suggested that maybe the plate was asbestos to avoid heating issues so we’ll keep an eye on that too.

Rotary Encoder – built into motor for Electric Car

My electric Dodge neon uses an AC motor and an industrial motor controller.  I upgraded from m 1984 motor controller to one less than 25 years old (actually less than 5.)

The new controller does much more than the old one and has the ability to do some fancy tricks.  At the moment I am running it in “sense vector” mode.  The controller senses the position of the armature by monitoring the current in the field coils.  This works great…   as long as the motor is spinning.  From a stop it tends to get out of sync, but there is a cure!

The controller can use a quaderature encoder so the encoder can read the position of the armature at any speed.

To add an encoder to the motor I decided to try a chip amde by Austrial Microsystem AS5040.  This chip senses a magnet near the chip and calculates the position of the magnet and can generate multiple output:  PWM, binary via I2C, and quadurature!

I bought a few of the chips and built a surface mount board to hold the chip and a few LEDs to display the output.  The first two version had a few problems but the 3rd time was the charm.



Thanks to Royce for working out the process for surface mount PCBs.


The final version had to be small enough to fit in a depression in the end of the motor cap.  The sensor centered and the whole board insulated (clear enamel)  since this is a grease pocket

for the rear motor bearing.





The magnet is mounted to a bolt that is threaded into a tapped hole in the back end of the armature.  It took a while to the position right (it needs to be within a few millimeters of the sensor) hence the nuts and washers.


The cable is brought out of the motor through a small threaded hole (it was an alternate location for the grease fitting.)  The hole is filled with epoxy and the wires go to a DB9 connector.   I built a small test board that shows the quadurature signals (4 round LEDs) and the status outputs from the chip (the two rectangular LEDs)






The motor controller puts out 15V to power an encoder and wants A and B as well as inverted A and B signals.  The circuit includes some NPN transistors along with a voltage regulator and a few capacitors to tie it all together.  I put the schematic for both the sensor and test board on one schematic so I could make both boards at the same time.

I installed it in the car today, but still need to put a few more parts together to run it.





It doesn’t work!

Ok, so the electronics work fine, it talks to the controller.

But it tops out at 256 pulses per revolution and the controller needs 1024.  It was a minor confusion between terminology.  The sensor detects 1024 positions, but to generate quaderature it uses 4 positions per pulse output.

Back to the drawing board.


I picked up a commercial shaft encoder on ebay for 50 that outputs 1024 PPR but it only works at 5V, so I’ll need a level shifter board and connector adapter.

Oh, yea, and I need to put the motor again, take out the old encoder, bring a shaft extension through the back grease pocket, add a grease seal and couple it to the encoder.