Controlling the controls

Luckily, the Kelly controllers, while requiring some fancier wiring (pre-charge resistors, their own power source) than other motor controllers, were extremely easy to program. It is all taken care of in the series of menus seen below. To connect to the computer, you use the provided adapter (rear of the controller, above) to the serial port of a computer. The controller needs to be powered on to connect, but through its own power connection, not the power connections for the motor. The controller is powered through the power and ground prongs of the rear plug. Above, we used a pair of 9V batteries in series to power them. They need a minimum of 12V DC  to operate.

Once you start the controller programming application on the computer, it helpfully asks you not to run the motor while connected, and then walks you through setting the controller up. Here, we chose whether the motor should run at half in reverse (didn’t matter because we hadn’t yet implemented a reverse switch), and whether to have the controller auto-off if started up with the throttle pressed. We decided that was a usefully safety feature to avoid unintended starts.
 While you can choose from two different types of throttle controller, ours responded to either setting. We did set it so that some throttle movement was ignored, hoping to compensate for vibration as we bounded along in our unsprung kart.



Here we set whether to cut out if the voltage went above or below a certain level–important if you have a system that provides more voltage than a particular component can handle. You can also choose from a light, medium, or heavy load. We figured that propelling a human around a race-track should count as heavy.

After setting up the controller, you cycle its power to ensure the new options are loaded, and you’re good to go.






Warning: We only did this with professional assistance. If there is one thing you copy about what we did, copy our getting professional help with the batteries.

One of several tools unfortunate enough to get too close to our go kart batteries.

In the beginning:

There were some elements in the ground. A very, very long time later, after being turned into a Toyota Prius, some of those elements ended up on a table in a classroom. Thus, we received the full, Nickel Metal-Hydride, propulsion battery from a (2nd?) generation model.

The initial form we received: the cells are encased in a metal shield as one long block. The white plastic at the near end of the battery is the end cap of the cells. Note the white blower motor on top (far end) for ventilation and spacious allowances on each side of the case for air passage. Also note the thick rubber gloves being used. While the electronics mounted on the end cap (assorted gray boxes) were necessary for using the battery in the Prius, they will be left behind as they have no function for use in our go kart.
The battery sans casing. We all watched a professional in action.
Removing all of the additional pieces from the cells. The orange plastic contains brass plates that link each cell to the one next to it. We used these same orange carriers and brass plates to make our smaller kart batteries. They are quite useful as they simultaneously connect cells electrically, and limit the points of contact by other objects or persons to the cells.
The battery removed from its case bottom. The wires remaining ran to sensors for the car and are unnecessary for the kart. Some acid or base had stained the casing. This particular battery had spent several years sitting unused.
All that is left from the Prius battery: its 34 cells, of which we would only use 5 per battery.

The cells are held together by long rods with threaded ends. Once we removed the nuts holding it all together we could marvel at the lightness of the individual cells-only a couple of pounds each.

Those plastic end caps and long rods are there for a reason: the battery cells will expand if not kept under compression. Thus, before you disassemble the big battery, start planning how to compress the smaller kart batteries you will make. We used threaded rod and nuts to hold slabs of acrylic against the side.

The cells have their contacts on opposite sides; the staggered arrangement of pieces sticking up from the top of the cells is due to their being turned end to end to facilitate arranging them in series. Each cell provides slightly less than 7.5V when fully charged. By sticking 5 in series, we obtained a 37V battery for our 36V kart.

The finished connections for a full battery. The nuts, too, are recycled from the original Prius setup. Both sides of the battery are setup in the same fashion, with one crucial difference: On one side a negative terminal is open (thus being the negative for the battery as a whole), on the other side, at the opposite corner, the un-connected positive terminal becomes the positive for the kart battery.
In this picture, the battery is the large grey slab at the right of the electrical system. The slabs are acrylic sheets held to the side of the battery by long bolts; washers help keep the acrylic from cracking as you use it to compress the battery’s sides.
In the final configuration, we ran two batteries in parallel, mounted together to save space. Running the batteries in parallel was accomplished with 8-guage connector wires, which can be seen by the black wire coming up from the middle of the pair of batteries. In use, it would connect to the ground of the right battery, which was then also connected to the kart’s system ground.

Breaking the Brakes

Given more time to research how bicycle disc brakes mount, I would have taken us in a different direction. When using a split rear axle ideally you need one brake to mount left and the other to mount right; bicycle brakes do not allow for this mounting orientation. What we ended up fabricating was several unique components that will be pictured below that allowed us to use the stock bicycle mounts but orient them correctly to the vehicle. We also, thanks to input from another professor, mounted the brake calipers on a flexible system to allow the brakes to float so that they wouldn’t bind if anything on the vehicle shifted. This mounting has the potential to slightly reduce braking ability, but at the speeds power racing series vehicles operate it was determined not to be a major issue.

First a custom mount had to be fabricated to allow the disc to attach to the axle. This mount had to be removable but also strong enough to hold when the vehicle was braking. I designed a weld able steel mount that fit the bolt hole pattern for the disc brake, then welded it to one half of a two piece shaft collar. This meant the screws could be tightened very tight which gave us a good friction lock to the axle.

Mounting the caliper was eventually accomplished using 90° angle iron and screws. The screws were placed in clearance holes to the calipers and springs were put behind them to create the proper tension. This allows the floating discussed earlier. As you can see, the angle iron was welded to the bars which the motor mounts and axle pillow blocks attach to.

Having each brake independent allows us to break each wheel separately which tightens the vehicle turning radius and doesn’t require one to entirely get off the gas when braking around a corner.

Waterjet Funtimes


I am lucky enough to have access to an abrasive waterjet at my work, which makes the fabrication of two dimensional parts significantly easier and faster. An abrasive waterjet works by forcing a large volume of water through a very small orifice creating a fine stream of very high pressure water (50,000+ P.S.I). This water stream is mixed with an abrasive powder of a specified grit (size of grain) to aid in the rapid cutting of material. A water jet can cut a variety of different materials ranging from insulation foam blocks to steel. It allowed us to rapidly fabricate the motor mounts as well as stronger driving sprockets, brake disc shaft mounts, and motor mount spacers.

20161122_150716The motor mounts were made from 0.5” 6061-T6 aluminum. Aluminum is a good material to use for this application because it has a high strength to weight ratio and is easy to machine. The design for the motor mounts was chosen to allow the pillow blocks to be mounted directly underneath without interfering. I would have liked to use thinner material to save on weight but 0.5” was used because slots were needed for the motor mounts to be adjustable with the axles. The front mount allows the motor to pass through and is then secured with a set screw to keep it from wiggling when in place. In the future I would modify the mount drawing to leave a bit more tolerance for the motors as they are not perfectly round or exactly the same diameter all the way through. I would also recommend prototyping the fit with a cheaper material as I wasted several large pieces of 6061 in getting the fit to each motor correct.


The original Drive Sprocket, the 3D prototype, and the waterjetted new drive sprocket.

The drive sprockets needed a little bit more post processing after being cut on the waterjet because they needed ramped angles to ensure the chain stayed on them and didn’t slip if it got a little loose. Daniel Raver accomplished this task carefully using a belt sander. They were first prototyped to ensure a good fit and then waterjetted and processed.

Given enough time and raw material many more components could be made with the waterjet. Balancing time and cost is important though and it is often more viable to just by a prefabricated part then to make it yourself, but what is the fun in that?


Product List

1 Go-Kart Frame


Steering System

1 bicycle handle bars w/ front steering column

1 3/8 bolt w/ nut

1 ¼ hexkey bolt w/ nut

¼ bolt w/ nut


Front Axis System

4 tie rod ends

2 front tires w/ bolts

2 steering spindles left and right

2 front axil bars


1 wooden front floor panel

1 battery under carriage panel

1 placement rib


Motor System

2 motors

2 front motor mounts

2 rear motor mounts

4 1 ¼ washers

4 ½ washers

4 5/8 washers

4 ½ bolts w/nuts

4 ½ hexkey bolts w/nut

4 ½ 1.5 inch bolt




4 3in ¼ bolts w/nut

2 wooden bumpers


Rear Axis System

2 pillow block spaces

4 pillow block mounts

4 3/8 2 ¼ bolt w/nut

4 3/8 2 ¾ bolt w nut

2 motor chains

2 rear tires

10 axil locks

2 11 inch rear axil rod w/ mount plate

2 braking disks

2 motor sprockets

8 hexkey 3/10 bolt w/nut


Braking System

2 handle brakes

1 5 ½ ft brake cord

1 4 ft brake cord

1 left brake

1 right brake

4 1in washer

4 ½ in. washer

4 springs

4 hexkey ¼ bolt

Mounting the Braking System

Mounting the MTB bicycle braking system. Front and Back braking kit by Gasbike.

Based on the steering configuration that we gone with, we decided to use a Front and Back braking kit by Gasbike. The steering bar which came from a recycled bicycle allowed the 2 braking handles to easily be fitted on.  We connected the braking cables to the braking handles.  With the braking kit that we ordered, one of the cables was shorter than the other.  We decided to assemble the shorter cable to the left hand brake handle.  The longer cable was long enough to be able to run down the steering mount and along the right side bar of the frame to where it was able to connect to the right brake already mounted on the back.  The left hand cable or shorter cable was ran down the steering mount and down the middle above where the battery was to be placed, with enough slack to place and remove the battery, and connect with the left brake already mounted on the back.

tmp_30066-20161117_1452211472576644When running the cable through the brake, we need to ensure that the cable is as tight as it possibly can be. To do this, it required 2 people one to hold down the handle for the steering brake handle and compress the braking arm on the brake mount, while the second person took pillars and pulled the cable as much as possible to ensure a tight fit. This configuration allowed us to make a dual braking system.

Seat & Frame additions

For this project, most of the frame additions such as the battery carrier, seat mount, and petal mount, we just simple scrap wood that was available.

For the battery mount, we simply cut several ribs made of wood that would act as sort of an under carriage.  The ribs were cut at the right measurements to where it would fit tightly against the frame.  The ribs were secured to each other by 2 pieces of ¼ inch wood board, that we drilled screws into.  The sides of the wood boards facing the battery pack would have to 2 notches cut out of them for the screw ends from the battery pack would fit into.

For the Petal mount, we cut a piece of wood board that matched the width of the go-kart frame.  This would allow us to drill holes into the wood board and allow us to secure the board to the frame with something simple such as zip-ties.  With the board that we used we had to cut a square out to allow room for the steering column to move freely. To mount the 5-volt foot throttle by Cisno we had to cut out a hole in the board to allow the pedal to comprise down.  We mounted the pedal by cutting 3 holes that matched the mounting brackets of the pedal and running 3 screws underneath to secure the pedal from the bottom.

The seating mount was assembled using several pieces of wood boards. Based on certain time constraints we were unable to use metal as the mount.  The seat that we used was a simple chair seat from recycled material.  At the bottom of this seat, there was a mounting plate that allowed us to bolt the chair to pieces of 2×4 which was the length of the go-kart frame.  To have a proper height off the frame for the driver.  We simply secured more wood boards together.  For us to mount the seating mount onto the frame, we drilled 2 holes on each side of the frame where the seat was to be mounted using a 3/8 drill.  We ensured that the all set of holes from the wooden mount and the frame matched up and tightly bolted down the seating mount to the metal go-kart frame.  To secure all the boards and seat together we used wood glue as well as 4 6 inch 3/8 bolts.

Initech Printer Electrical Systems


-Kelly Microcontrollers (2) nominal: 36V working: 24-48V

-On/off safety kill switch

-40 amp fuse

-1k Ohm, 10 Watt  pre-charge resistor

-NY1020 48v 1000w motor (2)

-DPDT switch for pre-charge resistor

– 3d-printed mount for resistor, DPDT switch, fuse

-DC-DC voltage converter (in 18-75vDC, out 12vDC): CUI Inc. PYB10-Q48-D12-U

-throttle pedal input: 5V output 0.8-4.2V

-37.7v 5-cell NiMH batteries from Toyota Prius (2)


8AWG wire, approximately 15 feet (~5m) before cutting: 2 colors to differentiate positive and negative/ground, 13 total wires

20AWG wire, approximately 18 feet (<6m): 5 colors–positive and negative, different signal lines such as throttle, 14 total wires

-electrical connectors: inordinate, just keep buying them until they’re out of stock (26 of 8AWG and less than 28 of 20AWG–some of the small gauge wires are for plugs in the Kelly controllers, which come with connectors to wire up the paired plug


-2 Power Systems: one to power motors, and one to power motor controllers (here taken care of by voltage dropping on 2nd set of wires (37-12 voltage converter)

-Pre-charge resistor: small amount of current allowed through resistor to charge capacitors before fully opening main power contactor


-Max: 1,440 Watts


-Reverse switch-perhaps in a future iteration

Necessary Thank yous: Cordell @ Baltimore Hackerspace, Kung @ RUGreasy, Harvey @ CCBC Automotive

Initial Decisions:

To start our build, we chose a 36 volt system. Thirty-six volt systems have been used extensively in the Power Racing Series, providing guidance in our system build. Equally important, with 48 volt motors, any over-voltage problems in our nominally 36 volt system would not damage the equipment.

Power Source:

The just-opened Prius battery case.

A Prius propulsion battery was donated to the team early in the build. With professional assistance, we parted out the battery into separate cells so that we could build individual ~36v batteries.

The Prius battery cells in one long block, held under compression by the rods and plastic end caps. All conduits between cells have been removed.
Five cells in series provide just over 36 volts at full charge.
Re-attaching connectors between individual cells to create the complete battery. The unconnected terminal will form one of the terminals for the battery as a whole.

Currently, we are uncertain about the total amp-hours available from our batteries, and so plan to run two batteries in parallel to reduce drain on the system until we have more detailed data.

Motor Controllers:

Though our initial decision to use Kelly motor controllers was cost-driven, they have proven to have several unanticipated strengths: For one, they can interact directly with a throttle rather than requiring a separate controller (such as an Arduino) to interpret between them. Second, though they require a serial port to communicate with for programming purposes, the interface is extremely simple and straight-forward, essentially amounting to answering a couple of questions.

On the other hand, the capacitors in the controllers require a pre-charge resistor to charge them before the system is actually “turned on” by the main power switch to prevent the initial current in-rush from either destroying the capacitor or shorting across the power connectors.

Left to right: 40 amp fuse, DPDT switch to control pre-charge resistor, and pre-charge resistor (under electrical tape). The box was designed and 3d printed in plastic.

Thus, there are actually two parallel circuits at the point of the main power contactor, the switch itself, and by its side, the smaller DPDT switch and resistor.

With some unrelated items, the motor controller system and its transparent plastic case. On/off contactor is at lower right.