How to Charge Any USB Device by Riding Your Bike

To start off, this project was started when we received a grant from the Lemelson-MIT Program. (Josh, if you're reading this, we love you.)
A team of 6 students and one teacher put this project together, and we have decided to put it on Instructables in hopes of winning a laser cutter, or at least a t-shirt.

What follows, is a compilation of our presentation and my own personal notes. I hope you enjoy this Instructable as much as we did.

I'd also like to thank Limor Fried, creator of the MintyBoost circuit. It played an key role in our project.


Jeff Brookins
Divine Child InvenTeam Member

Our original project was to develop a product that used the Faraday Principle to allow runners to charge their iPods while they run. This concept would generate electricity the same way those Faraday flashlights do.
However, we had a problem. To quote my team mate Nick Ciarelli,

"At first we considered using a design similar to one of those shake-up flashlights and converting it so that a runner could strap it on for a run and have energy to charge their iPod or whatever device they use. The shake-up flashlight gets its energy from the interaction of the moving magnetic field of the magnet in the flashlight and the coil of wire wrapped around the tube the magnet slides through. The moving magnetic field causes electrons in the coil to move along the wire, creating an electric current. This current is then stored in a battery, which is then available to use for the flashlight bulb/LED. However, when we calculated how much energy we would be able to get from a run, we determined that it would take a 50-mile run to get enough energy to charge one AA battery. This was unreasonable so we changed our project to the bike system."

We then decided to use a bike-mounted system instead.

We initially theorized the development and feasibility of a regenerative braking system for use on bicycles. This system would create a mobile power source to extend the battery life of portable electronic devices carried by the rider.

During the experimentation phase, the regenerative braking system was found to be incapable of fulfilling its dual functions simultaneously. It could neither produce enough torque to stop the bike, nor generate enough power to recharge the batteries. The team therefore chose to abandon the braking aspect of the system, to focus solely on the development of a continuous charging system. This system, once constructed and researched, proved fully capable of achieving the desired objectives.

To start off, we had to design a circuit that could take the ~6 volts from the motor, store it, and then convert it to the 5 volts that we needed for the USB device.

The circuit we designed complements the function of the MintyBoost USB charger, originally developed by Limor Fried, of Adafruit Industries. The MintyBoost uses AA batteries to charge portable electronic devices. Our independently constructed circuit replaces the AA batteries and supplies power to the MintyBoost. This circuit reduces the ~6 volts from the motor to 2.5 volts. This allows the motor to charge the BoostCap (140 F), which in turn supplies power to the MintyBoost circuitry. The ultracapacitor stores energy to continuously charge the USB device even while the bike is not in motion.

Selecting a motor proved a more challenging task.
Expensive motors provided the proper torque needed to create the braking source, however the cost was prohibitive. To make an affordable and effective device another solution was necessary. The project was redesigned as a continuous charging system, out of all possibilities the Maxon motor would be a better choice due to its smaller diameter.
The Maxon motor also provided 6 volts where as previous motors gave us upwards of 20 volts. For the latter motor over-heating would be a huge issue.

We decided to stick with our Maxon 90, which was a beautiful motor, even though its cost was $275.
(For those wishing to build this project, a cheaper motor will suffice.)

We attached this motor close to the rear brake mounts directly on the bike frame using a piece of a meter stick between the motor and frame to act as a spacer, then tightened 2 hose-clamps around it.

For the wiring from the motor to the circuit several options were considered: alligator clips for mock up, telephone cord, and speaker wire.
The alligator clips proved to work well for the mock up design and testing purposes but they were not stable enough for the final design.
The telephone wire proved fragile, and difficult to work with.
Speaker wire was tested due to its durability therefore becoming the conductor of choice. Although it was stranded wire, it was much more durable due to its larger diameter.

We then just attached the wire to the frame using zip-ties.

Tackling the circuitry was the most difficult challenge of the process. Electricity from the motor first travels through a voltage regulator which will allow up to a continuous five amp current; a larger current than other regulators would pass. From there the voltage is stepped down to 2.5 volts which is the maximum the BOOSTCAP can store and safely handle. Once the BOOSTCAP attains 1.2 volts, it has enough power to allow the MintyBoost to provide a 5 volt source for the device being charged.

On the input wires we attached a 5A diode so that we don't get an "assisted-start effect," where the motor would start to spin by using the stored electricity.

We used the 2200uF capacitor to even out the power flow to the voltage regulator.

The voltage regulator that we used, an LM338, is adjustable depending on how you set it, as seen in our circuit diagram. For our purposes, the comparison of two resistors, 120ohm and 135 ohm, connected to the regulator determines the output voltage. We use it to reduce the voltage from ~6 volts to 2.5 volts.

We then take the 2.5 volts and use it to charge our ultracapacitor, a 140 farad, 2.5 volt BOOSTCAP made by Maxwell Technologies. We chose the BOOSTCAP because its high capacitance will allow us to hold a charge even if the bike is stopped at a red light.

The next part of this circuit is something I'm sure you are all familiar with, the Adafruit MintyBoost. We used it to take the 2.5 volts from the ultracapacitor and step it up to a stable 5 volts, the USB standard. It uses a MAX756, 5 volt boost converter coupled with a 22uH inductor. Once we get 1.2 volts across the ultracapacitor, the MintyBoost will begin to output the 5 volts.


Our circuit complements the function of the MintyBoost USB charger, originally developed by Limor Fried, of Adafruit Industries. The MintyBoost uses AA batteries to charge portable electronic devices. Our independently constructed circuit replaces the AA batteries and supplies power to the MintyBoost. This circuit reduces the ~6 volts from the motor to 2.5 volts. This allows the motor to charge the BoostCap (140 F), which in turn supplies power to the MintyBoost circuitry. The ultracapacitor stores energy to continuously charge the USB device even while the bike is not in motion.

In order to protect the circuit from external elements, an enclosure was necessary. A "pill" of PVC tubing and end caps was chosen, with a diameter of 6cm and a length of 18cm. While these dimensions are large when compared to the circuit, this made construction more convenient. A production model would be much smaller. The PVC was selected based on durability, nearly perfect weather-proofing, aerodynamic shape, and low cost. Experiments were also performed on containers crafted from raw carbon fiber soaked in epoxy. This structure proved to be both strong and light weight. However, the construction process was extremely time consuming and difficult to master.

For the capacitors, we test two different types, the BOOSTCAP and a super capacitor.

The first graph depicts the use of the supercapacitor, which is integrated with the circuit so that when the motor is active, the capacitor will charge. We did not use this component because, while the supercapacitor charged with extreme speed, it discharged too quickly for our purposes. The red line represents the voltage of the motor, the blue line represents the voltage of the supercapacitor, and the green line represents voltage of the USB port.

The second graph is the data collected with the BOOSTCAP ultracapacitor. The red line represents the motor's voltage, the blue is the ultracapacitor's voltage, and the green line represents the USB port's voltage. We chose to use the ultracapacitor because, as this test indicates, the ultracapacitor will continue to hold its charge even after the rider has stopped moving. The reason for the jump in USB voltage is because the ultracapacitor reached the voltage threshold necessary to activate the MintyBoost.

Both of these tests were conducted over a period of 10 minutes. The rider pedaled for the first 5, then we observed how the voltages would react for the final 5 minutes.


The last picture is a Google Earth shot of where we did our testing. This picture shows that we started at our school, and then did two laps at Levagood Park for a total approximate distance of 1 mile. The colors of this map correspond to the speed of the rider. The purple line is approximately 28.9 mph, the blue line 21.7 mph, the green line 14.5 mph, and the yellow line 7.4 mph.

In order to make the device more economically viable as a consumer product, several improvements must be made in the areas of weather-proofing, circuit streamlining, and cost reduction. Weather-proofing is critical to the long term operation of the unit. One technique considered for the motor was to encase it in a Nalgene container. These containers are known for being waterproof and nearly indestructible. (Yes, we ran over one with a car to no ill effect.) Additional protection was sought against the forces of nature. Expansion foam would seal the unit, however the material has limitations. Not only is it difficult to position properly, but it would also prevent ventilation essential to the overall operation of the device.
As to the streamlining of the circuit, possibilities include a multitasking voltage regulator chip and a custom printed circuit board (PCB). The chip could replace multiple voltage regulators, this would decrease both the product's size and heat output. Using a PCB will provide a more stable base because the connections will be directly on the board and not floating beneath it. To a limited extent it will act as a heat sink because of the copper tracing in the board. This change would decrease the need for excessive ventilation and increase component life.
Cost reduction is by far the most important, and difficult, change that must be made to the design. The circuit itself is extremely inexpensive, however the motor costs $275. A search is underway for a more cost efficient motor that will still meet our power needs.

Thanks for reading our Instructable, if you have any questions feel free to ask.

Here are some of the pictures from our presentation at MIT.