Supercapacitor technology often looks like a revolutionary energy storage technology on the surface, but the actual performance numbers can be rather uninspiring. However, for rapid and repeated charge and discharge cycles, supercaps are hard to beat. [Tom Stanton] wanted to see if supercaps have any practical use on e-bikes, and built a DIY electric motor in the process.
One of the problems with supercaps is the rapid voltage drop during discharge compared to batteries, which can limit the amount of usable energy. In an attempt to get around the voltage limitation, [Tom] built his own axial flux motor for the bike, using 3D printed formers for the coils and an aluminum rotor with embedded magnets. He expected torque to be severely limited, so he also machined a large sprocket for the rear wheel. He built a capacitor bank using six 2.7V 400F supercaps, only equivalent to the capacity of a single AA cell. Although it worked, the total range was only around 100 m at low speed. When he hooked the motor up to a conventional battery, he did find that it was quite usable, if a bit underpowered.
The controller for the DIY motor was not capable of doing regenerative braking, so he fitted the capacitors to another e-bike that does have regenerative braking. Using this feature, he was able to reclaim some power while slowing down or going downhill. Since this type of charging cycling is what supercaps are suited for, it worked, but not nearly to the level of being practical.
[Tom Stanton] has been messing around with compressed air power for a few years now, and most of his work focused on piston engines. He likes using 2-liter soda bottles as lightweight tanks but their capacity is limited, so the nozzle can be a maximum of 1 mm in diameter if he wants to produce thrust for 30 seconds or longer using a turbine. Pelton turbines have been in use for a long time, especially for hydroelectric systems, and they use small diameter nozzles, so he decided to experiment with a pneumatic Pelton turbine. (Video, embedded below.)
Pelton wheels are water wheels with specially designed buckets to efficiently extract energy from a high-velocity jet of water. [Tom] 3D printed several geared Pelton turbines and started doing bench tests with a propeller and a load cell to gather empirical data. With the help of high-speed video of the tests, he quickly realized that the turbine efficiency is highly dependent on the load. If the load is too small or too large, the moving air will not come to a complete standstill, and energy will be wasted. [Tom] also suspected that some moving air was escaping from the bucket, so he created a version that enclosed the buckets with a ring on the outer perimeter, which increased the peak thrust output by 65%. Compared to his diaphragm air engine design, the peak thrust is higher, but the overall efficiency is less. [Tom] believes there is still room for improvement, so he plans to continue working on the Pelton turbine concept, with the hopes of building an air-powered model helicopter that can lift off. Continue reading “Pelton Turbine Development For An Air Powered Model Helicopter”→
One of the tricky parts of engineering in the physical world is making machines work with the available resources and manufacturing technologies. [Tom Stanton] has designed and made a couple of air-powered 3D printed engines but always struggled with the problem of air leaking past the 3D-printed pistons. Instead of trying to make an air-tight piston, he added a rubber membrane and a clever valve system to create a diaphragm air engine.
A round rubber diaphragm with a hole in the center creates a seal with the piston at the top of its stroke. A brass sleeve and pin protrude through the diaphragm, and the sleeve seals create a plug with an o-ring, while the pin pushes open a ball which acts as the inlet valve to pressurize an intermediate chamber. As the piston retracts, the ball closes the inlet valve, the outlet valve of the intermediate chamber is opened, forcing the diaphragm to push against the piston. The seal between the piston and diaphragm holds until the piston reaches its bottom position, where the pressurized air is vented past the piston and out through the gearbox. For full details see the video after the break.
It took a few iterations to get the engine to run. The volume of the intermediate chamber had to increase and [Tom] had to try a few different combinations of the sleeve and pin lengths to get the inlet timing right. Since he wanted to use the motor on a plane, he compared the thrust of the latest design with that of the previous version. The latest design improved efficiency by 366%. We look forward to seeing it fly! Continue reading “Diaphragm Air Engine”→
[Tom Stanton] has been playing Microsoft Flight Simulator a lot recently, and decided his old desktop joystick needed an upgrade. Instead of just replacing it with a newer commercial model, he built a complete controller system with a long joystick that pivots at floor level, integrated rudder pedals and a throttle box. You can see it in action after the break.
The throw of the joystick is limited by [Tom]’s legs and chair, with only 12° of travel in either axis, which is too small to allow for high resolution with a potentiometer. Instead, he used hall effect sensors and a square magnet for each axis, which gives good resolution over a small throw angle. The pivot that couples the two rudder pedals also makes use of a hall effect sensor, but needs more travel. To increase the size of the magnetic field, [Tom] mounted two magnets on either side of the sensor with their poles aligned. To center the rudder pedals and joystick, a couple of long tension springs were added.
A normal potentiometer was used in the throttle lever, and [Tom] also added a number of additional toggle switches and buttons for custom functions. The frame of the system is built with T-slot extrusions, so components can quickly moved to fit a specific user, and adjust the preload on the centering springs. All the electronic components are wired to an Arduino Micro, and thanks to a joystick library, the code is very simple.
At a total build cost of £212/$275 it’s certainly not what anyone would call cheap, but it’s less than what you’d pay for a commercial offering. All the design files and build details are linked in the second video if you want to build your own.