This is a battery powered motor for a small rowing boat. The motor used is from an electric wheelchair which ran from 24V. The motor has a built in reduction gearbox to provide a large torque to the wheels on the wheelchair. This gearing ratio was far to high for the propeller. The prop needs to turn much faster in order to generate the right amount of thrust to move the boat.

Fortunately the motor used had a small part of the main rotor protruding from the opposite end to the gearbox which was sufficient to attach a sprocket. This sprocket was linked with a chain to another sprocket on the main axle. This axle was then connected to a 90 degree 2:1 gearbox from a cheap angle grinder which was then attached to the propeller.

The propeller is from an old barge and is a little larger than necessary, but the motor provides just enough power to compensate.

The control system uses a pulse width modulation system for a smooth speed control and efficiency. The circuit used is based on the power pulse modulator but with the high voltage MOSFET being replaced by an array of lower voltage high current bipolar transistors so that the huge currents demanded by the motor could be tolerated.

The other switches on the control box are for forward/stop/reverse and to choose between 12V and 24V operation. These are simply large power switches that are used to directly switch the power as necessary.

A Homemade Magnetohydrodynamic Thruster (MHDT)

Based on the image on the left, it is possible to make a simple MHD thruster, just powerful enough to propel a toy boat.

This type of thruster generates magnetic fields by passing an electric current through a liquid conductor, such as sea water. Using another magnetic field, the liquid can be pushed in a chosen direction, therefore generating thrust. You can easily make one of these devices from household materials and a couple of neodymium magnets. In the diagram below the small arrows represent the intersecting electric and magnetic fields. the large blue arrow represents the flow of water.

Two opposing inner faces of a rectangular plastic tube are covered by metal strips. These are the main electrodes and should have a connection for a battery. Magnets are attached on the outside of the tube so that they are attracting each other, and are at 90° to the electrodes.

The metal strips here are cut from a thin Aluminium sheet, but you can just use foil although it wont last as long. The electrolysis and salt water corrosion soon eats away at the metal, but foil should last just long enough to see it working. The best sort of electrode would possibly be made from carbon. A good power source for this device would be a pulse width modulated supply such as our power pulse modulator. This would allow you to adjust the frequency and width of electical pulses so that you could get the optimum thrust from your design.

To learn how a MagnetoHydroDynamic thruster works, see the Propulsion section.

This setup shows the electrodes closer together than the magnets. The combination of magnets used in this device were not strong enough to move the electrified water with any decent force. At only 1.5V in salty water, large amounts of electrolysis occurs, but the water does generally flow in one direction.

The key to getting the best performance it to have the strongest magnetic field you can get across the gap between the electrodes.

Also make sure the magnets are behind the plastic or insulated somehow. You don’t want them shorting out the electric current in the water.

The valveless pulse jet engine or pulse detonation engine is the most simple type of jet and is therefore popular among hobbyists as a DIY project. it is often referred to as a ‘tuned pipe’ because its operation depends upon making the parts the right size and shape so that it fires, or resonates at the engines natural, fundamental frequency. This type of jet propulsion does not need any type of turbine, turbofan, or propeller, making it much less complex than a typical turbojet. In a turbojet the turbine or turbofan is used to compress the fuel/air mixture in the combustion chamber so that it is more efficient and powerful.

This jet engine has absolutely no moving parts and it relies on the simple shape of the combustion chamber and exhaust for it to function. The fuel to the jet is provided at a constant rate, but it is detonated in pulses. After each explosion there remains a lower pressure area inside the combustion chamber. This is immediately filled as air rushes back in and mixes with the fuel feed ready for detonation again.

This example of a homemade jet engine is about as simple as it gets, but it could not be used for propulsion purposes because it is only safe to operate for a short time. The main body of the pulse jet engine is made from copper pipes and various adaptors. The combustion chamber is made from two copper adaptors that have been cut and soldered together. Copper is an excellent thermal conductor which helps to spread the heat throughout the jet, but solder melts very easily so if the jet engine were allowed to run for more than a few seconds this part could come apart. This was enough to demonstrate the principles of operation which is all this DIY jet engine was designed for. If you wanted a working model for providing thrust, it would be necessary to consider different materials as a running jet will get very hot.

These images show the basic ‘tuned pipe’ without the spark plug and gas supply. Tuning was achieved by altering the length and width of the parts used. This was quite simple as there are wide range of plumbing parts that will easily fit together.

The fuel was provided from a cheap blow torch and was injected into the combustion chamber using fine brass tubes bought from a local hobby shop. This chamber also contained a tiny homemade spark plug. The spark rate could be controlled by varying the power to a HV capacitor connected in parallel with the spark plug.

The power supply for the tiny spark plug was made from a mini cold cathode PSU connected to a HV diode and capacitor. An alternative is to use an ignition coil and an ignition coil driver circuit.

The spark plug its self was just a single wire inside a small glass fuse. This wire was connected to one capacitor terminal (live) and the body of the jet engine was connected to the other terminal (earth). The spark would jump from the tip of the wire to the inside of the combustion chamber to ignite the fuel mixture.

The simple design and adjustability of this jet means that a wide variety of fuels can be used. The most common fuel used is kerosene and propane, but common lighter gas will work for this basic demonstration. Click here for More information on Jet Engines.

DIY Sensor Multiplexer

This device was created to allow more analogue sensors to be added to a homemade robot. It is simply a 8 port switcher for analogue or digital signals. The device is controlled by a PIC 12F675 microcontroler, and gives eight analogue outputs from one analogue input, and one digital input.

The K8000 computer interface board, used on the robot, is limited to four analogue inputs. A multiplexer was essential in order to capture more sensor data. The multiplexer can allow the computer to capture data from several sensors using a single analogue input on the K8000.

To do this it was necessary to create a device that could switch eight digital outputs one at a time and be controlled by 1 digital input from K8000. This is very similar to ‘running lights’ where several LED’s run in sequence and start again. To keep the multiplexer small it was decided that a PIC microcontroler would be used, as this would greatly reduce the number of required components. The chosen PIC was the 12F675 because of its small size (8 pin) and the low cost of the chip programmer. This PIC has only four digital outputs so these also needed to be multiplexed. The digital outputs when low can be used like a ground connection, allowing LEDs to be connected directly across two outputs as shown below.

The diagram above shows how four outputs of a PIC can be used to control 8 LEDs. If these LEDs are inside an optoisolator then they can be used to trigger transistors, allowing a sensor to pass data to the computer. The diagram below is intended to show the connections of the sensor multiplexer. This diagram does not show some components such as resistors and capacitors, and is intended only as a guide.

To capture sensor data from the multiplexer a subroutine named ‘GetPlexer’ was created in the robots main code. This loop can read the value of the analogue input on the K8000, store it in an array, and then switch the digital output to the multiplexer on and then off again. This process will loop until all 8 sensors have been read.

When power is connected the first sensor in the array will be active. It would be as if it were connected directly to the analogue input of the K8000. When the signal from the digital output of the K8000 goes from low to high the power to the first sensor is disconnected and the next sensor is activated. This process repeats until the last sensor is activated and then it starts again from the beginning. This was housed in a plastic box and jack sockets added to allow easy connection and removal of sensors.