3 Phase Power Generator

Hellohello! This was a two person project worked on by me (Vincent!) and Jeffery Phan. We are both mechanical engineering students at San Jose State University and worked on this with the help of the American Society of Mechanical Engineers student section on campus. For this contest, we had an unfinished gyroscopic damper project and an old robot on hand (Specifically this one!) to scrap for parts. From these, the most useful parts we found were a DC motor and 900 kV brushless motor from the two projects respectively. Though many ideas on how to use these came to my head, the one that we landed on was a 3-phase power generator. For those unaware, 3-phase power is essentially 3 phases of AC power that are out of phase by 120 degrees. As a result, the “peaks” of the voltage waveform occur way more frequently as opposed to a single phase AC signal. This has the benefit of being much more efficient as a power source than typical AC as the voltage input will never go to zero at any point of the cycle. 3-phase power is typically used in high power requirement applications like industrial machines and buildings. For a hobbyist, 3-phase power also has the benefit in that when rectified into DC, the output is a lot more smooth as compared to rectifying single phase AC. This makes filtering AC noise a lot easier.

For our project, the 3-phase power generator works by using a DC motor powered by a power source that is coupled to a brushless AC motor. The DC motor spins the AC motor, which generates three AC voltage outputs that are out of phase with each other by 120 degrees.

Mechanical Assembly

1. Power supply
2. 900kV Brushless motor (Arthur A2812 900KV Brushless Motor)
3. DC motor (I think this one)
4. M3 Screws
5. Super Glue

Electrical Assembly

1. 1N4148 Diodes
2. 1000uF Capacitor
3. Jumper Wires
4. Prototyping board (Breadboard, perfboard, etc.)
5. Three 100 kΩ resistors

For the mechanical assembly, this was quite uncomplicated as it really is just two motor mounts and a coupler. For the coupler, it was initially designed to have a heat set insert that can thread with the threaded shaft of the brushless motor. However, we actually found that the heat of inputting a heat set insert, caused the soft 3D-printed coupler to bend. This resulted in the motor coupling to be off-center and cause a lot of vibration in the system from the DC motor going all over the place. Thus, a simple cylinder with two holes that press fit into the shafts was used instead. This is not a perfect solution, but the two motors seemed to be aligned enough. For the motor mounts, the brushless motor had screws to mount onto on the back which made mounting easy. Though unfortunately, I could not find the actual dimensions online, so I had to resort to manual measuring. As a result, one of the screw holes was not aligned. Three screws seemed good enough however, so we continued with the setup. For the DC motor, there were no obvious ways to mount it, so a clamping action with bolts was used instead. To account for the slight off-center that was 100% going to happen with a 3D printed coupler, the mount was designed so that it could be slid back and forth before supergluing to make sure the motors were as aligned as possible.

The circuit consists of a power supply which powers the DC motor. The DC motor is coupled to the 900kV brushless AC motor transmitting mechanical energy which spins the motor. The mechanical energy is converted to electricity as the AC motor spins outputting 3 voltage waves out of phase by 120 degrees from one another due to the coiling within the motor. The three voltage outputs are connected to a 3-phase voltage rectifier consisting of 6 diodes to convert AC to DC. The DC signal was very rough with many bumps, so we added a capacitor to smooth out and lower the voltage drops to clean the signal.

To test and observe voltage waves on an oscilloscope we had to make a virtual ground: Y configuration of three resistors meeting at a junction. Due to the three voltage waves being identical but out of phase with one another by 120 degrees, the sum of the three voltages at any moment sum to 0. With the Y configuration the three voltage waves join at a junction cancelling out each other creating a 0V ground reference point thus a virtual ground.

We tested the voltage relative to each other and it showed that the voltages were exactly the same. When utilizing the virtual ground we tested each output on an oscilloscope to compare the voltage waves with one another. Unfortunately we were unable to observe all three waves at once since we were limited to a single channel oscilloscope preventing the observation of the phase differences of each output. When observing the voltage wave of each output it showed one output to be higher than the others. To understand why, we tested the resistance of each 100k ohm resistor with a multimeter. The resulting resistances during electricity output showed that one resistor preserved the 100k ohm resistance while the other two were halved. This may be due to the two resistors being considered a parallel connection but the issue overall remained inconclusive.

Additionally, one interesting aspect to note is that the setup works pretty well as a tachometer. Knowing that the brushless motor is 900kV, we know that it takes about 900 RPM to generate one volt. With a Vmax at 2.30V, this suggests that the motor is running at a little over 2000 RPM with 3V input. I've made a tachometer with an IR sensor for a previous project and got similar results using this motor which I thought was cool. This idea can probably be improved upon with a better coupler.

After the inconclusive results with purely 3-phase power generator, we moved onto rectifying the output for DC power. As said before, this is pretty useful as it is much easier to smooth out the rectified output as it is already pretty smooth compared to single phase power (see picture 2! sorry for the bad glare D:). To rectify the output, six diodes were used similar to how a single phase rectifier uses four diodes. For those wondering on how this works, as diodes can only current in the forward direction, voltage is always positive forward to the diodes and is zero in the reverse direction. As a result, the voltage difference is always positive between the two sides of the diode arrangement. To smooth out the voltage, a simple 1000 uF capacitor we had on hand did the trick. From this, we lit up an LED which was cool.

Generally speaking, the system worked much more smoothly than what was expected going in. However there are a few improvements that could have been made. For one, the couple being slightly off-center did introduce a lot more vibration than necessary into the system. The best solution to this most likely is to machine a coupler or simply buying one online. Additionally, the power from the 3-phase generator is not particularly useful. This is fine as the project was mainly to be used as a demonstrator, however rectifying 3-phase power from a brushless motor does have its uses as a low voltage + high current power source. These types of power sources are very useful for applications like welding and induction heating. Personally, I would like to use one for an electrolytic cell as high current + low voltage power supplies can get pricey. To make the power output useful, both the DC motor and brushless motor would need to be replaced. For the brushless motor, a new one with a lower kV rating would be needed to have a higher voltage output. I think one of those sub-300kV motors from electric scooters would do. In addition, the whole coupling a DC motor to a brushless motor assembly would most likely need to be replaced as there are a lot of efficiency losses in converting between mechanical and electrical energy back and forth. One could potentially just get a very powerful DC motor, but using something like steam power or another energy source seems a lot more practical. Overall however, we achieved our goals for the project and learned a lot! These improvements are not on the top of my priority list currently, but I do hope that I will be able to get to them at some point. Thanks for reading!