A buck boost converter is a simple device that employs the use of semiconductor switching devices, inductors, and capacitors, to either increase or reduce the current flow to maintain voltage at the output. Another important aspect of these converters is the power in equals the power out. Below is a schematic of the initial buck boost converter

Schematic 4. First design for the buck boost converter.

Figure 5. Voltage and current graphs for a few of the components.

Graph 3. Output waveform from the idea model. This output is for an input of 14 volts and a 50 percent duty ratio which with 2 volts of losses outputs at around 12 volts. This model looked very nice so the EE team went forward with constructing a circuit of this idea model. Unfortunately, no photos were taken of this circuit’s output but below are some photos of the circuit and the output of the rectifier.

Photo 1. Circuit built from following the ideal model design. Frequency generation device show an input waveform of 1Hz, which is as close to the 0.4 Hz waveform that this device could go. The oscilloscope shows the output of the rectifier before low pass filtration.

Photo 2. Circuit built from following the ideal model design. The oscilloscope shows the output of the rectifier after low pass filtration. The LPF has nicely formed a DC line.

The actual output was not nearly what it should have been. With a 50% duty cycle the output was nowhere near the input voltage level. With an input of 4 or more volts the output would never read higher than 0.7 volts at the greatest. After a very long debugging process it was determined that issue was stemming from the non-ideal nature of the real world transistor. The capacitance levels in the transistor we used, though very small, were creating discharges that greatly affected the waveform across the inductor. What occurred was a severe rippling effect every time the transistor turned on and the inductor began charging. This rippling affects lead to severe losses in the circuit dragging all of the power away from the output. Below is a similar representation found online.

Photo 3. Photo of inductor voltage ripple closely resembling the problem in the

physical circuit that was built according to the ideal model. The above photo shows a similar representation but not quite the same. The

difference was the ripple was occurring during the on cycle for the positively offset portion of the waveform. This issue was replicated with a new non-ideal circuit model before that model was than improved to create a working real world design. The schematic for this is shown below.

Schematic 5. Finalized real world design for the Buck boost converter

This design is very elaborate and took quite a long time to get right. This circuit includes full second order input filtration with damping, secondary input filtration, primary and secondary output filtration, and negative PID regulated feedback. The circuit itself has been added to the drop box and is available for running to see all of the individual scopes and waveforms. Below is the inductor voltage waveform and output voltage waveform.

Graph 4. Inductor voltage waveform of real world model

Graph 5. Output voltage waveform of real world model The inductor wave form still does not always look exactly like the ideal models of

the undergrad power electronics course depict but in doing research it seems the waveform often do not actually form perfect square waves. The waveform will change for different duty ratios. For a 50% duty ratio it actually does form a wave very close to a square wave with just a little over shoot on the beginning of the charging cycle. For other duty ratios it looks more like variations of the above waveform but the slowly oscillating form does transfer power to the output much more effectively. The distortion in the waveform is removed by the 2nd order primary filtration right before the output resistor. Graph 5 shows the output waveform very nicely staying within a few millivolts of the desired 12-volt output. The spikes are from differential discontinuities between the switching of the transistor and the charging/discharging of the inductor. They are extremely high frequency, remain stable, and do not even spike outside of the charging range for a 12v battery, which is typically between 11.5 and 12.7 volts so they would not cause any significant issues with the charging of the battery.

The duty ratio does not linearly adjust the output voltage, as it never would even

in an ideal model. Holding the duty ratio constant and linearly adjusting the input voltage does not have a linear effect at the output but in discussions with experienced grad students with these kinds of designs it appears that such behavior is very untypical of real world converters.

The Real World Test of the Generator

The real world test was performed by the ME students. No pictures were taken but information was conveyed that the output was as high as 60 volts and 1.2 amps.

This is an incredible 72-Watts created by the generator following the EE calculations that blew away all expectations for the power generated by this device. Below is a photo of the full-scale buoy.

Real World test photo 1.

This is the picture provided from the ME side who put the structure of the buoy

together. The only things missing are the center tube that holding the magnets and that the floating part of the buoy move up and down on as well as the coil of wires that sits on top of the center dark ring in the photo. The coils are enclosed in a capsule that attaches to this ring.

Real World Test photo 2. Pool test of linear generator. Structure held together well under testing. Conclusion To conclude specifically about the EE side of the project for the EE report, through out this year the EE side of the team made significant improvements in the power generation and electronics of the buoy. Better more precise calculations for the specific power generation involved with this technique lead to extremely significant improvements on the power output by better allocating and positioning the wire and

magnets. Significant improvements were also made in the power conversion and storage areas of this design. Advanced real world power conversion circuitry, normally a graduate level class subject was pioneered by the EE side for use in an area that has been barely tapped by modern engineering. The great difficulties with overcoming conversion of such a low frequency input waveform to such a precise output that is actually capable of charging a battery (Batteries require very precise voltage levels to keep charging) was overcome. Our legacy to the next team will be an amazingly strong 60 plus Watt wave power generator and a ready to go design for the power electronics with most of the parts already ordered. The next team will just need to follow this design and put it together with in the first few weeks of the year before considering the next aspects they desire to add on to the project. Project Timeline, Deliverables and Milestones

Project Deliverables

1.) Finalized Design of Linear Generator 2.) Small scale test model 3.) Large scale Buoy 4.) Finalized circuitry design

Project Milestones

1.) Hardware assembly and test run from ideal model 2.) Debugging of the ideal model to draw conclusions on how to make the real

model 3.) Fully functional design for the real world power electronics system

4.) Effective stable rectifier for the very low frequency input wave form before passing into the smaller more precise converter circuitry components

5.) Constructing full scale model of Linear Generator Project Collaborators: University of Connecticut Electrical Engineering

Charles Conway o Senior Design Team Member o University of Connecticut Electrical Engineering Major o charles.conway@uconn.edu

Patrick Vicente o Senior Design Team Member o University of Connecticut Electrical Engineering Major o patrick.vicente@uconn.edu

Dr. Peng Zhang o Faculty Advisor o peng@engr.uconn.edu

Dr. Shengli Zhou o Faculty Advisor o shengli@engr.uconn.edu

Taofeek Orekan o Graduate Advisor o taofeek.orekan@uconn.edu

University of Connecticut Mechanical Engineering

Steve Haldezos o Senior Design Team Member o University of Connecticut Mechanical Engineering Major o steve.haldezos@uconn.edu

Patrick Kalagher o Senior Design Team Member o University of Connecticut Mechanical Engineering Major o patrick.kalagher@uconn.edu

Dr. Bryan Weber o Faculty Advisor o bryan.weber@uconn.edu

Additional Acknowledgements

Joshua Ivaldi o University of Connecticut Electrical Engineering Graduate Student o Helped with the programming of the ATmega168 microprocessor and implementation of circuitry to best protect the processor and transfer its voltage effectively to drive the transistor. o University of Connecticut Mechanical Engineering Major o steve.haldezos@uconn.edu

Appendix A Experimental results

When performing the experiments, linear velocity was held at about .14 by making strides in each direction of about 2cm and keeping the frequency (as can be seen on the oscilloscope) at about 3.5Hz. This means the magnets traveled a total of 4 cm during each period to give the .14m/s speed. The following are the results of the program for several different values, which the small scale model closely followed. Videos of the small scale tests corresponding to these outputs for voltage and speed based on frequency and a ruler showing distance are available on the website under the “Project Documents” tab. The results were strikingly similar to the program.

Matlab program ‘“linm” output for experiment in video 1 Linear motor design program is assumed to use 22 gauge aluminum wire This program also assumes an average wave height of 1.25 meters Strength of magnetic field (Tesla's): .0004 Radius of coils (meters): .0155 Number of turns of the coils in the z direction: 20 Number of coils for each row of turns (overlaps of wire): 1 Linear velocity of generator (meters per second): .14 The generator produces 1.140224e-02 Volts At 2.909775e-03 Amps For a power generation of 1.327118e-04 Watts And delivers 4.883795e-05 watt seconds per wave This uses 1.030400e-02 percent of the total seconds to capture energy in each wave Length of wire used is 1.947787e+00 meters Matlab program ‘“linm” output for experiment in video 2

Linear motor design program is assumed to use 22 gauge aluminum wire This program also assumes an average wave height of 1.25 meters Strength of magnetic field (Tesla's): .0004 Radius of coils (meters): .0155 Number of turns of the coils in the z direction: 20 Number of coils for each row of turns (overlaps of wire): 2 Linear velocity of generator (meters per second): .07 The generator produces 2.234963e-02 Volts At 2.793702e-03 Amps For a power generation of 2.497528e-04 Watts And delivers 9.190904e-05 watt seconds per wave This uses 1.030400e-02 percent of the total seconds to capture energy in each wave Length of wire used is 3.976502e+00 meters Matlab program ‘“linm” output for experiment in video 3 Linear motor design program is assumed to use 22 gauge aluminum wire This program also assumes an average wave height of 1.25 meters Strength of magnetic field (Tesla's): .0003 Radius of coils (meters): .0155 Number of turns of the coils in the z direction: 20 Number of coils for each row of turns (overlaps of wire): 2 Linear velocity of generator (meters per second): .07 The generator produces 1.676222e-02 Volts At 2.095276e-03 Amps For a power generation of 1.404860e-04 Watts And delivers 5.169883e-05 watt seconds per wave This uses 1.030400e-02 percent of the total seconds to capture energy in each wave Length of wire used is 3.976502e+00 meters