Renewable energy has always been interesting to me. The idea of harnessing energy from existing, nearly endless, natural processes is amazing. Our Sun puts out more energy than we could ever use every day[1]. With current photovoltaic technology, only a small fraction (<1%) of the land mass on earth would be needed to supply global electricity needs[2],[3],[4]. Adding in wind, hydroelectric, tidal and geothermal renewable power seems reasonable. Also the cost of photovoltaic cells has dropped dramatically[5]. When I first became interested in solar energy systems in high school, building your own solar panels was much cheaper than buying manufactured ones. Currently the dollar cost is nearly equal (personal knowledge). The extra effort and lower quality makes buying commercially a better option now. But I still plan to build a few panels since its been a long running goal. However, renewable energy has the issue of being intermittent. There are only a few hours of Sunlight and the wind varies. Which is where storage technologies come in: batteries or mechanical systems (gravity powered) can store the extra supply for use when the wind stops and the sun goes down. But battery technology needs to advance further, for complete transfer to renewable sources.
Around the same time I was becoming interested in solar panels, I discovered a typical permanent magnet DC motor could be used to generate power. This seemed incredible at the time. I had been experimenting with small DC motors since elementary school, but I never thought of turning the motor and producing power. I made a brief attempt at building a generator with a 12 volt motor from an RC car. This motor was low amperage; the power would be minimal unless the it was run at high RPM. However, doing so would reduce the life of the motor; between excess heat and bearings being rated for particular speeds. In this setup, I decided I would run the motor above the RPM rating. The RC car was already broken and I wanted to see what I could produce. There are two ways to produce high RPM: direct drive or a gearing system. The motor was already part of a gearbox from the car. I reused that in reverse by turning the wheel axle to turn the motor. I would be able to run the motor very fast with a lower speed input. The way this gearbox was designed it also required increased torque to use the gear box in reverse. Good for a driving a car, not so good for a generator. This removed the possibility of using wind. I decided to create a belt driven system by connecting it to an exercise machine called a bicycle trainer (at right). Even in this setup, it was very difficult to drive the motor. At that point, I decided to put the project aside. I was going to wait until I progressed further in school and had more knowledge.
in robotic applications for their ability for precise rotation control and ability to hold a position under load. Stepper motors also make good generators (and can be considered small alternators). They are typically high torque in a small form factor, provided by the multiple magnetic poles and corresponding wire coils. In terms of generators, this means they can create a relatively high power output at low rotation speed. Compared to regular DC motors with only 2 magnetic poles and one coil, a much greater rotation speed is required for the same level of power output as a stepper motor. The output is also much less smooth. With 2 poles a pulse of power comes every half revolution. In a stepper motor, a pulse will be generated every quarter turn -at least- depending on the design. The downside is the output is AC power at a high frequency, but it can easily rectified into very smooth DC power. The upside, however, is the motor can turn in either direction and the polarity of the DC power will remain the same. If you have an interest to read more, I have created a document explaining stepper motors as generators.
After graduating from college I found myself with extra free time. Since my windmill project in my first college design course I wanted to return to my work building a generator like I discussed in the previous paragraph. My original intention was to build a wind turbine, but I didn't have the expertise, funding or tools to create what I wanted. My degree provided me with a knowledge base and experience to build on. My design uses stepper motors (example below). They are integral to the design and it required I learn a lot before I used them. Stepper motors are often used
I am using three 35 volt, 2.8 amp, 2 phase, 8-lead Pacific Scientific PowermaxII stepper motors. The 8 leads are from the four stator coils. Motors with fewer leads, either have less stator coils, or the coils are connected internally. 2 phase refers to the number of independent AC circuits. In 2 phase motors, the coils always run with opposing polarity. If you were to directly connect the coils together, the output would always be zero. Before you can use a stepper motor (as a generator or otherwise) you need to know which leads correspond to which coil. Each coil works independently. In order to operate, properly the wires must be connected a certain way. My motors actually have four coils (4 phases), but in use, two of the coils are connected in series. This is possible with four coils because two will be the same polarity at any given time. Normally when a motor is purchased, a wiring diagram is provided. But mine were used and many years old. I needed to determine the wire configuration with experimentation. There is a very simple method to do this. In any given stepper motor, some leads are connected and others are not. The coil in a motor is very long and will have a measurable resistance value (roughly the same for each coil in a particular motor). An 8-lead motor is the easiest. All the wires are in pairs, each lead only has one connection to one other lead. In a 6-lead motor (with four coils), two coils are internally linked. In one of the coil sets, there is a lead connected in between the two coils (the common) and one on each end, making 3 leads. Here is where knowing the coil resistance matters. Between the common and either end the resistance is the same (this is the coil resistance). Between the two ends, the resistance valve is double (two coils). Any leads not connected will have infinite resistance.
Once the coil connections are known, the power can be rectified. A rectifier circuit uses semiconductor diodes to direct current in one direction. Turning AC into DC requires half the AC pulses get blocked. This creates pulses of current with one polarity, which is DC. There are two circuits designs to convert the AC to DC: a basic rectifier and a bridge rectifier. The first is not efficient, used more to test the output levels of the motor. Only half the power is used in every rotation of the motor. The rest is thrown away. In the second circuit, the magnitudes of the pulses are added (right, picture 1). The positives are combined and the negatives are flipped, all the power is used. Eight diodes are used, four for each phase (right, picture 2). In an 8-lead motor, 2 sets of leads must be connected to replicate the wiring in a 6-lead motor. It is possible to rectifier each coil and combine the output, but this doubles the diode count and losses. All diodes have a certain voltage drop, like any electrical component. Using a typical silicon diode, the voltage drop is 0.7V. In the rectifier circuit, this becomes 1.4V (power passes through two diodes). This is significant for a low voltage application. However, there is an alternative called a Schottky diode. A Schottky diode
is a combination, silicon and metal, diode (instead of all silicon). The voltage drop on those is only 0.2V. There is also the optional ripple capacitor (to dampen the rectified DC). This capacitor will charge an discharge with every pulse. It charges from an active pulse and discharges when the pulse ends, filling the millisecond gap in power. Depending on the application, the DC power won't be "clean" enough to use directly. This matters for electronics where "dirty" power can affect the function of a CPU or damage other board components. If the pulse frequency is low enough it will be seen as a flicker in lights, which is annoying and possibly dangerous for those with seizure disorders. But in general, the multiple phases and revolution speed puts the frequency well above 100Hz; essentially unnoticeable.
Now that I had understood the motors and could handle the power output, I was able to begin creating the physical design of the wind turbine. I was going for a basic design: low cost, boxy shape and minimum parts. This means the motors would be driven directly, no gearbox and I would use wood for a majority of parts. If anything, it would be the first prototype that I could refine in the future. I used Solidworks modeling software to allow me to play with design options before spending money on materials. The models are true scale. I measured the real motors, created a model, then built everything else around them. The models are functional as well. Not with complete accuracy, but the blades can rotate and will rotate the motors. Further, I created a motion study for each which I rendered into a video (available in the DropBox folder).
My first step was creating a box for motors. The motor box would need to fit the motors and wiring but I made it much larger. This being the first design, I thought having extra room would be beneficial for testing. I had also thought the extra weight would help balance the weight of the blades. I came up with a few designs (at left), reusing the same motor box for simplicity. One is the traditional, horizontal type, where the blades need to rotate into the wind. The other two are vertical; where the blades can catch the wind from any direction. A motor is 2 inches in height (excluding the shaft), with a length and width of 2.5 inches. The shaft has a diameter of 1/4 inches and length of 4 inches. Each motor shaft is connected to the motor in front of it using a shaft coupler. The front-most motor is then connected with the blade shaft, creating a continuous shaft. I increased the size of the blade shaft to 1/2 inches to handle the material weight and rotational forces. This required a shaft couple adapter to link the 1/4-inch shaft to the 1/2-inch shaft. The box is 17 inches long, 7.5 inches wide and 9.5 inches high and made
A horizontal style windmill requires an extra part in order to function: the slip-ring (right, picture 1). This component translates electric power from the rotating platform to the stationary structure. It does this through carbon brushes (like in a motor) that touch a cylindrical metal contact. The cylindrical contact is stationary and the brushes rotate around it. I could have purchases this item, but it seems they are fairly costly. The design of a slip-ring is simple on its own. It uses only a few parts that can be made easily. So I built my own (right, picture 2,3). The model differs from the actual because cutting a perfect circle is difficult. The shape is also irrelevant as long as it fits in the motor box and holds the components I need. The base is a 1/8-inch PVC sheet in the shape of an octagon 4 inches across (right, picture 4). I chose an octagon because it's easy to make and centering components is easier. I used a scroll saw to cut a 4-inch square. Then I calculated the length the side for a 4-inch octagon. I drew a center-line in both directions (normal to the edge) on the face of the square. I divided the length of the octagon side by 2 and made a mark of that length from the center-lines on the square. Then the marks were connected to draw a perfect octagon. After completing the cuts on the scroll saw, making the center hole was easy. I drilled a pilot hole at the center of the octagon (using the center-lines I drew) and used a 1.5-inch hole
saw to cut the middle out. The brushes I purchased are 7/16 by 1/4 by 3/4 inches (right, picture 5). They did not come with a sleeve. I cut and bent 22 AWG sheet steel into brush sleeves (right, picture 6, 7). I drew an unfolded view of the sleeve in Solidworks, without a tolerance. Then increased the dimensions to account for bend radius and fit clearance. I drew those dimensions onto the steel and cut out the shape. I do not own a sheet-metal bender so I used two pairs of pliers and a hammer to shape the metal. Securing the brush to the PVC base was a different challenge. I considered glue or a fastener, but I didn't think glue would last and the sleeve didn't allow for a fastener. I chose to cut small rectangles from the same PVC sheet to enclose the sleeve. These rectangles were glued in place (above, picture 3). The carbon brushes came with the spring and brass contact connected. The brass contact was actually wider and longer than the brush, I had to trim it to fit the sleeve. I could have only wired the sleeve body for the power connection since the brass contact would be touching the sleeve. But steel is not the best connector and I couldn't prevent oxidation on the brass or steel, I decided to solder a wire directly to the brass contact. The wire from each brush would be connected together to a single spade quick-connect terminal. Creating this link first involved soldering two brush wires to a ring terminal, plus another wire which would connect to the spade terminal. The brush wire was 18 AWG and the secondary wire is 12 AWG. I
Since my models were drawn to scale, I could use them directly to fabricate parts. Building the motor box involved creating drawing files for each board in the model. Those were printed to aid in placing the cuts on the boards. I wanted to avoid "knots" in the wood and other damage It would look better, but also make the box stronger. With the cuts marked, I used a table saw, in combination with alignment tools to make straight cuts for a perfect fit. The thicker boards (2x8-inch) were used where the greatest forces were. The base would hold all the weight. The front would be supporting most of the blade weight. The backside would support the rudder (not shown in the model). I then had to decide where to place screws to assemble the box. First, I marked the locations where the boards would overlap. Then I marked the center between the overlap and the edge of the boards (the line where the screws would go (left, picture 2,3). I still had to decide the spacing of the screws. I
from standard lumber (1x8 inch and 2x8 inch boards). The internal width was exactly 6 inches, which is important for the custom mounting plate for the motors. Those would be made from 16 AWG sheet steel. The motor has four mounting holes that the plate would bolt on to. 3/4-inch angle brackets would then be used to mount the plates inside the box. I had researched available sizes for sheet metal. A local hardware store sold 16 AWG steel in 6-inch by 18-inch sections. Since I needed three plates, I could use one sheet to make them. The mounting plate is symmetrical in both directions, making it easy to install. The bow-tie shape is to allow room for wires, above and below the motors. 16 AWG metal was chosen for the strength, availability and relative ease which it could be shaped.
All three designs use blades made from PVC in some way. The vertical designs both use 4-inch PVC pipe. The wheel paddle design uses half sections of pipe bolted to a support structure. That support structure adds significant weight, cost and fabrication complexity. A solution to that issue was the helical spiral of stacked 4-inch PVC pipe. Each level of the spiral (a blade) is a single section of pipe. One end is cut down the middle -along the length- stopping 2 inches from the middle. The procedure is repeated on the other end but the opposite side of the pipe is halved. Then a hole is drilled through the center of the pipe (where the halved sides face forward). This hole will be the outer diameter of 3/4-inch PVC pipe. After cutting multiple blades, they can be stacked on a length of 3/4-inch PVC pipe. PVC glue would be used to secure the blades to the central shaft at some spiral angle. A 1/2-inch steel shaft is then put inside the 3/4-inch PVC pipe and the two are bolted together. This steel shaft is connected to the motors. The blades for the horizontal model required additional research. My original idea was based on the blades from my college design course. The blade hub would have slots that the flat blades (PVC sheet) would fit into and then bolted in place. But any blade longer than 2 feet would wave and bend under their own weight. I had heard of PVC pipe being cut into aerodynamic blade shapes, but I had no idea how that would look. Since the internet is a wealth of knowledge, I was able to find some examples with dimensions and fabrication instructions. Once understanding the basic shape I tweaked the design (shown in green above) in Solidworks. Modeling those was a new challenge. I had little experience creating complex shapes in CAD software. I started by drawing a half (along the length) section of pipe. I then drew the projection of the blade shape above the half pipe (on a parallel plane). Solidworks has a feature called 'Extrude Cut'. This allows you to draw a 2D shape, which will cut through a solid body. The cut will either be area inside the 2D shape or all the area outside it. I used the latter. The 2D blade projection was what I wanted, so I removed everything else. The blades could then be bolted to a central hub.
There are two different designs to support the blade shaft, one for horizontal blades and one for vertical blades. In both, there is a double support for the shaft. I would use bearings pressure fitted into the wood. I wanted to avoid putting any force on the motor shaft. I knew a single wood support could not contain the forces from the blades. The horizontal blades will want to rotate downward at the exterior support. The second support will counteract this force. With the vertical blades, there is an axial force downward from the weight of the blades. The motors are not built to support weight in that way. Therefore, I added a thrust bearing on the interior support. The blade shaft would have a shaft collar which sits on the thrust bearing. The bearing would sit on a steel plate imbedded in the interior support. A vertical design would be most effective. The physical space it uses is smaller since the blades point upward and the whole system doesn't rotate. It likely capture more wind power because wind from any direction can push the blades. However, going against all those positives I am building the horizontal model. It is what I planned for so long and it looks the part.
Fabricating the motor mounting plates also involved using drawing files from the model. I printed out the shape in 1:1 scale and traced the shape onto the steel sheet (right, picture 1). All the holes and cutouts were marked as well. I cut the raw shape out using a scroll saw and a blade with the greatest tooth count I could find (25 tpi) (right, picture 2). The trapezoidal cutout was not removed. Scroll saw blades are not made for metal cutting. I had already used six blades cutting out the three plates. Additionally, the procedure for removing a cutout without starting from the outside has added difficulty. Further, the cutout was not a necessity. In my design, the cutout would be to save weight and improve internal air flow. But neither of those are a major concern for this prototype design. With the raw shape cut from the sheet, I hand filed the edges to the final shape. The next step was the holes to mount the motor. I used a drill press and a 1/16-inch drill bit to make a pilot hole at each mount hole and the central hole for the shaft. The final size for the central hole was 3/8 inches and the mount holes were 3/16 inches (the same as on the motor) (right, picture 3,4). I made the central hole much larger than the shaft to provide clearance in the event the hole wasn't perfectly centered. The next step was the holes for the angle brackets (which mount the plates to the motor box). These were not included in the model because I had not decided on the exact method, and I would be
In the background information, I mentioned I would be using a bridge rectifier to convert the AC output from the motor to DC. I wanted a professional-looking solution that would have a long life and would be easy to connect or remove. Therefore, I created my own circuit board (at right). Many years ago I purchased several sizes of prototyping circuit boards. In this case I wanted to make a board are small as possible. I started by drawing the components and their connections on a paper to get an idea of how I could create a circuit board. The diodes I would be using would require the most space. However, I realized by inserting the diodes vertically, instead of laying them horizontally over the board I could reduce the space they used by a third. The board is half of the largest board I had. I found the components would fit exactly. The diodes are 1N5822 Schottky diodes which can handle 3 amps and 40 volts. Since there are two groups of diodes, each set only sees the output of two coils, which is half the total output, putting the load on the diodes below their limits. The leads from the motor are 0.1-inch spacing pin contacts, a common type. It is very easy to purchase connectors to fit those leads. I wanted a specific connector cable, female to female, that went from 8x1 (motor) to 4x2 (board). A 4x2 male connector was used on the board because
As I mentioned in the design the blade shaft is larger than the motor shaft. The motors came with shaft couplers for 1/4-inch shafts (left, picture 1). Initially I looked up existing solutions for coupling different sized shafts, but they were surprisingly expensive for my budget. It also didn't have the size I needed. I would be required to modify it anyway, so I decided I would make my own (left, picture 2,3,4). I used the dimensions of the commercial product to choose the stock material I would use to fabricate the part. The commercial product I was replicating was aluminum with a diameter of 1-7/8 inches. I felt nearly 2 inches across was much bigger than necessary. The 1/2-inch shaft was the limiting factor. There would need to be enough material to handle the rotational forces. Therefore, I bought a 7/8-inch square bar of 6061 aluminum. The price for a 2-foot bar was half the price of a single commercial couple. I would only need 1.5 inches to make a shaft couple. I chose square over round because I would need to mark the exact center of the cross section to drill the shaft holes. It is very easy to mark the center of a square. I cut length of bar I needed for the couple with a handsaw. This left the surface very uneven. I used a hand file to smooth and level the surface as much as possible. I marked the center on both sides of the block. To aid the accuracy of drilling the holes I used an adjustable vise
purchasing angle brackets. I could not determine the dimensions until I bought them. Having the physical plates made deciding on the particular bracket much easier. I chose 3/4-inch steel angle brackets. There would be two on each side of the plate, four in total (above, picture 5). The angle bracket is bolted to the plate and wood screws are used to attach the plates to the box. These angle brackets would have worked for the original plate as well (with the cutout removed); another reason I chose those them. I placed each bracket 3/4 of an inch from the corner of the plate. The bracket needed to be placed exactly at the edge of the plate (the side that would touch the box). The internal width of the box would not allow for anything wider than 6 inches. I used a bracket held against the plate, with both held against a very flat surface, to mark the holes. This allowed me to make sure the bracket did not extend over the end of the plate.
started by marking the first screw to start 1 inch from the edge (still on the center-line). A screw too close to the edge in wood will cause it to split. I marked another screw location at the other end, 1 inch from that edge. I desired the screws in between be evenly spaced. The distance in between could be divided by a certain number (length) which would give a whole number answer for screw placement. This number varied depending on the board size. On smaller ones, the distance only allowed one screw (which was centered on the board). For longer boards I needed a balance between adequate support and too many screws. On connections between the face of a 1-inch thick board and the edge of a 2-inch thick board I used fewer screws. The thicker board would be able handle screws driven very tightly. For connections between the face of a 1-inch board and a 1-inch edge, I added more screws. The thinner wood would split if the screws were made very tight. In order to have the same connection strength, the screw spacing was smaller so more screws could use a lesser tightness. After marking all the screw holes I drilled them out as accurately as possible so the boards would align correctly when fastened together.
mounted on a drill press (above, picture 5). I used a 1/16-inch bit to create a starter hole. Then with a 1/8-inch bit I drilled all the way through the block. A 1/4-inch bit was used to widen the hole. Only half the block needed to fit the 1/2-inch shaft. I marked the depth on a 1/2-inch drill bit and drilled the hole. On the 1/2-inch shaft I could drill through the shaft and the couple then use a bolt to secure the shaft (left, picture 3). But the 1/4-inch shaft was too small to drill. I decided to replicate the compression fitting of regular shaft couples. This required I cut the couple at the 1/4 - 1/2 interface and then split the side of the couple on the 1/4-inch section. I drilled two holes near the edge of the couple on the 1/4-inch section (above, picture 4). I also drilled the surface to fit the bolt head to make it flush. On the opposing side, I performed a similar operation to imbed the nuts. By tightening the bolts, the couple would flex slightly and grip the shaft.
used four brushes to minimize the chance of losing a connection when the motor box is rotating. They also share the power load, but any one brush can carry the full output. The maximum combined current from the motors would be 9 amps (the motors are wired in parallel) and the voltage will peak at 35 volts. 18 AWG wire can easily handle 9 amps at 35 volts, and 12 AWG wire can handle even more. The other section of the of the slip-ring (the stationary tube) (left, picture 1) has a much shorter procedure. In my search for materials to replicate my model, I discovered the outer diameter of 1/2-inch PVC is almost exactly the size of the inner diameter of a 3/4-inch copper pipe couple. The main structure is a length of 1/2-inch PVC pipe. The electrical contacts are from the 3/4-inch copper pipe couple (left, picture 2). Each piece is a 5/8 inches long, which I cut with a scroll saw. The separator for the two contacts is a 1/4-inch piece of 3/4-inch PVC pipe. The inner diameter is just slightly smaller than the outer diameter of the 1/2-inch PVC. I had to grind down the inside of the 3/4-inch PVC. A 12 AWG wire was soldered to the inside of each of the copper contacts (left, picture 3). I wanted to limit the added thickness Then a slot was cut in the 1/2-inch PVC pipe to fit the wire and the extra thickness of the wire on the contact (left, picture 4). The lengths for the copper contacts and PVC separator are such that the brushes will touch the middle of each contact when installed.
it was a smaller form factor and to make it easier to wire the various pins to the diodes. No such cables existed, so I bought the components to create my own. I have built with connectors of this style before; I had no trouble making these. I would use the same style pins on both ends for consistency and reduced cost (I could use the same internal contacts for the female connectors). I already had the wire (stranded 22 AWG), which is capable of handling the power. Each wire would see no more than 1 amp or 17 volts. I used eight different colors to match each lead from the motor. The proper function of the rectifier depends on each wire from the motor being connected correctly. I built three identical cables (right, picture 4). Each connection requires crimping and soldering a connector contact to a wire then sliding it into the connector. The holes on the board are designed to fit through-hole components like the 4x2 pin connector. But the diodes and spade terminal are much bigger. I had to drill out the holes where I wanted those components to go. The spade terminals are for the DC output. I soldered the diodes in first. All the other connections would be made to those. The next component was the pin header. Since the board only had holes, with no connection between them, I used short lengths of wire to connect the pins to the diodes. The final components were the spade terminals. I added extra solder around those because the spade connectors require a fair amount of force to attach and remove. I would wire each side of the output to single connector that connects to the slip-ring. In this way, the motor outputs are in parallel, increasing the amperage at the same voltage. Each motor has nearly the same output voltage so very little power is lost between them from differing voltages.
My wind turbine is not fully complete yet. The progress has stalled partially due to lack of funds to buy the PVC pipe for the blades. To create the proper curvature of the blade the PVC needs to be either 6 or 8-inch pipe. A 10-foot section is over $100. I never considered larger PVC pipe would be so expensive. All the sizes smaller are less than $20. The other issue is the shaft adapter couple. The one I have created is too flawed to function. On the 1/4-inch side, the hole is just slightly larger than the bolts can squeeze. In other words, the coupler cannot grip the motor shaft. I require more precise tools to fabricate a new part. The tool I believe I need is a band saw. It can create the precise, straight cuts necessary for the part to function better. This one flawed part also prevents me from installing the motors in the motor box and assembling the box. The alignment of all the shafts is very critical for the best function of the turbine. The best way to align the blade shaft with the motor is to connect them together before fixing the motors in place.
The size of this turbine limits the usefulness, but it was my intention to learn rather than build a substantially powerful system. The theoretical maximum output would be 150 watts. Not exactly small but not huge. The wind is also not very consistent. So I expect I would use the turbine to charge batteries which I could then use to power something like a small electronic device. This is why I have chosen to wire the motors in parallel. In a low wind speed I will be likely to produce half the voltage capacity (about 15V). That is enough to charge a 12 volt battery. Having the outputs in parallel there will be increased current output. Current is the factor that affects the charging speed in batteries.