August 19, 2015


The challenges of working on a project that was begun by someone else and has been on hold for five years include restoring items that have deteriorated over time.

For instance the rotors that had already been assembled had become somewhat rusted and the magnets were covered in metal dust and filings.

Another challenge is in knowing what work has already been done. Unless the previous person kept very detailed records of their work, the new person won't always know what they did or why they did something a certain way.

After learning what I could of what had been done and restoring the parts that had deteriorated, I had on hand one cast rotor, one uncast rotor, and a wound but uncast stator. Their construction, along with the basic workings of the axial flux alternator have been covered in previous articles:

  • Axial Flux Alternator: exploring an evolving technology by Walt, March 2008

  • Assemblying the Axial-Flux Alternator: Machining Rotors and Linking to the Steam Engine by Oana, October 2008

  • Axial-Flux Alternator Update: The first stator and stator mold by Oana, 2010
  • The Rotors

    Casting the rotor in fiberglass resin.

    The resin cracking less than a day after it was removed from the mold.

    Before casting the second rotor, I cleaned it thoroughly. The dirt came off using compressed air, but the metal dust was a stickier problem.

    Some of it could be cleared off by using your finger to collect it at one corner of the magnet and lift it off.

    For the rest, I used the tip of a thin metal rod, which once magnetized by the magnets on the rotor would lift the metal dust off of them.

    In casting the uncast rotor, I made the mistake of not adding a sheet of fiberglass over the magnets. At the time it didn't seem necessary, but lacking that support the resin started to crack around the edges of the magnets almost as soon as it was dry.

    For now, I covered the cracks with lacquer to provide the magnets with protection against corrosion. If and when this alternator will be in continuous operation, a sturdier solution might need to be found, but for the purposes of research it should do fine.

    The sanding board with a partially sanded rotor.

    A wooden platform to hold the rotor in place for sanding.

    To get the rotors as close to the stator as possible, I needed their surface to be as flat as I could make them. I used a sanding board to sand away excess resin and make the surface of the resin on the rotors flat.

    As in many of the considerations involved in this project, there are two competing issues here.

    The thinner the layer of resin over the magnets, the closer together the two rotors can be placed, and the stronger the resulting magnetic field. But if the resin layer is too thin, then even the resin on first rotor cracked despite the reinforcement of a fiberglass sheet.

    Since it is necessary to handle metal tools and bolts around the rotors, no matter how much care one takes some will impact the magnets, and the resin will crack or even break off if it's too thin.

    Another limitation on how thin to make the resin over the magnets is that if they become thin enough, the resin starts to bend outwards, as if an air bubble was forming between the resin and the magnet.

    The Stator

    The stator sealed in its mold. Tape covers the hole used to release air bubbles.

    The stator was built before I came to Windward, and its technical details are in the previous articles. It was missing one of the terminals, I used a brass knurled nut for the last terminal, and soldered the last pair of copper wires to it.

    The next step was to cast the stator. When inserting the stator into the previously prepared mold, I found that the coils made a circle with a diameter of 14".

    Since the diameter of our rotors is 12", in this configuration a significant portion of the copper copper coils would not be directly between the steel plates, meaning the magnetic field across those parts of the coils would be weaker than the field directly between the steel plates.

    The cast stator ready to be removed from the mold.

    To fit all the coils within a 12" circle, I modified the mold, reducing the central wooden disc's diameter from a 6" to 4 3/4". This solved that problem, but as you will soon see, created another.

    To hold the terminal buttons in place during the casting, I drilled 3 holes in the mold, where screws would hold the terminal buttons flush against the bottom of the mold.

    I also drilled a 2'' hole at the top of the mold to release air bubbles and add more resin during the casting.

    With the mold adjusted, it was time to pour the resin. I based the casting off Wind2Volts YouTube tutorial (see online resources below for links).

    This project is the first time I poured resin, so casting the stator was the first opportunity I had to apply the lessons I learned when I cast the rotor. When I cast the rotor the resin filled the slots in the screw heads, making it very difficult to unscrew them.

    This time I covered the slots with small pieces of tape. I likewise covered the exposed ends of the terminal, so that resin wouldn't fill the hole in their center.

    After the cracking resin on the rotor I cast, I made sure to use fiberglass sheets on both the top and bottom of the stator. This produced much better results, and the resin has not cracked even a little in the weeks since the stator was cast.

    Of course, there were further lessons to be learned. Despite my attempt to tap out the air bubbles as shown in Wind2Volts video, quite a few remained, especially trapped under the mold's central disc. For next time, perhaps it would help to rotate the mold while tapping out the air bubbles.

    Since the stator won't be moving, it will be under far less stress than the rotors. So most of the air bubbles shouldn't be a problem. Some, however, are large enough that the copper wire is exposed. Those I filled with more resin or covered with epoxy.

    Putting it all together

    The rotor assembly mounted on a shaft connected to the steam engine.

    Our alternator will be mounted on a shaft that sits beside the steam engine on a wheeled table. With the mount for the shaft already built, I mounted the steam engine on wooden supports to bring the steam engine's shaft in alignment with the alternator's.

    I added a layer of rubber to the mounts to allow the steam engine some flexibility, and bolted everything to the table.

    The stator supported by two of its L braces.

    The collars that attach the rotors to the shaft and the bearings both grip the shaft with set screws. When a set screw is tightened against the shaft, it digs into the steel and creates protrusions that make it difficult to remove the bearings from the shaft.

    To get around this, I sanded flat the areas on the shaft where the set screws will rest.

    The aforementioned bolts, along with the two small hex head bolts that hold the collar to the stator.

    Since I want the rotors to spin as close to the stator as possible, I needed to secure the stator to the table such that it wouldn't sway into contact with the rotors when the steam engine starts vibrating.

    I cut out pieces of 3/4" steel square tubes, and Walt welded them together to make L braces that will be bolted to the table and the stator and hold it steady.

    The rotors are held apart by four bolts which start in threaded holes on the steel plate of one rotor, pass through the central hole on the stator, and rest against the collar of the other rotor.

    Given their strong attraction to one another, it isn't really necessary to hold them together, but two bolts go through both steel plates as an extra layer of protection.

    The heads of the two small hex head bolts were also meant to fit into the central hole on the stator. However, since I reduced the diameter of that hole to 4 3/4", the bolt heads no longer fit. Since it is no longer possible to make the hole bigger, those protruding bolt heads must be eliminated.

    Drilling the countersinks in one of the collars.

    To eliminate the protruding bolt heads we used flat head (or countersunk) bolts. Since the angle of the countersink is 82 degrees, and standard drill bit's point is sloped at 118 degrees, we sent off for a 82 degree countersink.

    Once it arrived we drilled the countersinks so that the tops of the flat head bolts sat flush against the surface of the collar.

    Because the table the alternator is mounted on has a rubber surface over the metal, bolting the stator to the table caused it to sink into the rubber, which caused the stator's central hole to be just a little to low for the bolts that hold the rotors apart to pass through it.

    I added 1/8'' thick pieces of tire rubber under the stator and its L braces, and that got it to the appropriate height.

    Aligning the rotors is tedious, but not difficult. To minimize the distance between them, it's important to align the rotors such that their surfaces are parallel to the stator.

    Even when the set screws are tightened, the rotors can tilt slightly in relation to the shaft. By tightening or loosening some of the 4 bolts that hold the rotors apart, the rotors can be made parallel to the stator.

    The countersunk bolts compared to the original bolts.

    With everything lined up, it was time to give the alternator its first test run. Each terminal is wired to a bridge rectifier, and those are wired to a 12V headlight. I brought the speed up gradually so that the alternator wouldn't generate too much voltage and burn out the headlight. If it were to pass 12 Volts, more headlights could be wired in series with the first one.

    Wiring diagram with simplified alternator wiring.

    The good news is that mechanically everything holds together well. The rotors and the stator don't wobble despite the steam engine causing the table to shake.

    Since the rotors in the final configuration will be less than 1/8'' from the stator, even a slight wobble would cause them to brush up against each other.

    The bad news is that the alternator did not generate any electricity, nor did I expect that it would. After the stator was cast, it no longer conducted electricity between the terminals.

    The stator has three terminals through which electricity flows. As the stator is wired in a delta configuration, each terminal is connected to each other terminal.

    The completed alternator. Protruding from under the flywheel is our homemade RPM sensor (an article on its construction will be online shortly).

    Before casting the stator, I used a multimeter to check the connections between each pair of terminals, and all showed a closed circuit. After the casting, however, when I performed the same test (along with several others), no pair of terminals showed a closed circuit.

    For now, I'm trying to figure out what went wrong and if it can be fixed. I may discover that this is a fixable problem, or an entirely new stator may have to be built.

    Such is the cost of building something complicated for the first time when you are simultaneously learning about many of the processes involved in such an undertaking.

    Helpful online resources about Axial Flux Alternators: - various calculators for magnetic forces. They're also a good source to email for questions about neodymium magnets. - an explanation of the equations governing alternators - specifications of the bridge rectifiers we're using.

    Wind2Volts tutorial videos: - building the stator - casting the stator - building the rotors