Axial Flux "Pancake" Motor

As part of our school project, we had to design and build a generator that would generate net positive power in a wind tunnel. It's a design and build from scratch. Our goal was to make one of the biggest and baddest generator for the project :) There's a lot of room for improvement for this design and hopefully, in the future, I can make improvements to this one. This project is pretty complicated and time-consuming, and requires quite a lot of learning to do.

As a note, this is a guideline of how we made our project. There’s definitely plenty of room for improvement! The results you may get following this may be better or worse than ours!

This instructables is submitted for the Magnet Contest.

Theory

Our generator design was a permanent magnet, brushless, axial flux direct current generator. It had 15 slots/coils and 16 North-South poles. The permanent magnets we used were 16 x “wedge” (arc segment) N42 magnets, rated for a surface field of 13,200 Gauss with a 39.89822475mm (1.57079625in outer arc and a 19.9491122mm (0.78539812in) inner arc each. Combined, they form a ring with an outer diameter of 203.2mm (8in) and an inner diameter of 101.6mm (4in); each segment represents 22.5˚, and they are 6.35mm (0.25in) thick, magnetized through the thickness. These magnets represented the North and South poles of our magnetic field.

In the axial flux design, the magnets and stator are arranged with the poles parallel to the shafting, and the flux directed axially (hence the name). The off-set in the number of poles P=16 and the number of coils C=15 creates three distinct phase peaks, A, B, and C, separated mechanically and electrically by 120˚.

While we had intended from the beginning to use steel for both the rotor and the stator, we simulated the interactions of the magnets and the steel to attempt to predict the effect on the magnetic field magnitude and direction inside our rotor. We used Finite Element Method Magnetics (FEMM) to simulate the effect of steel and placement of magnets, as well as distance between rotor magnets and stator. The images generated using FEMM are in two-dimensions, so they represent a side view of our wedge magnets if viewed from the outside diameter, which will appear as a rectangle; for reference, see the figure below, which has an angle on it aid visualization.

28 AWG Magnet Wire

N42 22.5˚ Arc Magnets; 13,200 Gauss https://www.magnet4sale.com/n42-wedge-neodymium-m...

16 gauge G90 high carbon steel, 9” diameter; spiral cut to disrupt eddy currents

16 gauge G90 high carbon steel 2

Bearings 0.25” ID, 0.50” OD

1 Shaft 0.25” x 9”

18 Bolt Sets 10-24 x 2” bolts plus washer and nuts

8 Hex Screws 12 x 1-1/2”

2 Rails 80/20 Aluminum T-Slot Rails, 6” L

Epoxy

Rectifier

As for the tools:

Plasma Cutter

3D Printer

Hand Drill

Soldering Iron Kit

Most of our supplies and tools were purchased on Amazon or we had available on hand.

We first bought the wedge-shaped magnets, intending to use them as the basis for our design. We constructed jigs for wire-wrapping based on the size of the magnets using Solidworks. We had an assortment of sizes that we made so we could test later on the voltage that was produced with the different sizes.

Coils were sized based on magnet size, stator size, and required number, then wound on a break apart jig, which was mounted onto a hand drill. We wound a single test coil very carefully, counting to 320 turns, then used that coil’s mass and measured resistance to size additional coils. This process was arrived at after making multiple coils which were irregular in mass and resistance, although supposedly careful turn counting had been observed. The coils were shaped on the jig, taped together, then removed. We planned to have 5 coils per phase, so we would end up with a total of 15 coils. This was probably one of the most tedious parts of the project, counting exactly 320 turns per coil.

We constructed our metal parts, because we had immediate access to the CNC plasma cutting bed, and experience with drafting parts in the DesignEdge software package. We cut several parts out with 9-inch diameters. Part of the reason why we chose to utilize steel was to increase the permittivity of material, therefore theoretically increasing the voltage results. Also, it looks cool.

We spent a large majority of time learning Solidworks and 3D printing parts for our project.

Final assembly of the rotor was simple; the shaft was epoxied to the collar, and the collar bolted to the rotor, and the magnets epoxied to the rotor.

Final assembly of the stator was much more involved. Mounting the coils was done with epoxy, but we made the mistake of removing the tape from the coils, which made them very unstable. We kept their shape, but the process became much more delicate and time-consuming because of this choice. After fixing the coils to the stator, we then epoxied them to help them retain their shape and pressed them under a weighted plate to ensure uniform height, then traced and labeled the phase wires for each coil group. We had intended to then dip the stator in epoxy to fix everything in place, but ended up drilling new holes through the steel and ran zip ties through to fix the coils in place; this was done to ensure that if we somehow generated unexpectedly large currents, the resulting force would not unseat the coils. At this point, the phase coils were wired in series, the negatives soldered together to form the Wye neutral, and then positive ends stripped and soldered to our rectifier.

Final assembly of the generator was to take the individual parts and mount onto 1/2” plywood. At this point, the mounting bolts for the 80/20 rails were mounted, as well. Final fit of the air gap (rotor distance from stator, magnet distance from coils) was established using locking collars on the shaft, and reinforced with stacked washers on the shaft (to maintain the air gap).

Note for the wiring diagram, we used an online resource to help us create and analyze it.

Of the three major problems, numbers 1 and 2 would have been relatively easy to ameliorate, if not fix, by simply changing the air gap; problem 3 would have been unfixable without changing the number of slots (i.e. winding new coils, designing a new stator, etc), and; problem 4 would also have required a stator redesign and new coils.

We also suspect that we basically left performance on the table by installing our magnets touching each other, and that this was a mistake. Spacing out the magnets so that they were not in contact with each other would have been the better choice, even if it grew our design a bit.

I would like to thank all the help and available resources for making this project and my lab partner(s).