Rewinding the Armature

As it turns out, my reasoning was flawed concerning the windings and number of poles. According to the text books I was able to find, the coils should span 180 magnetic degrees, or from the north pole to the adjacent south pole. By adding magnets, but not rewinding the coils, I was using coils meant for 4 magnets, and thus spanning 9 armature slots (37/4 = 9 and omit the remainder). This meant that each coil is now spanning 5 changes of polarity or 900 magnetic degrees. Turning the 20 pole generator without rewinding the armature yielded poor results. I thought it over, read a lot and decided I'd have to rewind the armature anyway.

A word about the steel spacers. I originally was going to put them facing the housing. However, in testing I discovered that the magnetic pull from the steel was considerable more than that from the ceramic magnet that it was glued to. I wanted that strength facing the armature. I also figured that if the air gap got too close, it would be easier to grind the steel down than the ceramic. As it turned out, the air gap is small, but when the casing is snugged up, the armature doesn't rub on the pole pieces.

Adventures in rewinding

Before attempting to rewind the armature, I went to the local university library and took out a few books. The best of these IMHO was Electrical Machines, Direct and Alternating Currnet by Charles S. Siskind (McGraw-Hill, 2nd Edition 1959). Reading this helped me visualize what I had to do.

To begin with, I had to strip the old wire from the armature. This is easier said than done. I had hoped to trace the path of the old wiring, but this proved difficult. The wires were heavy and were varnished in place, making for a physically challenging task. It took the better part of a day to pull all the wires from the armature and my right arm was stiff for days afterwards.

When it was finally stripped of all wire, it looked like this:

The stripped armature showing the large copper commutator and the armature slots ready for new coils.

I left the plastic inserts in each slot to protect my own coils. I used the jig shown below to wind the new coils. I had no idea how many turns I'd be able to use and did a little trial and error before settling on 60 turns using 21 guage wire. I was hoping 21 guage copper wire would carry 10 amps or so at 60 to 90 volts retaining the 600 watt rating of the original motor at a much slower speed. Revisiting the tables from above, I made some rash assumptions and postulated the following, which turns out to be rashly optimistic, especially the part about gaining, not loosing, wattage.






Wire guage


Original - 12







Assuming - 12







Maybe - 60 to 90


612.5 - 900





Only testing would tell if I was totally out to lunch. It turns out that I was way off. See the test results to see how far off this expectation really was.

In order to make 37 coils, I made the jig shown in the next two pictures. It is made from 1/8" masonite board, scrap 2x4 lumber, a few screws, a " dowel and a bolt. Epoxy was used on the end piece only, since it had to be disassembled 37 times to extract the coils.

The motor serves only to hold the spool of wire. It isn't plugged in and does not run. Here are two other photos of the winding jig. View 1 and view 2.

37 coils ready to be installed. Note the tape for holding the shape before installing. Each coil is 1/8" thick to fit into the armature slots. The tape is removed on installing so the wires can be moved around inside the slot. The coils are longer than the armature slots so they can go around each other.

The first six coils, lower ends installed. Tape secures the coils until they are installed in the slots.

View showing that each coil skips a slot. Wood strips secure each slot. Each slot contains the front of one coil and the rear of another. Room must exist for the coils belonging to the slot in the middle to cross between the two coils.

The semi finished back side of the armature. Note how the coils are bent over on each other to conserve space.

Commutator side of the armature, showing the connections to the commutator segments. There are two wires attached to each commutator section. When the last segment is connected, every segment is connected to every other electrically. I used this wiring diagram to determine the pitch (spacing) of the wire ends.

Side view before securing connections. The connections were secured by pounding on the slots holding the wires with a nail punch. I tried soldering, but my heat source wasn't strong enough. Since the slots in the commutator sections originally held much larger wire (15 vs 21 gauge) I wrapped the two wires together, pulled them through one of the two slots, and pounded a lot of brass to get a solid connection.

Next - finished generator