I am experimenting with various schemes to rewind AC induction motors to change their characteristics for special purposes. I have successfully rewound several single phase motors for three phase, and I have changed their speed by varying the number of poles. I am winding them with much heavier wire than originally used for the 120 and 240 VAC machines, as I want to use them for low voltage battery powered applications.
The most successful motor so far is my first motor, although I rewound it several times before I got it to work properly. It was a 1/2 HP capacitor start pump motor with a 24 slot stator. I wound it using #14 AWG THHN wire, which required it to be threaded through the slots rather than prewound and pushed in. The winding scheme I used was a serpentine path which went in slot 1, out slot 7, in slot 13, and out slot 19 (for phase A), with a total of 8 turns. The next slot consisted of 4 turns of phase A, and 4 turns of reverse polarity C (C-). Then 8 turns of C-, and then 4t C- 4t B, etc. Thus I had two pole pairs and the motor runs at 1740 RPM with about 4 VAC phase to phase and a current draw of about 5 to 10 amps per phase. The motor seems to have good torque and runs fairly quiet, even though my three phase source is imperfect.
I also built a modified sine wave generator using a PIC and a six-MOSFET bridge, and was able to run the motor at several speeds, with current draw of only 1 or 2 amps on a 12 VDC battery.
My next attempts used 36 slot stators, and I tried to get the highest number of poles possible for slowest speed and highest torque. I wound the stator also in a serpentine pattern but with only A, -C, B, -A, C, -B windings, for a total of 12 pole pairs. I expected it to run at about 550 RPM, and it actually runs about 510 RPM. However, it does not want to start itself, and it draws about 25 A on two phases, and only about 2 amps on the third phase, at about 5-8 V/phase. I think this imbalance is because of my test setup, but it does seem like this motor is much less efficient, and seems to have much less torque than the previous motor.
For my third attempt, I got some #15 AWG magnet wire and wound the motor in the more conventional and faster way, although it was hard to work with the stiff wire, and I had to force it into some of the slots. The pattern I tried here was 8 turns completely around pole piece 1, 7, 13, 19, 25, and 31 for phase A, then 8 turns on 3, 9, etc for B, and the same for C on 5, 11, etc. Then I wound 8 turns on the even numbered pole pieces with opposite polarities -C on 2, 8, 14, etc, -A on 4, 10, 16, etc, and -B on 6, 12, 18, etc. I had a total of 12 wires out of the motor, labeled A1, -A1, B1, -B1, C1, -C1, A2, -A2, B2, -B2, and C2, -C2. I connected -A1, -B1, and -C1 together, joined -A2 to A1, -B2 to B1, and -C2 to C1, and hooked power to A2, B2, and C2. I was surprised that the motor ran at about 235 RPM, rather than about 550 as I had expected for what I thought was a 12 pole motor. It also produced an impressive display of internal sparks cause by damaged insulation on the magnet wire. Then I connected the power only to A1, B1, and C1, and the motor ran at about 115 RPM. In this case, I was only energizing the positive polarity (North) poles, so that the South poles were created by the consequent pole phenomenon that is used in some multi-speed motors.
I made one more attempt on this same motor, after removing the damaged wire and removing sharp edges so I could rewind more safely. Here, I wound 14 turns completely around the odd pole pieces 1, 3, 5, etc, in a pattern A, B, C, in an attempt to make a 12 pole motor, but I got about the same result as my first attempt, with about 234 RPM and very poor torque.
I have done some analysis of what might be happening, although my knowledge of electromagnetic fields is rudimentary. It seems that, in this last motor, the vector sum of currents in the slots next to the pole pieces that were not explicitly wound, makes a pole of positive polarity, so the actual sequence of poles is A, C, B, etc., for 36 pole pieces, so it is essentially a 24 pole motor that would be expected to run at 3600/12 = 300 RPM, so 234 is reasonable. I suppose the negative polarity consequent poles do not exist in the pole pieces, but perhaps deep in the slots, which makes for a very wide magnetic gap and consequently very poor torque.
The purpose of this long discussion is to solicit ideas and recommendations so that I can design a practical, efficient, high torque, low speed motor to run on about 6 to 12 VAC at 60 Hz, and be capable of running up to 3600 RPM (or higher) by means of a V/Hz PWM drive. I have some ideas for a 12 pole motor where I may wind pole pieces 1 and 2 with phase A, 3 and 4 with phase B, and 5 and 6 with phase C. Then I will wind 2 and 3 with -C, 4 and 5 with -A, and 6 and 7 with -B. If I am correct, pole piece 1 will have a vector sum of (A-B), which will be at -30 degrees to A, with amplitude about0.86 compared to purely A, and the pole sequence will be correct for a rotating field. If this is correct I may rewind a 1 HP 8 pole or 2 HP 2 pole motor (both 3 phase). I have one of each, and they are almost identical in size and weight, yet the HP is double in the faster motor. I am curious why this is so, and I will pay special attention to the winding scheme. Then I will concentrate on a PWM controller.
I look forward to your comments. Thanks.
Paul E. Schoen, President P S Technology, Inc.