Comparison of RPM and Load on the Efficiency on An Franklin-Thomas Company StarPower16 Zero Cogging Generator By Joe D. Shepard Chief Engineer, CEO, Chairman Franklin-Thomas Company, Inc. Abstract This paper discusses the evaluation of Franklin-Thomas Company s patent-applied-for StarPower16 generator s efficiency at different RPM and with different loads. The generator was subjected to various loads and operated at varying speeds to fully evaluate the generator s performance. This particular permanent magnet generator was hand-wound with double strands of 18 gauge magnet wire with a combined ampacity of 32 amperes. There were 36 coils interleaved into 36 slots. Adjacent coils were phased together and rectified using a 25 amp 1,000 volt bridge rectifier. The coil insulation was 410 Dupont Nomex paper. The filter capacitor was a 1,000µ 100 volt electrolytic wired across the bridge. This arrangement created 18 separate rectified DC sources. The 18 rectified DC sources were series-ed together to form a single DC output. The output wire was 12 gauge stranded rated at 600 volts. The generator stator and rotor were made from 24 gauge M19 non-grain oriented electrical steel laminates cut by a laser. It was bolted together using ½ -13 threaded rods. The magnets were 1 inch by 2 inch by ½ inch N42 neodymium. The case was made from ¼ inch 6061 aluminum. The load consisted of 9 water heater elements. There were seven 3,500 watt elements, one 4,500 watt element, and one 1,500 watt element. These elements were suspended in a 5 gallon pail of water. The driving force was a 3 hp three phase electrical motor. The motor speed control was a WEG CFW-10 capable of driving a 3 hp motor. The torque sensor was a Himmelstein 48292V11-31NN, S/N 48202V-4112-473 rated at 7,000 RPM at 1,000 lb/in. The torque sensor had calibration traceable to NIST. The software was Himmelstein-supplied DT482 V.3.142.1. The meters were Kline MM2000 True RMS. The serial to USB cable was a B&B Electronics model (P/N232USB9M, S/N 0170833012) supplied by Himmelstein. The temperature sensor was a Ryobi IR001 noncontact infrared thermometer. The test was performed using all procedures recommended by the various manufacturers.
Introduction The StarPower16 generators have been demonstrating some remarkable performance. The efficiency has been so high that electrical engineers and physicists disputed the validity of the numbers. While I understand the concerns of the educated community, the numbers I saw kept coming up. To begin, with a common wire-wound generator, creating the magnet field dynamically consumes 30% to 40% of the energy used turned the generator. This generator can never achieve high efficiency. Putting magnets in a generator that cogs introduces resistance as the magnets line up and are attracted to the steel between the slots. This cogging also causes vibration which would limit the rotor speed. With large magnets, the cogging generators will not turn. When I invented true zero cogging by changing the ratio of the magnets to coil slots, I was able to insert ever larger magnets into the generator without the problems and resistance associated with generators that cog. Using electrical steel, the effect of the eddy currents in the stator was dramatically reduced. Thus, we had a generator that spun as though it had no magnets at all. While zero cogging made the generator free-spinning, the alignment of the coils to the magnets made it difficult to generate any power because the magnets were cancelling the energy between the coils. To solve this problem, I rectified each coil individually and then combined the DC output from all the rectifiers. Not satisfied, I looked at how adjacent magnets lined up on the slots. In interleaving the coils, you end up with two coils in each slot. By phasing the two coils and then rectifying their combined output, I was able to increase the energy level. The illustration to the right shows the concept. The left side of the left coil is influenced by the north magnet while the right side of the right coil is influenced by the south magnet. Combined as I have done here, we retain the value of the zero cogging while enhancing the power generated. A side benefit is that the generator consists of 18 individual power supplies that can be combined in a variety of ways. They all can be series-ed for high voltage. One can be used as a low voltage supply. You could group them in sets of six to triple the current while lowering the voltage --- as long as you employ some sort of load sharing circuitry.
While the generator can be made to produce lethal voltages, the voltage across the individual coils is comparatively low. This means the only high voltage issue would be on the output, which can be easily controlled. The Testing Process Here is the actual image of the test bench I used. It s not laboratory pretty, but it shows I make do with what I have available. To determine power in watts supplied to the generator, we instruct the Himmelstein software to read the transducer once per second. This gives us a fairly consistent number and avoids any issues. Once the Himmelstein has settled and we have written down the value, we read the RPM of the generator using the non-contact tachometer. Only then to do we read the generator output voltage and output current.
The formula for converting the collected numbers to power in to the generator is: (((Torque * RPM) / 5252) * 745.6) = watts in We compare that value to the absolute power shown on the current and voltage meters to determine the power produced. We vary the load and the RPM to get an understanding of the efficiency of the generator. The speed of the motor is varied manually so it is difficult to get exactly the same speeds for each test. Another variable is the belt slipping on the motor at high loads. This affected the test adversely because we would take a reading on the Himmelstein and by the time we took the RPM data, the belt would slip enough to lower the output voltage and amperage. We waited between tests for the belt to cool because at the higher power, the belt was literally smoking. Test One: 51 ohm load Test was halted at 690 RPM because the power in exceeded the capacity of the 3 phase drive. Generator temperature 89º F. Motor temperature 141º F. Load 51 Ohms RPM Torque Power In Volts Amps Power Out Efficiency 84 0.63 7.51 7.70 0.15 1.16 15.37% 204 3.55 102.81 50.80 1.04 52.83 51.39% 250 4.99 177.10 70.20 1.41 98.98 55.89% 308 8.20 358.55 107.00 2.22 237.54 66.25% 320 8.97 407.50 114.30 2.36 269.75 66.20% 359 11.90 606.49 147.00 3.01 442.47 72.96% 387 14.52 797.74 165.00 3.40 561.00 70.32% 437 18.60 1,153.92 201.00 4.16 836.16 72.46% 482 21.50 1,471.18 228.50 4.73 1,080.81 73.47% 512 23.50 1,708.12 248.00 5.14 1,274.72 74.63% 560 25.50 2,027.26 272.80 5.65 1,541.32 76.03% 587 26.60 2,216.67 285.90 5.93 1,695.39 76.48% 637 27.80 2,514.00 324.00 6.31 2,044.44 81.32% 690 29.00 2,840.72 374.00 6.73 2,517.02 88.61%
Test Two: 100.6 ohm load Test Two was halted at 892 RPM because Power In had exceeded capacity of the 3 Phase Drive. Generator temperature 87º F. Motor temperature 136º F Load 100.6 Ohms RPM Torque Power In Volts Amps Power Out Efficiency 80 0.55 6.25 11.22 0.11 1.23 19.76% 188 2.10 56.05 57.70 0.57 32.89 58.68% 219 3.00 93.27 75.80 0.75 56.85 60.95% 277 5.10 200.55 118.00 1.17 138.06 68.84% 311 6.80 300.23 148.10 1.47 217.71 72.51% 360 8.75 447.19 182.60 1.81 330.51 73.91% 402 10.23 583.71 211.80 2.10 444.78 76.20% 450 11.46 732.18 238.90 2.37 566.19 77.33% 495 12.56 882.62 265.60 2.64 701.18 79.44% 547 13.54 1,051.45 290.00 2.88 835.20 79.43% 590 14.37 1,203.62 312.30 3.11 971.25 80.69% 639 15.20 1,378.88 337.10 3.35 1,129.29 81.90% 680 15.90 1,534.93 356.80 3.55 1,266.64 82.52% 773 17.00 1,865.56 405.00 4.05 1,640.25 87.92% 850 18.33 2,211.89 448.00 4.45 1,993.60 90.13% 892 18.87 2,389.56 471.20 4.68 2,205.22 92.29%
Test Three: 161.5 Ohms Test was halted at 98.79% to avoid issues. Generator temperature 88º F. Motor temperature 137º F. Load 161.5 Ohms RPM Torque Power In Volts Amps Power Out Efficiency 82 0.47 5.47 14.60 0.09 1.33 24.27% 147 1.63 34.08 61.00 0.38 23.18 68.03% 215 2.53 77.22 90.30 0.56 50.80 65.79% 271 4.05 155.81 134.60 0.84 112.88 72.45% 318 5.33 240.62 171.50 1.07 183.25 76.16% 380 6.59 355.45 212.30 1.32 280.82 79.00% 431 7.37 450.95 240.80 1.50 361.28 80.11% 491 8.20 571.58 273.40 1.70 465.72 81.48% 542 9.64 741.75 315.00 1.96 618.22 83.35% 610 11.09 960.38 367.30 2.29 840.56 87.52% 672 11.58 1,104.74 402.00 2.50 1,006.88 91.14% 741 13.32 1,401.21 459.00 2.86 1,312.65 93.68% 810 14.43 1,659.33 507.00 3.16 1,601.55 96.52% 850 15.41 1,859.53 543.00 3.38 1,837.07 98.79%
Conclusions It is obvious based on these tests that the StarPower16 Zero Cogging generator can produce significant power and will do so across a wide speed range. The efficiencies have gone much higher, but because of the resistance on the technical community to accept that data, we have left it out. The StarPower16 Zero Cogging can be run to 7,000 rpm easily and produce power through the entire speed envelop. The efficiency is remarkable and can be assumed to be tied directly to RPM and load. We feel the performance of this generator will be enhanced and improved upon. Engineers will craft solutions using the StarPower16 Zero Cogging generator for years to come.