December 2014
Characterisation of the Turnigy 150cc equivalent
brushless out runner DC motor
In summary the motor performed well and is recommended. It has high efficiency 85 - 88% over a wide operating range, is nicely made and mechanically smooth. A larger one of similar design (say longer stator) would be welcomed.
Design and construction
This motor is the largest available from Turnigy under their Rotomax range, being cited as 150cc equivalent. Intended for large RC model planes, the motor also finds potential application in manned electric aircraft and small electric vehicles. The motor was selected as being the newest design and also has windings only close to the rotor with an open area in the centre. This design feature is also found in premium drone motors with very large diameter and short air gapped rotors. Turnigy also sell two 100cc equivalent motors of similar weight and size.
45V 170A 167rpm/V 2.07Kg 70V A? 150 rpm/V 2.73Kg
It is not clear why the motor chosen is considered 150cc equivalent as the older CA120 appears similarly capable. The 100cc looks similar to the 150cc with a shorter stator.
The out runner design is almost universally used for electric model aircraft and drones due the high efficiency and power density. For a given speed, torque and hence power is proportional to radius2 of the air gap. A high pole number and high drive frequency keep the magnetic path short and allow a relatively thin back iron outer rotor cylinder. The brittle rare earth magnets are much stronger loaded in compression inside the cylinder than in tension in a conventional in runner design. The disadvantage of the out runner is balance is critical and forces also follow a radius square law. Also the bearing arrangement is more demanding as the stator feed and mounting need to be inside the bearing diameter. Larger bearings then see a higher speed. The low cost and small size of the electronic speed controller (ESC) also contribute to the drive performance, thanks to the advances in low voltage MOSFET performance, micro controllers and Chinese volume manufacture.
The motor has noticeable cogging in line with smaller types though this quickly smoothes out once spinning at any significant speed. The external field is lower than other motors the author has experience of such as the older C80100 design and C5045 but still plenty for the hall sensors. "Back iron" for the magnets is a compromise between performance and weight. Thinner iron will weigh less but links less field reducing the field strength.
Published specifications
Maximum voltage 52V (14S Lipo)
Maximum current 190A (DC)
Maximum power 9.8KW
Emf constant ("KV") 6.67V/1000 rpm (150 rpm/V)
Winding 8 turns
Stator poles 24
Rotor poles 20
Stator length 38mm
Stator dia 110mm
Lamination thickness 0.2mm
Resistance 11mOhm
Weight 2.52Kg
Measured/derived specifications
Emf constant ("KV") 6.77V/1000 rpm
Resistance 4.7mOhm (per phase assumed star)
Inductance 8 - 14uH depending on rotor position
Torque constant 0.06Nm/A
Weight 2.46Kg (with back plate but no prop h/w)
Test set up and drive electronics
To test the motor two were purchases and coupled together as a motor/generator pair. If the generated power is then feed back to the motor, only total losses have to be supplied externally. Also if the machines are considered fully bidirectional then the losses are shared near equally between the two. Speed is primarily controlled by the voltage supplied across the motor by an external variable DC PSU (PSU2). Torque/current is primarily controlled by a second PSU (PSU1) in series with the generating machine in turn connected in parallel with the voltage one. At low speed and high current the power supplied by PSU1 supplies all the losses and PSU2 drops to zero. The motor voltage is then not regulated so some additional load is required on the generator to give a little load on PSU2, regulate the constant speed and achieve the desired motor current.
The machine chosen as the motor is driven by a simple 3 phase bridge voltage source inverter. Hall sensors picking up stray rotor field provide commutation pulses to the bridge via some simple logic. The sensors are Allegro A3290KVA latches, that is they switch output around where the field reverses so give an approximate 50/50% duty cycle independent of field strength. The positions are adjusted to make the 3 outputs space at 120 deg. Also the overall position is adjusted to determine the motor direction and obtain the cleanest drive waveforms and lowest idle current.
Upper bridge switches are nominated A, C and E. Lower B, D and F. Sensors X, Y and Z. X means inverse of X. The logic of the drive is then:
A = X + Z C = X + Y E = Y + Z
B = X + Z D = X + Y F = Y + Z
Each phase is driven by a 2/3 duty square wave spaced at 120 deg. So each switch is on for a duty of 1/3. Though there is no cross conduction risk in each bridge leg there remains a brief interval where each phase is shorted out so time delays in each switch drive shorten the on time slightly. For each switch, two International Rectifier FS3107s were paralleled giving a switch on resistance of around 1 mOhm. 75V in a multi leg To220 size package. These were driven by Agilent HCPL 3120 driver/opto couplers with floating rails for the upper switches. Switching losses were ignored and a simple i2R correction used for the driver loses assuming the DC current flows through two switches.
In normal operation the ESC would control the motor speed in response to a demand signal and load conditions. The upper switches would be further PWM modulated at a high carrier frequency typically 8 or 16KHz to control the current, with a fixed DC supply (battery) supplied to the ESC.
For the generator a 3 phase rectifier was made up using 24 MBR3060CT Schottky diodes, 4 diode pairs per rectifier. Losses were measured for two rectifiers in series as the theoretical duty is 1/3 as for the driver switches.
The motor can be run "open loop" as a synchronous motor. This was done initially to help set up the Hall sensor positions. The motor had to be started manually and over driven to remain lock. If the drive was gradually reduced it would drop out at about the same current as it naturally draws unloaded in a true brushless drive scheme.
Waveforms
Left is the unloaded EMF from the generating machine, at 10V 250Hz, approx 1500 rpm. The trapezoidal waveform matches the drive reducing harmonic currents (maybe better than sine wave).
Motor/generator pair unloaded at 10V 250Hz, approx 1500 rpm. Lower waveform is phase current, 10A div.
Below left unloaded at 20V, 500Hz 3000 rpm Loaded for motor at 100A DC.
Above coupled motors, drive electronics and rectifier.
Losses and cooling
Machine losses depend on voltage/current and therefore speed/torque. Windage and bearing friction depend primarily on voltage/speed. Copper losses depend on current and magnetic losses a combination of the two. The rotor was noticeably heated just by speed, the generating machine's got just as warm as the motoring one unloaded. This indicates the rotor iron sees an alternating magnetic field as the magnetic path changes as the magnets pass each pole. In the test set up the DC voltage controls speed so there is no additional PWM switching and hence loss.
The motor features slots in the stator with chamfered edges which appear to act like an extraction fan, giving some cooling at the upper speed ranges. For the test fixture this was supplemented by a 12W 24V 90mm fan pressurising an enclosure around the coupling. This then blows air though the machines supplementing the motor's own fan feature. In applications where high load can be placed on the motor at low speed such as electric vehicles, additional cooling is desirable, reducing copper losses and winding temperature. In it's intended application driving propellers, cooling would be provided by air from around the roots of the blades.
Ultimate performance is limited by maximum speed and current. Speed is primarily limited by mechanical strength with the published limit being around 7500 rpm. Current comes down to cooling. Generally electrical machines have high peak capability relying on their thermal inertia though with small high speed machines like the Turnigy 150cc the time is relatively short.
Performance
The key performance factor is efficiency, idle current was also measured for both machines coupled and just one as a motor.
Efficiency with current at various voltages:
10V / 20V / 30V / 40V / 50V0
10
20 / 75.5 / 70.2 / 64.4 / 58
30 / 82.8 / 79.3 / 76 / 72.3 / 68
40 / 85.3 / 83.7 / 81.5 / 78.7 / 74.9
50 / 86.5 / 85.9 / 84.5 / 82.3 / 79.3
70 / 86.9 / 87.8 / 87.1 / 85.9 / 83.9
100 / 85.5 / 87.7 / 88.2 / 87.8 / 86.4
130 / 77.9 / 87 / 87.5 / 88 / 86.8
160 / 81 / 84
Peak efficiency of 88% was achieved around 30V 100A. 85% maintained over a wide range.
Light electric motorcycle application
A light 4 stroke motorcycle, say 125 cc 10KW/10Nm, would achieve 60 mph + and good acceleration with 3:1 gear spread. For the motor a torque constant of around 0.06Nm/A was derived, so 9Nm is possible at 150A. 3 motors mechanically in parallel could then produce 27Nm, roughly equivalent to the internal combustion engine with gears. Maximum motor speed is 7500 rpm @ 50V, a little lower than an engine but enough for 60 mph. The inverter would need to be 22.5KVA, supplying 3 x 150A at rest 0V, tapering to 200A total at 50V, maintaining 10KW. Base speed above which the torque starts reduce is then 22.2V.
Paul Bennett BSc Eng MIET
2