Introduction
The aim of this article is not to provide ultra-accurate absolute efficiency numbers for both mentioned motor controllers, but to give a basis for comparison in a typical solar racing setup between the two controllers.
As you can see for yourself, the controllers perform very similarly.
Test setup
Both inverters were measured in the same test-bench with the same electrical and mechanical setup, except the inverters themselves.
The motor used in these tests is a typical solar-racing motor.
(If anyone would like to lend us a Mitsuba motor, we will gladly repeat the measurement. Please contact us at info@dq-drives.de)
The motor is loaded by a load machine, with torque measured by an HBM torque transducer in the shaft. The DC bus is supplied by a bench power supply of the Elektro-Automatik EA-PSB series. A Hioki PW3390 power analyser measures the three-phase output by the two-wattmeter method and the DC input on a separate channel. Its internally averaged values are logged every 200 ms for the duration of each test.
Each inverter is run through the same grid of operating points: 8 speeds from 300 rpm to 1000 rpm in 100 rpm steps, and 9 torque levels per speed spanning roughly 3.5 Nm to 20 Nm. Speed is the outer loop; at each speed the torque is swept upward through all levels. Every point is ramped to its setpoint and then held while the bench settles.
Both inverters are mounted on equivalent heatsinks with forced air cooling. Their temperatures are kept between 20°C and 40°C.
The XL's 12V supply runs from its own integrated HV → 12 V converter, so its auxiliary consumption is drawn through the DC bus and is contained in the measured P_dc. The WaveSculptor 22 runs from a separate external 12 V supply, so its auxiliary consumption doesn't pass through the measured DC bus and is absent from its P_dc.
See section "Comparison calculation" for details.
Comparison calculation
Operating points are extracted from the continuous log by discarding the ramps: a sample counts as settled only where neither the speed nor the torque channel is changing, the resulting intervals are eroded at both ends to remove the settling transient, and each remaining interval is averaged into one operating point.
Because of the auxiliary supply difference, the measured P_dc values are not directly comparable. Both are therefore normalised to the same scenario - each inverter has an auxiliary 12V DC supply, which is powered by a 90% efficient DC/DC from HV to 12V.
The Wavesculptor P_dc is therefore adjusted to contain its constant 4.5W consumption measured on the 12V rail transmitted through a 90% efficient DC/DC converter, resulting in 5W added to P_dc of the Wavesculptor.
As the XL-Classic supplied itself through its integrated HV to 12V DC/DC, which consumes 3.5W on the HV bus. When the XL-Classic is supplied through a separate 12V source, it consumes 2.5W on the 12V rail with the applied switching frequency (See: XL Solarracing auxiliary power consumption). These 2.5W result in 2.78W on the HV bus through a 90% efficient DC/DC.
Overall, 3.5W are subtracted from the XL's P_dc and 2.8W are added , resulting in an overall subtraction of 0.7W.
The resulting efficiency map difference is shown in the following figure:
Conclusion
As seen in the provided figure, efficiencies of the XL-Classic Solarracing Edition and the Wavesculptor22 are in close proximity for the entire operating envelope. It is visible that at higher speeds and higher phase currents (high torques), the XL-Classic slightly outperforms the Wavesculptor22. This is most probably due to the XL-Classic using latest generation Power-MOSFETs, which have reduced on-resistance in comparison to the Wavesculptor's last generation Power-MOSFETs. The XL's higher efficiency at higher speeds and low torques may be explained by its overall lower auxiliary consumption. It should be noted though that the efficiency benefit at low torques is almost negligible considering the small powers transferred at those operating points.
Further tuning of switching frequency, observer and motor parameters are possible for both inverters to optimise the absolute efficiency of the drivetrain. The Xl-Classic's VESC® based firmware gives the users lots of options for fine-tuning, while the Wavesculptor software is known to be limited in this regard.
The Wavesculptor22 is a long-time proven inverter for solar racing applications, yet the XL-Classic Solarracing Edition provides similar efficiency at better usability, additional features and a lower price point.