Analyzing Change in Magnetization State

The MDA810 captures the three-phase line-line voltage, line current, and other signals and calculates associated waveforms (Figure 1). The acquired line-line voltage and line current acquired signals are not “typical” motor drive pulse-width-modulated (PWM) and sinusoidal current outputs. The voltage signals are drive output-controlled energy pulses applied to the stator winding to change the magnetic flux of the variable magnet, and the applied line-line voltages cause line-current signals to flow in the stator winding. Figure 1 illustrates a relatively short capture to show the detail, but the capture time could be much longer if the user desires.

Figure 1:

Three-phase line-line voltage, line current, and other associated acquired signals

The following describes the various acquired signals:

  • Channel 1, or C1 (yellow trace, top left) is the VR-S voltage, C2 (magenta trace, second from top left) is the VS-T voltage, and C3 (blue trace, third from top left) is the VT-R voltage. Teledyne LeCroy HVD3106 HV differential probes were used to acquire these signals.
  • C4 (green trace, top left) is the IR current, C5 (light green trace, second from top left) is the IS current, and C6 (purple trace, third from top left) is the IT current. LEM current transducers (IT 700-S ULTRASTAB) were used to acquire these signals, and we rescaled the voltage output of these transducers to Ampere units in the MDA810’s channel setup dialog.
  • C8 (orange trace, bottom left), a probed output signal from the control system, is a timer for the start of the switching period.
  • C7 (red trace, third from top right) is the acquired output of a resolver converter system that shows rotor angle over time (described as a variation in voltage over time).

The following describes the calculated waveforms:

  • At top right of Figure 1 is a per-cycle Waveform of the three-phase output power to the motor stator (orange trace, named P(ΣRST)). This waveform shows each calculated three-phase power value, calculated once per switching cycle and plotted in a vertical scale versus time. This waveform is time coincident with the other acquired waveforms.
  • Below that is the calculated integral of the P(ΣRST) waveform (yellow trace, F1).
  • The green trace below F1, DriveOut Sync (abbreviated as DrvOutS), is a calculated signal using C8 as an input. The DriveOut Sync waveform defines the period during which the MDA810’s algorithms perform power calculations. In this case, the instrument makes calculations once per power semiconductor device switching period. A transparent overlay on the DriveOut Sync signal identifies the various switching periods.

The cursors bound the approximate period during which the magnetization state change occurs. The cursor values in the F1 descriptor box (below the grid and above the Numerics table) clearly define the total energy applied to the machine terminals during the process of changing the magnetization state (shown as Δ26.171 Joules in the F1 descriptor box). The acquired resolver signal (C7) determines whether the rotor angle location is correct for application of the magnetization state-change pulse. The P(ΣRST) waveform displays the three-phase power of the pulses over time, with one value calculated per switching period. This determines whether the peak power delivery exceeds DC bus capacity or other parameters.

The Numerics table (below the waveform display grid) displays the mean value of multiple switching periods. Thanks to a line-line to line-neutral voltage conversion, voltage values display on a line-neutral basis, and calculation of per-phase power values proceeds accordingly. The Numerics table also displays the mean power value over the full acquisition (shaded light purple). If desired, the per-cycle phase power Waveforms could be displayed like in like manner to the P(ΣRST) waveform.

Analysis of Per-Phase Change in Magnetization State

Figure 2:

In the future, the Motor Drive Analyzer series will convert acquired line-to-line voltage waveforms to a line-neutral waveform and allow further per-phase calculations

Figure 2 illustrates some capabilities currently in development for the Motor Drive Analyzer series:

  • Left column: the three line-neutral voltage waveforms calculated from the acquired Line-Line waveforms.
  • Middle column: the three acquired line-current waveforms.
  • Right column: instantaneous power waveforms calculated from the line-neutral and line-current waveforms.

The acquired and calculated waveforms shown in Figure 2 foster understanding of the per-phase power delivery that enables the magnetization state change. In this example, we did not employ the MDA810’s three-phase power measurement capabilities, but rather a (future) line-line to line-neutral voltage-waveform transformation capability. The example also leverages the MDA810’s inherent capability to perform math on the acquired waveforms.

DC Bus Behavior during Changes in Magnetization State

Of course, the magnetization state-change pulses have an impact on the DC bus, thus making it necessary to ensure the DC bus carries enough energy to support them. Figure 3 illustrates the following waveforms:

Figure 3:

Determining the fitness of the DC bus for delivering magnetization state-change pulses

  • The three acquired line currents (C1, C2 and C3, top left grid).
  • The three acquired line-line voltages (C5, C6, C7, bottom left grid).
  • The acquired DC bus voltage (Z8, bottom right grid). Z8 is a vertical zoom of Channel 8 (C8) and the vertical zoom trace better displays the change to the DC bus during a very short magnetization process.
  • The calculated three-phase total real (PΣRST), apparent (SΣRST), and reactive (QΣRST) power per-cycle Waveforms. These are calculated for each device-switching period (99 periods total), yielding deeper understanding of power behaviors for each switching perioda unique capability of the MDA. These appear in the top right grid.
  • Also acquired, but not shown in Figure 4, is the digital output device-switching period from the control system (C4). This acquisition used a passive voltage probe. As described above, we calculate power quantities over this short switching period.

We used a 500 MS/s sample rate for this test, which is far in excess of that provided by a typical power analyzer. The test determined that the DC bus had adequate energy to support the magnetization state-change process.

Correlation of Control System Signals to Changes in Magnetization State

It is important to understand the correlation of the control system with the operation of the drive system. Correlation testing proceeded similarly to that of the DC bus, although the magnetization pulses were of shorter duration and different channels for the voltage and current acquisitions. Figure 4 shows the following waveforms:

Figure 4:

Testing correlation of control system signals with changes in magnetization

  • The three acquired line currents (C4, C5 and C6, top left grid).
  • The three acquired line-line voltages (C1, C2, C3, bottom left grid).
  • The control system’s calculated ID (d-axis current, C7, bottom right). This was a voltage input to the MDA and was re-scaled to current in the MDA channel setup dialog.
  • The calculated three-phase total real (PΣRST), apparent (SΣRST), and reactive (QΣRST) power per-cycle Waveforms. These are calculated for each device-switching period, yielding deeper understanding of the power behaviors for each switching perioda unique capability of the MDA. These appear in the top right grid.
  • The switching frequency digital output from the control system (C8, bottom right) acquired using a passive voltage probe and used for the Sync calculation (DrvOutS). As before, we calculate power quantities over this short switching period.

We determined that the control system’s behavior was consistent with expectations and requirementsthe ID magnitude and behavior were well within design boundaries.


The Teledyne LeCroy MDA810 Motor Drive Analyzer provides unique capabilities to analyze power quantities during very high-frequency events (e.g., a power-semiconductor’s switching period), show the dynamic changes of these values over time, and permit correlation to other measured power and control system signals.