Testing and debugging techniques for brushed DC motors

Article By : James Lockridge

Here are tips for debugging stepper motor systems as well as general bench testing advice.

The tips and tricks when testing motor-driver ICs on the lab bench can help engineers reduce time spent on evaluating and prototyping motor systems. The first installment of this two-part article series focuses on debugging brushed-DC motor systems, starting with how to set up bench for brushed-DC motor testing.

  1. Setting up bench for brushed-DC motor testing

Get a current probe. It’s almost always the first debugging suggestion to resolve issues such as “Why does my motor not spin?” or “Why does my motor spin in this weird way?” In brushed-DC motors, torque is a function of current in the winding, so looking at the current waveform gives you an understanding of why the motor might not be spinning, or why it might be spinning abnormally. Many of the examples in this article require a current probe to debug. However, before using a current probe, remember to degauss and zero the probe to ensure the correct current measurement.

Use a bench supply that can source enough current. The current limit on the bench supply may clamp the supply rail voltage when trying to drive large motor currents caused by inrush or stall. Be sure to select a robust supply and set the current limit high enough for the motor under test.

  1. Familiarize with proper brushed-DC motor current profiles

When debugging motors, it helps to know what current profile to expect. Figure 1 shows a typical profile for a brushed-DC motor. Brushed-DC motors will have large inrush or startup currents when initially energized. As the motor comes up to speed, the back-electromotive force (back EMF) increases, causing a reduction in current. The back EMF is a voltage that the motor generates to oppose the terminal voltage. When the motor stalls, the current will increase to a level equal to the terminal voltage divided by the motor winding resistance. A stall can occur from a mechanical failure, or the load reaching its end of travel.

Figure 1 Scope shot of current in a brushed-DC motor driven at 100% duty cycle. Source: Texas Instruments

If the brushed-DC motor acts strangely while connected to the driver, disconnect the motor from the circuit and power it directly from a bench supply. When connected directly to the supply, the current profile should look like Figure 1. If the current does not look like Figure 1, there may be an issue with the motor. If the current looks like Figure 1, then check the motor driver settings or the MCU firmware to ensure everything is operating as intended.

  1. Current regulation schemes in brushed-DC motors

Sometimes, what causes strange behavior is a feature of the driver that becomes misconfigured. Figure 2 shows the current waveform for a brushed-DC motor driven by a driver with the current regulation level set to limit the inrush and stall currents. While the current regulation feature is meant for this exact purpose, the motor may not produce enough starting torque to spin the motor if the current regulation level is set too low.

Figure 2 The current waveform for a brushed-DC motor driven by a driver IC. Source: Texas Instruments

Sometimes, the current regulation scheme can interact with the pulse-width modulation (PWM) signal sent to the motor-driver inputs. Generally, motor drivers will prioritize the current regulation response by going into a slow-decay state for a fixed off time rather than following the logic table for the input pins. The motor-driver datasheet will have specific details for how the current-regulation scheme operates.

  1. PWM issues for brushed-DC motors

Figure 3 shows how current flows through the full bridge of a brushed-DC motor driver during PWM. Typically, the system will have better performance when the PWM-off state uses slow decay. Table 1 shows the driver logic table for the DRV8251A driver IC. In order to drive with 50% PWM in the forward direction, configuring the inputs as IN1 = 1 and IN2 = 0 during the on time and IN1 = 1 and IN2 = 1 during the off time recirculates the current in the low-side FETs and enables the motor to maintain its torque during the off time.

Figure 3 This is how currents flow through an H-bridge. Source: Texas Instruments

Table 1 The PWM logic table is shown for the DRV8251A driver IC. Source: Texas Instruments

If IN1 = 0 and IN2 = 0 during the off time, then the outputs disable, and the current quickly flows to the supply rail—and into the bulk capacitor—through the FET body diodes; hence the description, “fast decay.” For motors with low inductance, the current can quickly go to 0 A during the off time, and the motor may lose its torque.

Some applications such as voice coils or galvanometers may require fine control of the PWM duty cycles. For those applications, the system may benefit from using fast decay during the PWM off time. The off time can also be a mix of fast decay and slow decay for better control of the current waveform.

Another potential issue related to driving brushed-DC motors with PWM is a sudden reduction in the PWM duty cycle. When this occurs, the driver can act like a boost converter and pump up the voltage on the supply rail to the full bridge. Figure 4 shows an example waveform of this behavior, which can damage the motor driver and other ICs on that supply rail. The amount of supply pumping depends on load inertia and speed. It’s possible to mitigate supply pumping by reducing the PWM duty cycle slowly, or keeping the motor in a slow-decay state until the rotor fully stops.

Figure 4 Supply rail pumps up when the duty cycle drops from 100% to 50%. Source: Texas Instruments

  1. Capacitors on brushed-DC motor terminals

Brushed-DC motor manufacturers often embed capacitors inside the motor to help with electromagnetic noise and transients caused by the brushes making and breaking contact with the commutator. These capacitors may be embedded on the rotor, connected between the motor terminals, or connected from the motor terminals to the motor case. Some engineers add their own capacitors externally as well.

Figure 5 shows capacitors inside of a car window motor. When this motor is fully assembled, the capacitor leads make electrical contact with the motor case.

Figure 5 The closeup of the capacitor (above) and brushes and commutator (below) are shown when capacitors are embedded inside a car window motor. Source: Texas Instruments

For brushed-DC motors controlled by a motor driver, the capacitors will briefly draw large currents at each PWM or current regulation cycle. The capacitor current-voltage relationship—i = C × dv/dt—explains that when the capacitor voltage changes drastically in a brief amount of time, it will draw a large current. These large capacitor currents may trip the motor driver’s overcurrent protection or current regulation functions causing inconsistent operation of the motor system.

Solutions include:

  • Adding inductors in series with the motor.
  • Adjusting the drain-to-source voltage monitor in gate drivers. This is the overcurrent protection feature in gate drivers.
  • Choosing a different motor.
  • If adding externally, choosing capacitors (typically <3 nF) so that capacitor inrush transients last less than the deglitching time of the motor driver’s overcurrent protection. Many motor drivers with integrated FETs have an analog current limit that keeps the magnitude of these transients within the safe operating area of the full-bridge FETs.
  1. Gate drivers

Motor drivers may either integrate the power MOSFETs or provide the gate-drive signals for the hardware engineer to drive their own MOSFETs. Compared to integrated MOSFET motor drivers, gate drivers have the added design challenge of routing gate-drive and sensing signals to MOSFETs. On top of that, parasitic inductances and capacitances can cause transients that impact electromagnetic interference (EMI) and may damage the MOSFETs or driver, as described in the application report, “Understand Smart Gate Drive.”

The gate-drive signals are the first place to check when debugging a gate driver. Figure 6 and Figure 7 show examples of proper signals from the DRV8706-Q1 gate driver datasheet.

Figures 6 The gate-drive signals are shown during the PWM operation. Source: Texas Instruments

Figure 7 The gate-drive signals are shown during the motor startup. Source: Texas Instruments

The main advice for debugging gate-drive signals is to measure with the probe near the pins. If the driver experiences failures, measure near the driver pins. For the best measurement, connect the probe ground to the nearest driver GND pin. If the MOSFETs fail, measure near the MOSFET pins. Additionally, try to reduce loops in the probe measurement.

Figure 8 shows the tip-and-barrel method for reducing loop area. Transients can couple to the probe when using the alligator clip probe ground with a long wire. The extra inductance of the ground wire can add ringing to the measurement that is not actually there.

Figure 8 The setup provides a view of the tip-and-barrel measurement method. Source: Texas Instruments

Also, consider using a differential probe to measure the MOSFET gate-to-source voltage or gate-to-drain voltage directly. The differential probe is especially helpful when debugging high-side gate-drive signals.

If the gate driver needs further debugging, probe the charge pump pins. Most gate drivers have external charge pump pins—VCP, CPH and CPL—connected to capacitors. The charge pump provides the gate-drive voltage for the high-side MOSFETs. Gate drivers have additional fault modes related to the charge pump and gate-drive signals.

Alternatively, some gate drivers, such as DRV8770, use a bootstrap architecture to provide the high-side gate-drive voltage. In this case, probe the pins where the bootstrap capacitor connects.

Motor behavior understanding

Having the proper bench equipment and understanding motor behavior is an important starting point for debugging brushed-DC motor systems. If the motor acts strangely, look at the voltages and currents on the motor terminals. If those signals look strange, check the motor-driver supply, input signals and feature configurations. Also double-check the bench equipment to make sure that you are using proper current limits and sampling resolutions.

Editor’s Note: In this two-part article series, the follow-up article will delve into stepper-driver debugging, offering tips for accelerating evaluation and staying safe when testing motor drivers on the bench.

The author wishes to dedicate this article to Rick Duncan, his mentor in the early days as applications engineer in TI motor drive designs. He also thanks Clark Kinnaird, Ryan Kehr, Pablo Armet and Pedro Arango Ramirez for suggesting some of the tips mentioned in this article.

This article was originally published on Planet Analog.

James Lockridge is a system engineer for motor drives at Texas Instruments.

 

Join the Conversation

  1. JosenDong says:

    Professional, very professional, learning…