BLDC Fan Current

Motors and Measurements

Today’s small fans and blowers depend on brushless DC (BLDC) motor technology for their operation. Here, Ed explains how these seemingly simple devices are actually quite complex when you measure them in action.

By Ed Nisley

The 3D printer Cambrian Explosion unleashed both the stepper motors you’ve seen in previous articles and the cooling fans required to compensate for their abuse. As fans became small and cheap, Moore’s Law converted them from simple DC motors into electronic devices, simultaneously invalidating the assumptions people (including myself) have about their proper use.

In this article, I’ll make some measurements on the motor inside a tangential blower and explore how the data relates to the basic physics of moving air.

Brushless DC Motors

Electric motors, regardless of their power source, produce motion by opposing the magnetic field in their rotor against the field in their stator. Small motors generally produce one magnetic field with permanent magnets, which means the other magnetic field must change with time in order to keep the rotor spinning. Motors powered from an AC source, typically the power line for simple motors, have inherently time-varying currents, but motors connected to a DC source require a switching mechanism, called a commutator, to produce the proper current waveforms.

Mechanical commutators date back to the earliest days of motor technology, when motors passed DC power supply current through graphite blocks sliding over copper bars to switch the rotor winding currents without external hardware. For example, the commutator in the lead photo switches the rotor current of a 1065 horsepower marine propulsion motor installed on Fireboat Harvey in 1930, where it’s still in use after nine decades.

Fireboat Harvey’s motors produce the stator field using DC electromagnets powered by steam-driven exciter generators. Small DC motors now use high-flux, rare-earth magnets and no longer need boilers or exhaust stacks.

Although graphite sliding on copper sufficed for the first century of DC motors, many DC motors now use electronic commutation, with semiconductor power switches driven by surprisingly complex logic embedded in a dedicated controller. These motors seem “inside out” compared to older designs, with permanent magnets producing a fixed rotor field and the controller producing a time-varying stator field. The relentless application of Moore’s Law put the controller and power switches on a single PCB hidden inside the motor case, out of sight and out of mind.

Because semiconductor switches eliminated the need for carbon brushes, the motors became known as Brushless DC motors. Externally, they operate from a DC supply and, with only two wires, don’t seem particularly complicated. Internally, their wiring and currents resemble multi-phase AC induction motors using pseudo-sinusoidal stator voltage waveforms. As a result, they have entirely different power supply requirements.

The magnetic field in the rotor of a mechanically commutated motor has a fixed relationship to the stator field. As the rotor turns, its magnetic field remains stationary with respect to the stator as the brushes activate successive sections of the rotor winding to produce essentially constant torque against the stator field. Electronically commutated motors must sense the rotor position to produce stator currents with the proper torque against the moving rotor field. As you’ll see, the motor controller can use the back EMF generated by the spinning rotor to determine its position, thereby eliminating any additional components.

Figure 1
The blower motor current varies linearly with its supply voltage, so the power consumption varies as the square of the voltage. The motor speed depends on the balance between torque and load.

I originally thought Brushless DC (BLDC) motors operated much like steppers, with the controller regulating the winding current, but the switches actually regulate the voltage applied to the windings, with the current determined by the difference between the applied voltage and the back EMF due to the rotor speed. The difference between current drive and voltage drive means steppers and BLDC motors have completely different behaviors.

Constant Voltage Operation

The orange trace along the bottom of Figure 1 shows the current drawn by the 24 V tangential blower shown in Figure 2, without the anemometer on its outlet, for supply voltages between 2.3 V and 26 V. The BLDC motor controller shapes the DC supply voltage into AC waveforms, the winding current varies linearly with the applied voltage and, perhaps surprisingly, the blower looks like a 100 Ω resistor.

Figure 2
An anemometer measures the blower’s outlet air speed and a square of retroreflective tape on the rotor provides a target for the laser tachometer. If you are doing this in a lab, you should build a larger duct with a flow straightener and airtight joints.

The blower’s power dissipation therefore varies as the square of the supply voltage, as shown by the calculated dots in the purple curve. In fact, the quadratic equation fitting the data has 0.00 coefficients for both the linear and constant terms, so it’s as good as simple measurements can get.

As you saw in March (Circuit Cellar #332) and May (Circuit Cellar #334), a stepper motor driven by a microstepping controller has a constant winding current and operates at a constant power. Increasing the supply voltage increases the rate of current change but, because the controller applies the increasing voltage with a lower duty cycle, it doesn’t directly increase power dissipation. …

Read the full article in the July 336 issue of Circuit Cellar

Don’t miss out on upcoming issues of Circuit Cellar. Subscribe today!

Note: We’ve made the October 2017 issue of Circuit Cellar available as a free sample issue. In it, you’ll find a rich variety of the kinds of articles and information that exemplify a typical issue of the current magazine.

Stepper Motor Back EMF

Supply Voltage vs. Current Control

Continuing with the topic of stepper motors, this time Ed looks at back electromotive force (EMF) and its effects. He examines the relationship between running stepper motors at high speeds and power supply voltage requirements.

By Ed Nisley

Early 3D printers used ATX supplies from desktop PCs for their logic, heater and motor power. This worked well enough—although running high-wattage heaters from the 12 V supply tended to incinerate cheap connectors. More mysteriously, stepper motors tended to run roughly and stall at high printing speeds, even with microstepping controllers connected to the 12 V supply.

In this article, I’ll examine the effect of back EMF on stepper motor current control. I’ll begin with a motor at rest, then show why increasing speeds call for a much higher power supply voltage than you may expect.

Microstepping Current Control

As you saw in my March 2018 article (Circuit Cellar 332), microstepping motor drivers control the winding currents to move the rotor between its full-step positions. Chips similar to the A4988 on the Protoneer CNC Shield in my MPCNC sense each winding’s current through a series resistor, then set the H-bridge MOSFETs to increase, reduce or maintain the current as needed for each step. Photo 1 shows the Z-axis motor current during the first few steps as the motor begins turning, measured with my long-obsolete Tektronix Hall effect current probes, as shown in this article’s lead photo above.

Photo 1 Each pulse in the bottom trace triggers a single Z-axis microstep. The top two traces show the 32 kHz PWM ripple in the A and B winding currents at 200 mA/div. The Z-axis acceleration limit reduces the starting speed to 18 mm/s = 1,100 mm/min.

The upper trace (I’ll call it the “A” winding) comes from the black A6302 probe clamped around the blue wire, with the vertical scale at 200 mA/div. The current starts at 0 mA and increases after each Z-axis step pulse in the bottom trace. Unlike the situation in most scope images, the “ripple” on the trace isn’t noise. It’s a steady series of PWM pulses regulating the winding current.

The middle trace (the “B” winding) increases from -425 mA because it operates in quadrature with the A winding. The hulking pistol-shaped Tektronix A6303 current probe, rated for 100 A, isn’t well-suited to measure such tiny currents, as you can see from the tiny green stepper motor wire lying in the gaping opening through the probe’s ferrite core. Using it with the A6302 probe shows the correct relation between the currents in both windings, even if its absolute calibration isn’t quite right.

Photo 2 zooms in on the A winding current, with the vertical scale at 50 mA/div, to show the first PWM pulse in better detail. The current begins rising from 0 mA, at the rising edge of the step pulse, as the A4988 controller applies +24 V to the motor winding and reaches 110 mA after 18 µs. The controller then applies -24 V to the winding by swapping the H bridge connections. This causes the current to fall to 40 mA, whereupon it turns on both lower MOSFETs in the bridge to let the current circulate through the transistors with very little loss.

Photo 2
Zooming in on the first microstep pulse of Photo 1 shows the A4988 driver raising the stepper winding current from 0 mA as the motor starts turning. The applied voltage and motor inductance determine the slope of the current changes.

The next PWM cycle starts 15 µs later, in the rightmost division of the screen, where it rises from the 40 mA winding current set by the first pulse. It will also end at 110 mA, although that part of the cycle occurs far off-screen. You can read the details of the A4988 control algorithms and current levels in its datasheet, with the two-stage decreasing waveform known as “mixed decay” mode.

Although the H-bridge MOSFETs in the A4988 connect the motor windings directly between the supply voltage and circuit ground, the winding inductance prevents the current from changing instantaneously. The datasheet gives a nominal inductance of 4.8 mH, matching what I measured, but you can also estimate the value from the slope of the current changes.. . …

Read the full article in the May 334 issue of Circuit Cellar

Don’t miss out on upcoming issues of Circuit Cellar. Subscribe today!
Note: We’ve made the October 2017 issue of Circuit Cellar available as a free sample issue. In it, you’ll find a rich variety of the kinds of articles and information that exemplify a typical issue of the current magazine.