Fullscreen HTML5Med Res (??MB)HTML4

Feature by Andrew Levido Motor Control Part 2: Brushless DC Motors Last month, we examined traditional brushed DC motors that use mechanical commutation to reverse the direction of current in the armature windings, producing sustained torque at the shaft. Now we will look at electronically commutated (brushless) DC motors and permanent-magnet synchronous motors. B Fig.1: a BLDC motor operates on the same principle as a conventional permanent-magnet brushed DC motor, but the windings are fixed to the stator and the magnets rotate with the rotor. to rotate, it is the reactive force that produces the torque. This time the rotor direction is anticlockwise, but to sustain this rotation, we still have to reverse the direction of current in the winding as the magnet poles pass the centre line. In a BLDC motor, we do this electronically, reversing the winding current every time one of the rotor poles passes a rotor position sensor at the 12 o’clock position, represented by the black dot. If we assume that the current directions shown in Fig.1 correspond to positive current, we should flip the current to negative when the south pole passes the position sensor and back to positive when the north pole passes it. If we reversed this and flipped the current negative when the north pole passes the sensor, and positive when the south pole does so, the motor would run in reverse. A single-winding motor like this is not very practical, as it will have high torque ripple. Practical BLDCs therefore use three or more windings, as shown in Fig.2. Three winding (or multiples of three) is the most common configuration; these are often called three-phase BLDC motors. The three windings, which I have coloured red, green and blue, respectively, are connected in a ‘star’ arrangement, with the common terminals of each winding marked C. The common connection is not usually brought out of the motor, so three-phase BLDC motors generally have three power terminals. Because the windings are connected in a star arrangement, current must always flow through two of them at once (for example, in through the U Australia's electronics magazine siliconchip.com.au rushless DC motors have everything in common with their brushed counterparts in terms of the method of torque production, electromechanical model and the equations that govern their operation. As a reminder, I have included the cross-section of a conventional brushed permanent magnet DC motor on the left of Fig.1. A torque is produced by the BIl force on the rotor windings acting over the radius of the armature. The force produced by the winding current flowing perpendicularly to the magnetic field of the permanent magnets is proportional to the rotor current, I, the field density, B, and the length of the winding, l. I have shown this force as blue vectors, plus the equal-but-opposing reaction force vectors in yellow. The reaction force is a product of Newton’s third law: when one object exerts a force on a second object, the second object exerts an equal and opposite force on the first. The best example I can think of is the force the 46 Silicon Chip floor exerts on the soles of your feet when you are standing. The force of gravity is pushing you down onto the floor, but at the same time the floor is pushing back with the same force. If it did not, you’d be in trouble! The commutator and brushes ensure that the motor current – coming out of the page in the left winding and into the page in the right winding – reverses as the windings pass the motor centreline. This way, the pulses of force, and therefore the torque, are always in the same direction, producing clockwise rotation. Brushless DC (BLDC) motor principles The brushless DC motor simply inverts this paradigm, as shown on the right of Fig.1. The windings are fixed to the stator and the field magnets are attached to the rotor. When current flows in the windings, a force is produced as before, but because the winding is fixed and the magnet free Fig.2: if the commutation points are determined by the rotor position, it is a BLDC motor, but if commutation is externally driven, it is a permanent-magnet synchronous motor (PMSM). terminal and out through the V terminal). With three windings, we need three position sensors set at 120° angles from each other, again shown by the black dots. In some motors, the sensors are positioned 60° apart and some additional logic is used to decode the rotor position. While the rotor is in position A, the U-phase is connected to the positive supply and the W-phase is connected to the negative. The rotor is forced to move clockwise, and as the centre of the north pole passes sensor 1, the U-phase is disconnected and the positive supply is applied to the V-phase. This state persists while the rotor moves through the 60° arc shown in diagram B until the south pole of the rotor passes sensor 3. Now the W-phase is disconnected and the negative supply is connected to the U-phase for the arc segment shown in diagram C. At the end of the C-segment, the north pole of the rotor passes sensor 2, the V-phase is disconnected and the W-phase is connected to the positive supply. This process continues, with the polarities of the windings changing every 60°, as shown in the diagram at the lower right of Fig.2. Is it really a DC motor? Before we get into the details of driving BLDC motors, we need to address the elephant in the room. One glance at the drive waveforms in Fig.2 confirms that we are absolutely driving this motor with a three-phase AC voltage – isn’t this therefore an AC motor, specifically a permanent magnet synchronous motor (PMSM), not a DC type? siliconchip.com.au The short answer is yes. There is no significant difference between the two in terms of the physical arrangement of the motor. The difference between a BLDC motor and a PMSM comes from the way you drive it. Purists might point to the shape of the back-EMF, but this difference is minor. In a motor designed as a BLDC type, the back-EMF tends to be trapezoidal in shape because the motor is meant to be driven by a square waves, while in a PMSM designed for mains operation, the magnetics are shaped so that the back-EMF is more-or-less sinusoidal. It is a subtle distinction, and in practice, it makes little difference. As an aside, I think the terms DC motor and AC motor are a bit outdated and unhelpful. In some respects, all motors are AC motors – the brushes and commutator in a conventional DC motor are converting DC to AC mechanically. Permanent magnet synchronous motors If you come at the same motor shown in Fig.2 from the PMSM perspective, you can choose to interpret the diagrams in another way. If you ignore the rotor for a moment, you can see that, if fed with the voltage waveforms on the right, the stator windings produce a magnetic field that ‘steps’ around the motor in each successive diagram. In diagram A, the field is horizontal, with the north pole pointing left. In subsequent steps, it rotates clockwise, 60° per step, always aligned with the unenergised winding. If the excitation were sinusoidal rather than stepped, the field vector would rotate smoothly around the stator once every Australia's electronics magazine cycle. This is exactly like the stator of an induction motor. Now consider the rotor. It is drawn with its magnetic field perpendicular to the stator field, the position where it experiences maximum torque. The rotor therefore is dragged around the stator, ‘locked’ to the rotating magnetic field. With a two-pole rotor such as this, the speed of rotation will be one turn for every excitation cycle. We can generalise this for any (even) number of poles Np to ω = 2πf ÷ Np in radians per second or n = 120f ÷ Np in rpm. A twopole motor fed with 50Hz mains therefore rotates at 3000RPM (157rad/s). The rotor speed is determined only by the frequency of the excitation voltage – hence the ‘synchronous’ term. The torque-speed characteristic is shown in Fig.3. As you might expect, for the nominal operating frequency fn, it is a vertical line, since the speed is unaffected by load up to the ‘pull-out torque’ (Tpo), where the rotor can no longer keep up with the rotating magnetic field and gives up trying. This is normally about 50% higher than the maximum rated continuous torque, labelled Tn in the diagram. If the torque is negative (the green part of the graph), the PMSM is acting as a synchronous generator. Given the November 2026  47 Fig.3: the torque-speed characteristic of a PMSM is a vertical line for a given AC frequency. The rotor current and therefore the torque produced is proportional to the torque angle, δ. speed equation above, it stands to reason that the only way to vary the speed of a synchronous motor is to change the frequency of the excitation. I have shown the torque-speed ‘curves’ for 25%, 50% and 75% of nominal frequency as dotted lines. The synchronous speed is reduced proportionally. The equivalent circuit looks a lot like an AC version of the DC motor’s equivalent circuit. In this case, R is the effective winding resistance and Xs is the effective inductive reactance of the stator windings. I write “effective” because in a balanced three-phase system, the magnitude of the flux produced by a given current is 1.5 times that which would be produced by a single winding. Xs is therefore 1.5 times the inductive reactance of one phase. Like a DC motor, the torque is proportional to the winding current (T = km • I), where I is the RMS value of the phase current and km is the constant of proportionality. Things get a bit more complicated from here. Firstly, we cannot ignore the inductive reactance like we did for the DC motor. In fact, it is usually much higher than the resistance, so we ignore that instead. Secondly, we have to contend with a potential phase difference between the excitation voltage and the back-EMF. This phase shift comes about because the angle between the magnetic axis of the stator and that of the field varies as the load torque changes. At low load, the angle between the rotor and the field is near zero, but as the torque increases, so does the angle. In Fig.2, it is shown at 90°, the angle it would be at maximum torque (just before pullout). This angle, δ (delta), is sometimes called the torque angle. The current flowing in the equivalent circuit is therefore the difference between the supply voltage and the back-EMF, divided by the inductive reactance and multiplied by the sine of the torque angle to account for the phase shift. Putting this all together gives us an expression T = km•([V – Eb] ÷ Xs)•sin(δ). For a given frequency, the back-EMF is fixed (because the motor is synchronous), so the only variable we can easily control is the voltage. We can actually control δ using a complex technique known as field-oriented control, but I will save that for another article, because it applies to other AC motor types too. If we apply a fixed voltage and frequency to the motor, it runs at some constant speed. As the load torque varies, the torque angle, and therefore the current, changes. One important thing this equation tells us is that, for a given torque/current, we have to reduce the applied voltage as the frequency drops. This is because the inductive reactance and back-EMF both reduce with frequency. If we don’t reduce the voltage more-orless in proportion, the current would increase and the motor would quickly burn out. Because of their rotor inertia, synchronous motors can be hard to start unless the frequency is ramped up gradually so the rotor can ‘keep up’ with the accelerating field. Industrial-sized permanent magnet synchronous motors – those intended for direct connection to the mains – usually have conductive bars embedded in the rotor and actually start as induction motors until the rotor gets near enough in speed to lock into the rotating field. BLDC/PMSM drivers Interestingly, you can drive this one type of motor in two different ways and get two different sets of performance characteristics. If you use the rotor position to determine the commutation points, the motor has all the characteristics of a DC motor that we described last week, including the equivalent circuit, equations and torque-speed characteristics. If you perform the commutation independent of rotor position, the same motor has the characteristics of a synchronous motor with the equivalent circuit, equations and torquespeed characteristics described above. This is not always made clear in either the textbooks or the application notes for motor drivers. Fig.4: the power electronics to drive three-phase PMSMs and BLDC motors is identical. You can use either voltage-mode or current-mode modulation with a variety of modulation schemes. 48 Silicon Chip Australia's electronics magazine siliconchip.com.au The power electronics required to drive these motors are identical for either case, so let’s get that out of the way first. Fig.4 shows a typical power electronics drive circuit for a threephase BLDC or PMSM. This is just a classic three-phase full bridge that is capable of driving current into the motor windings in either polarity. I don’t have much more to say about this except to note that the voltage, or current in the case of a current-mode modulator, is normally controlled by pulse-width modulation from a fixed DC supply, as shown at the right of Fig.4. I have shown two types of modulation: square modulation, which is identical to that shown in Fig.2 except for the PWM to control voltage, and trapezoidal modulation, which more closely matches the back-EMF waveform and produces smoother torque and often lower acoustic noise. You can use sinusoidal modulation or field-oriented control too. The rest of the modulator is where the difference lies. In the case of a BLDC driver, the commutation is controlled by position sensors, usually Hall-effect sensors embedded in the motor or a resolver on the shaft. The modulator input is a signal proportional to the desired operating voltage for a voltage-mode modulator, or current for a current-mode modulator. In the case of a PMSM, the modulator input is the desired frequency/ voltage. Fig.5 shows the typical control schemes for both types of drive. These are both very similar to the DC drive controllers we looked at last time, with an inner current control loop and an optional outer speed control loop. You can also add a position control loop around both to create a servo drive. The BLDC modulator requires position feedback from the motor, adding complexity, but this can also be used to provide speed feedback. You will need an additional speed sensor for a PMSM if you want to close the speed loop. If you are using a motor with rotor position sensors, you can use them for speed sensing, even though they are not being used for commutation. You can potentially derive the commutation points for a BLDC motor by looking for a zero-crossing in the voltage across the unenergised winding (assuming you have access to the winding star connection). I’m not sure siliconchip.com.au Fig.5: a BLDC modulator requires rotor position feedback, but a PMSM modulator does not. Rotor position feedback also provides speed information. Fig.6: these extracts from the L6235 data sheet show how the current-mode modulation works and how easy it is to add a speed control loop. Australia's electronics magazine November 2026  49 how practical this is, given you can get more-or-less any motor characteristic you want out of a PMSM with field-­ oriented control, with or without rotor position sensors. A practical example The L6235 BLDC motor control chip is a good example of an integrated driver designed for BLDC applications with Hall-effect position sensing. I have not personally used this chip, but I think is serves as a good example of what’s out there. It includes a three-phase Mosfet bridge that can supply 2.8A continuous (5.6A peak) with a supply in the range of 8-52V, so it is suitable for 12V, 24V or 48V motors. It includes the Mosfet drivers and most of the charge pump necessary to power the high-side drivers. You do have to provide a couple of diodes, a couple of capacitors and a resistor to complete the high-side power supply. There are three inputs for Hall-­effect or other position sensors, which can be either arranged 120° apart like those I have shown in Fig.2, or 60° apart. The IC includes the additional decoding necessary for 60° sensors. The L6235 contains a monostable multivibrator with a programmable pulse width that is connected to one of the Hall-effect sensor inputs. This provides a fixed-width pulse train at a frequency proportional to rotor speed, which is very handy for speed feedback. The Hall-effect sensors manage the commutation but the output is modulated by a current-mode controller, as shown in the diagram taken from the data sheet, at the top of Fig.6. The Mosfet bridge current is measured across the external shunt resistor labelled Rsense and is used to drive a constant off-time current controller. Recalling that only two windings in series are energised at any time – so only one low-side Mosfet is on at any given time – the voltage across the shunt will be proportional to winding current. When this rises to a level determined by the voltage on the Vref pin, the flip-flop is set and the external capacitor (Coff) begins to charge from a 5mA constant current source. ▶ 50 Silicon Chip Fig.7: BLDC and PMSM motors come in many different configurations, including ‘outrunners’, where the rotor spins outside the stator. Australia's electronics magazine During this time, the Mosfet drive is inhibited, so the winding current begins to fall as the bridge freewheels. When the voltage on Coff reaches 2.5V, the comparator resets the flipflop, and the Mosfet drive is switched on again. Coff discharges via Roff, ready for the next cycle. The current waveform is therefore triangular, with the peak-to-peak ripple determined by the programmed off-time and the peak current determined by the voltage on the Vref pin. This chip therefore incorporates the modulator and the current controller, so it is good value for money. The L6235 data sheet shows that it is easy to add a speed control loop, as shown at the bottom of Fig.6. This simple Type-I compensator has a single pole at a frequency determined by C1 and R4. I am not 100% sure what Cref1 is doing – another pole is not strictly necessary – but I suspect this network may just be to provide a bit of high-­frequency filtering at the reference input. Another nice feature of this chip is that it uses synchronous rectification during the freewheeling phase. For example, when a low-side Mosfet switches off, the current it was carrying would normally commutate to the body diode of the corresponding highside Mosfet. The L6235 switches this high-side Mosfet on after a 1µs delay, to reduce the freewheeling losses since the voltage drop due to the on-resistance of the Mosfet is much lower than that across the diode. That’s a nice touch. There are plenty of other drivers out there that suit both BLDC and PMSM applications. Many (perhaps most) include just the bridge driver and some protection circuits, leaving the production of the drive waveforms up to an external microcontroller. Motor configurations Just like with DC motors, BLDC/ PMSM motors come in an almost bewildering variety of configurations, optimised for various applications. The wound-stator type, shown at the top of Fig.7, is generally used in highspeed, high-torque applications. The rotor slots are skewed so that torque production is relatively smooth and the rotors can be small so their moment of inertia is low. I have shown the rotor with four siliconchip.com.au poles, although in reality these motors are often made with many more. The majority of electric vehicle motors are of this type, so a lot of design effort has gone into them, particularly in optimising the magnetics and efficiency. Their performance is mind-­boggling. A typical EV motor can deliver 200kW or more at the shaft, while weighing just 15kg. They usually operate from 400-800V and can draw several hundred amps per phase. Even with their high efficiency, at this power level, they require water or oil cooling. Less-demanding applications often use a salient-pole stator, like that shown in the second part of Fig.7. Each winding is replaced by a pole-pair on opposite sides of the rotor. Magnetically, this is more-or-less identical to the wound-stator design, but it is cheaper to produce. On the downside, salient-pole motors don’t make quite as good use of the space inside the motor, so they tend to be a little less efficient. Photo 1 shows a typical high-performance salient-pole BLDC motor manufactured by Maxon. This model is 32mm in diameter and has integrated drive electronics. The motor shaft is just visible behind the edge of the PCB. There is no law that states that the rotor has to be inside the stator. In many cases, it is convenient to have the stator inside the rotor – an arrangement known as an outrunner. These can be wound-rotor types, like the third diagram, but more often they have salient poles, like the fourth diagram. Out-runners are often used when the main requirement is torque rather than speed, although some are capable of pretty high speeds too. The rotor has a relatively high moment of inertia due to its larger radius, so this type of motor is mostly used in less dynamic applications, like cooling fans. This type of motor has become the go-to motor for drones because its speed and torque are well matched to propellers, and they are relatively lightweight. You also see salient-pole outrunner PMSMs used in the hub motors of e-bikes and e-scooters. All the motors I have shown here are three-phase types, but there is no reason why that always has to be the case. Fig.8 shows one common alternative configuration. This is a single-­ phase motor with two pole pairs, like that often used in small cooling fans found in electronic equipment. siliconchip.com.au The four windings are all connected in series in such a way that a positive current produces the current directions and resulting magnetic polarities as shown. A negative current would produce the opposite current direction and magnetic polarity. These motors generally have one Hall-effect position sensor (the black dot) that drives the commutation. The current direction is reversed every time one of the four permanent magnet poles passes the sensor. If the rotor is moving, its moment of inertia ensures that it passes each pole and triggers the sensor so it is attracted to the next pole in turn. The challenge with single-phase motors like this one is starting, or at least starting them in the right direction. In the position shown in the figure, there would be no torque on the rotor if the poles were symmetrical. A symmetrical pole has an even distribution of flux density across it, so both sides of the winding see the same BIl force and they would cancel out. This would happen for all four poles, so no net torque would be produced. This type of motor therefore has the asymmetrical poles you can see in the diagram. The flux density across each pole is slightly stronger on the thicker side. This means there is slightly more BIl force on one side of each winding in all four poles. You can use the right-hand palm rule to satisfy yourself that these net forces all point counterclockwise with the current flowing in either direction, so there is always a small clockwise reactive torque bias on the rotor. This is enough to get the rotor moving in the right direction. The asymmetry means that this type of motor can only run in one direction, but that is not a problem for cooling fans. These motors are a wonder of lowcost design. Photo 2 shows the stator of a motor taken from a 40mm fan that gave its life in the interests of science. You can clearly see the asymmetric salient poles and the windings, which are just one length of enamelled copper wire wound around each pole in succession. At the top of the motor is the combined Hall-effect sensor and driver chip. This has just four pins, two for the incoming power and two for the stator windings. That is all there is on the single-­sided phenolic PCB – there Australia's electronics magazine Photo 1: this is a nice example of a salient-pole BLDC motor with integrated drive electronics. Source: Maxon International AG. Photo 2: this wonder of low-cost engineering is the stator from a 40mm cooling fan. Note that one end of each stator pole is slightly thicker than the other. Fig.8: single-phase BLDC motors, such as those used in cooling fans, have asymmetric poles to provide a small torque bias. This ensures they start reliably and spin in the right direction. is not even a decoupling capacitor in sight! Brushless DC and permanent-­ magnet synchronous motors are rapidly taking over from brushed DC motors – and that’s no surprise given their versatility. One motor type that can be used for anything from a powerful and advanced EV motor to a superlow cost fan motor. Next month, I will look at stepping or stepper motors – a very popular motor for position control applications, like in many 2D and 3D printers. They operate on a different principle: SC that of variable reluctance. November 2026  51