In the previous article, a total of four methods were explained for current regeneration in PWM driving of DC motors. Here, losses and switching methods are explained as matters to be noted when using PWM driving.
Because PWM driving is pulse driving, the power consumed is simply the average of the power consumption in the motor voltage-applied (on) period and the power consumption in the current regeneration (off) period of a single cycle. More rigorously, as indicated by the diagram below, we can think about power consumption in three separate states, which are the voltage-applied period (red), the current regeneration period (blue), and during transitions (yellow). In general, losses in the steady state are conduction losses, and losses during on-off transitions are switching losses.
●Power consumption during voltage-applied periods
(a) shows the state of switches (MOSFETs) during voltage application. Current flows in the path in which the two MOSFETs are turned on; losses in this case are the sum of the on-resistance of the turned-on MOSFETs × the square of the current.
●Power consumption during current regeneration periods
As explained in the preceding article, there are four methods of current regeneration; the current paths differ, and so the losses are also different.
(b) and (e) are current paths in which two MOSFETs are turned on; the losses are the sum of the on-resistance of the turned-on MOSFETs × the square of the current.
In (d), the current passes through the parasitic diodes of the two MOSFETs; the losses are the current × the sum of the parasitic diode VF.
In (c), the current passes through an on MOSFET and the parasitic diode of an off MOSFET; hence the losses are the on resistance of the turn-on MOSFET × the square of the current + the parasitic diode VF of the turn-off MOSFET × the current.
●Losses during transitions
This is somewhat more complex. In general, switching losses can be determined using the following equation.
Losses during transitions =0.5×Ea×I×(tr+tf)×fsw
Ea: Applied voltage, I: Current, tr: Rise time, tf: Fall time, fsw: Switching frequency
Here tr and fr are found from the actual waveforms; it must be remembered that the actual waveforms are not necessarily simple lines as in the figure. Moreover, actual measurement of the applied voltage and the current will make possible more accurate results.
●Losses and noise during transitions
As is clear from the above equation, losses during transitions--that is, switching losses--are smaller for faster rise times tr and fall times tf. The faster the switching speed (slew rate), the smaller are the switching losses; consequently, power consumption is smaller so that efficiency is improved, but on the other hand switching noise is greater. This represents a tradeoff between efficiency and noise, and a compromise between the two must be sought out.
●Switching timing of high/low-side switches
This is control that is essential even in synchronous rectification switching regulators. Turn-on and turn-off of the high-side/low-side switch pairs of an H-bridge (in the above diagrams, Q1/Q2 and Q3/Q4) must be controlled such that there is no occurrence of a period during which the high-side switch and the low-side switch are turned on simultaneously. If both are turned on at the same time, the power supply and GND are shorted, the result is a large current, known as a shoot-through current or the like, culminating in device destruction.
In order to prevent such occurrences in which devices are simultaneously turned on, a control circuit is needed that can perform switching with a period known colloquially as "dead time", during which both switches are turned off, always inserted. However, such a dead time period results in losses, and so must be kept to a minimum; and extremely sophisticated control is necessary to achieve both the simultaneous-on prevention and high-efficiency and high-speed switching (slew rate). It is in fact difficult to configure such a control circuit as an external circuit, and a simultaneous-on prevention circuit provided in motor driver ICs is generally used.
When the PWM frequency is higher, that is, when the cycling between voltage application and current regeneration is faster, switching losses increase. This too is clear from the above equations. When there is a need to alleviate current ripples by raising the frequency, there will be a tradeoff with efficiency.
●Confirmation that a motor driver IC supports PWM driving
To state the obvious, if it is not stated that a motor driver IC "supports PWM driving", then it may be that PWM driving is not possible. If there is a pressing need to use the driver IC in PWM driving, the manufacturer must be consulted.
If a driver that does not support PWM driving is used in PWM driving, it is highly likely that the driver will not operate, and even if it does work, it will malfunction, or external components or circuits may be needed for operation. As a rule, it is best not to attempt such operation.
・Losses in PWM driving of a brushed DC motor should be considered from voltage application, current regeneration, and transition.
・When switching is made faster (the slew rate is raised), efficiency rises, but noise increases.
・If the PWM frequency is raised, current ripples can be reduced, but efficiency is lowered.