Electric motors provide many functions inside today’s vehicles—be they cars, planes or military vehicles. Most are designed for simple operation—press a button and a mechanism—a seat, a window, a latch— moves in a desired direction. Some motor functions are currently computer-connected and remotely-controllable, like door latches. Future functions with controllable motors, like re-configurable seating showcased at this year’s Consumer Electronics Show, are coming quickly. These features will require computer-control and improved design over today’s simple motor operation.
One reason improved motor control is necessary is because today’s simple ON-OFF motor operation may cause problems with tomorrow’s autonomous communication. When a seat adjustment motor is turned ON, the immediate change from 0 to 12 volts (see Figure 1) causes a high in-rush of current to the motor. This in-rush current increases the electromagnetic noise the motor normally emits.
In the pre-autonomous world, where the computer network is contained within the car’s body with well-shielded wires, this hasn’t been a problem. However, a few years from now, when communication extends beyond your car to other cars within the vicinity, critical information could be lost. The antenna that receives and transmits crucial messages (like “black ice ahead”) will also be open and susceptible to unwanted electrical signals. This means the “black ice” message could be obscured by, and lost to, motor noise. (These types of EMI interference scenarios are familiar to the aerospace industry; some have resulted in deadly crashes.)
Motor turn-OFF presents other problems. Stopping a seat-adjustment motor, designed to move 300 or more pounds, requires detecting a sudden current rise at a hard-stop. The current rise signals the controller and power is shut off. Not only can this be an abrupt stop, the hard-stop itself must be either designed into the seat structure or be an additional, very strong, designed and tested, component.
Counting motor revolutions can also signal a computer to turn off a motor at a desired position. However, a motor running at full speed, whether stopped by a hard-stop wall or a count, will stop abruptly if there is no rotational slowdown.
Many of the above problems can be eliminated or greatly reduced if motor speed can be controlled. With autonomous vehicles, motor speed control can be an add-on function of the embedded computers used to automatically control features like seating reconfiguration or privacy partitions.
Electricity conservation, extending motor life, reducing motor heat are additional benefits of controlling motor speed.
PWM Controls Motor Speed
PWM, or pulse width modulation, is a technique that allows a computer to control motor speed.
A computer operates with 2, and only 2, voltage levels—a logic “1” level and a logic “0” level (5V for “1” and 0V for “0” is typical). Since no other voltage level is used, a computer cannot directly control the speed of a DC motor (which uses voltage to determine speed). So, rather than controlling voltage ‘height’, a computer can change motor speed by controlling voltage ‘width’.
With PWM, a computer controls the speed of the motor by controlling the length of the logic “1” it outputs. As the voltage pulse is fed to the motor, the motor starts its motion. Its own inertia, however, prevents it from getting to maximum speed immediately. When the pulse ends, the motor ‘coasts’ until it receives the next pulse.
This push-and-coast action causes the motor to rotate slower than if 12V was continuously applied
As an approximation, if the pulse is ON for 25% of the frequency of the pulse train, the motor “sees” 25% of the maximum voltage. Thus, the speed of the motor’s rotation is 25% of its 12V speed.
One useful speed control implementable with PWM is a ramp or gradual acceleration. A ramp allows a mechanism to smoothly get to speed, or slow to a gentle stop.
A PWM ramp is achieved by stepping up (or down) the width of the driving pulse in small increments. As a pulse width gradually increases (see Figure 3) a motor gradually increases its speed in steps. The motor’s inertia helps smooth out the voltage steps “seen” by the motor. The result is a smooth acceleration, or deceleration, much like stepper motors in robotic applications.
By combining a ramp-up with a ramp-down to a coasting speed (see Figure 4) a smooth mechanism movement, from start to finish, is possible. This PWM profile also helps reduce or eliminate problems common to abrupt ON-OFF motor operation. I found that peak in-rush current can be reduced by ramping up a motor’s turn-ON. (To optimally minimize in-rush current, you’ll need to play with the ramp-up time.) This means transmitted EMI can also be reduced. And, by ramping down toward the end of a motor’s travel, a mechanism can stop gently with little-to-no positional overshoot or strain on a hard stop.
A variation of the ramp-up-and-down PWM profile is depicted in Figure 5. This represents the PWM voltage profile for a motor that pulls a cable and keeps it open until signaled to release. A cable-pulling motor needs a high voltage, say 12V, for only a short time. After pulling, only a small voltage is needed to hold the cable, depending on the attached mechanism. By applying the least holding voltage possible, less heat will be generated thus extending motor life. When the cable is finally released and the mechanism resumes its resting or latched position, the low-to-no-voltage transition makes it possible to minimize any re-latching click.
Another good use for a ramping voltage is a motor that is periodically turned ON and OFF (see Figure 6). An intermittent massage motor is an example. Again, a ramped acceleration can reduce EMI, as well as provide a smoother, more appealing mechanism sound or tactile response.
I believe as autonomy acceptance increases, so too will customer expectations for better and safer operation. A solution like PWM, that both increases possibilities and decreases problems, helps move autonomy forward.