OE-5.4 Digital Computers as Controllers & Actuators

What this is and why it exists

Putting a computer in the loop changes the problem in one fundamental way. A continuous process becomes a sampled one.

The computer does not see the process. It sees a sequence of snapshots, at whatever rate you chose, and everything between two snapshots is invisible to it. That is the design question this unit poses, and it is not obvious until stated.

The actuators at the end are worth real attention, and stepper motors most of all, because they offer position control without a position sensor.

The vocabulary

• Sampling — reading the process at discrete instants.
• Sample rate — how often that happens.
• Interface electronics — what converts between process signals and computer signals.
• Analog-to-digital conversion — turning a voltage into a number.
• Solenoid — an electromagnet producing linear or rotary motion.
• Direct-current motor — speed roughly proportional to applied voltage.
• Alternating-current motor — speed set largely by supply frequency.
• Stepper motor — moves in fixed increments, one step per command.
• Open-loop position control — commanding a position without measuring the result.

The mental model

A computer in a control loop sees samples. Between them it knows nothing. So the sample rate has to be fast relative to how quickly the process can change. Sample too slowly and the controller responds to a picture of the past, which is a good way to build an unstable loop.

The interface is the other half of the design. The computer needs the process quantity as a number, which means conversion. The process needs the computer's decision as a signal an actuator will accept. Both directions need thinking about: what range, what resolution, how often, and what happens if a conversion is wrong or late.

Then actuators. Solenoids give simple linear or rotary motion, essentially on or off. Direct-current motors give speed roughly with applied voltage and are convenient to control. Alternating-current motors are robust and their speed follows supply frequency, which is why controlling them means controlling that frequency.

Stepper motors are the ones worth understanding well. A stepper moves a fixed increment per command pulse. Count the pulses and you know where it is: position control without a position sensor. That is a genuinely attractive property, and why steppers are everywhere in printers, small machine tools and instruments.

But open-loop position control has a matching danger, and it is worth stating as plainly as the benefit. If the motor is asked for more torque than it can deliver, it skips steps. The controller keeps counting pulses and now believes a position the machine is not at. Nothing reports the error. Everything after that point is wrong, confidently, until something is homed against a physical reference.

That is the trade: no sensor, and no way to know when you have lost. Which is why serious systems still home periodically, or add a sensor after all.

What you should now be able to explain or do

Say what changes when a computer enters a control loop, and why the sample rate matters. Describe what the interface must carry in each direction. Choose between solenoid, direct-current, alternating-current and stepper actuators for a task. Explain how a stepper gives position control without a sensor. State the skipped-step failure and why nothing reports it. Say what homing is for.

Check yourself