61  Processes & Systems: Control Types

In 60 Seconds

Open-loop control executes a fixed program regardless of output (e.g., toaster timer), while closed-loop control measures output and adjusts input to maintain a setpoint (e.g., thermostat). Closed-loop costs more in sensors and complexity but handles disturbances automatically – use it whenever output accuracy matters more than component cost.

61.1 Learning Objectives

By the end of this chapter, you will be able to:

• Differentiate between open-loop and closed-loop control systems and identify their defining characteristics
• Trace the signal flow through a feedback loop from setpoint through error calculation to actuator output
• Evaluate trade-offs between control strategies (cost, complexity, accuracy, energy) for IoT applications
• Design self-regulating systems that automatically adjust actuator output to maintain a desired setpoint
• Select the appropriate control architecture for specific IoT scenarios based on precision and safety requirements

61.2 Prerequisites

Before diving into this chapter, you should be familiar with:

• Processes & Systems: Core Definitions: Understanding system decomposition and input-output transformations
• Sensor Fundamentals and Types: How sensors measure physical phenomena for feedback
• Actuators: How actuators produce physical outputs in response to control signals

61.3 Getting Started (For Beginners)

Open-Loop vs Closed-Loop: The Core Concept

This is the most important concept in control systems:

Type	How It Works	Example	Problem
Open-Loop	Set it and forget it	Toaster with timer	Burns toast if bread is thick
Closed-Loop	Continuously adjusts based on feedback	Thermostat	More complex, more reliable

61.3.1 Real-World Comparison

Control System Comparison:

Type	Device	Command	Behavior	Result
OPEN-LOOP (Dumb)	Space Heater	“Run for 2 hours”	Runs regardless of temperature	Maybe too hot or cold!
CLOSED-LOOP (Smart)	Smart Thermostat	“Keep room at 22°C”	Checks actual temp and adjusts	Always comfortable!

61.3.2 Feedback Loop Explained

The Shower Temperature Problem

Everyone has experienced this control system:

Goal: Get water at perfect temperature (38°C)

Step	Component	Action
1	Desired Temp (Setpoint)	You want 38°C
2	Valve (Actuator)	Turn valve to adjust
3	Water Temp (Output)	Water comes out
4	Your Hand (Sensor)	Feel water temperature
5	Feedback Loop	Compare to desired → Adjust valve → Repeat!

This is a CLOSED-LOOP control system—you are the controller!

61.3.3 IoT Examples of Control Systems

Example 1: Smart Irrigation

Smart Irrigation Comparison:

Type	Behavior	Problem/Benefit
OPEN-LOOP (Basic Timer)	Water lawn every day at 6 AM for 30 minutes	Waters even when it rained yesterday!
CLOSED-LOOP (Smart)	Moisture Sensor → Controller → Sprinkler Valve → Soil wetter → Feedback	Only waters when soil is actually dry! Saves 30-50% water

Example 2: Smart HVAC

Smart HVAC Control System:

Component	Details
INPUTS	Temperature, Humidity, Occupancy, Time of day, Energy price
PROCESS	HVAC Controller
OUTPUT	Comfortable Room!
FEEDBACK	Sensors report actual conditions back to controller

Flow: INPUTS → PROCESS → OUTPUT → FEEDBACK → (back to PROCESS)

Key Takeaway

In one sentence: Closed-loop control systems that measure output and adjust inputs automatically are the foundation of smart IoT - turning “dumb” timers into intelligent devices that respond to real conditions.

Remember this: If your IoT device just runs a timer without checking actual results, it is open-loop and will fail when conditions change - add a sensor and feedback to make it smart.

61.4 Feedback Control Systems in IoT

⭐⭐ Difficulty: Intermediate

Understanding feedback control systems is essential for IoT device design. Most IoT applications require systems that can automatically maintain desired states without constant human intervention—this is the domain of closed-loop control.

Knowledge Check

Test your understanding of control system concepts.

Quiz 1: Open-Loop vs Closed-Loop Control

Quiz 2: Feedback Loop Architecture

61.5 Open-Loop vs Closed-Loop Systems

The fundamental distinction in control systems is whether they use feedback to adjust their behavior:

Figure 61.1: Open-loop control system block diagram with reference input flowing through controller to plant/process with no feedback path back to controller, showing one-directional control flow.

Closed-loop control system block diagram with reference input to summing junction, error signal to controller, control signal to plant/process, sensor measuring output and feeding back to summing junction to enable self-correction.

Open-loop control (top) operates without feedback; closed-loop control (bottom) continuously measures output and adjusts to maintain the desired setpoint.

Detailed Comparison:

Feature	Open-Loop	Closed-Loop
Feedback	None—operates based on predetermined inputs	Continuous—measures output and adjusts
Accuracy	Low—cannot correct for disturbances	High—self-corrects to maintain setpoint
Complexity	Simple—fewer components, easier to design	More complex—requires sensors and control logic
Cost	Lower—no sensor feedback required	Higher—additional sensors and processing
Reliability	Susceptible to drift and external changes	Robust to disturbances and component variations
Speed	Can be fast (no feedback delays)	Slower (must wait for feedback measurements)
Energy	Less power (no continuous sensing)	More power (continuous sensing and adjustment)
IoT Example	Timer-based irrigation (waters for fixed duration)	Soil moisture-based irrigation (waters until target moisture reached)
Use Case	Predictable environments, low criticality	Variable environments, high accuracy needs

Real-World Examples:

Figure 61.2: Open-loop microwave oven example: user sets 3-minute timer, magnetron runs at fixed power, food heats without any temperature feedback — the oven cannot adjust if food cooks unevenly.

Closed-loop thermostat example: desired 22°C setpoint compared to actual sensor reading, error signal drives controller, heater output warms room, temperature sensor feeds back to close the loop and maintain setpoint automatically.

Open-loop example (top): Microwave runs for fixed duration regardless of food temperature. Closed-loop example (bottom): Thermostat continuously measures and adjusts to maintain 22°C.

Key Insight: Open-loop systems are predictable but inflexible—they cannot adapt to changing conditions. Closed-loop systems are adaptive but require more resources (sensors, processing, energy).

61.6 Real-World Example: Smart Factory Temperature Control

⭐⭐ Difficulty: Intermediate

Industry Context: A semiconductor manufacturing facility requires precise temperature control in clean rooms to maintain product quality. Even ±0.5°C variations can cause $50,000+ in defective wafer batches.

System Components:

Component	Specification	Cost	Function
Temperature Sensors	±0.1°C accuracy, PT1000 RTD	$85 each × 12	Measure zone temperatures
HVAC Actuators	Modulating dampers, 0-10V control	$450 each × 4	Regulate airflow
Edge Controller	Siemens S7-1200 PLC	$800	Execute PID control loops
Cloud Dashboard	Azure IoT Central	$200/month	Monitoring and analytics

Process Flow:

Figure 61.3: Clean Room PID Temperature Control System with Feedback Loop

Performance Metrics (Before vs After Closed-Loop Implementation):

Metric	Before (Manual Control)	After (Closed-Loop Control)	Improvement
Temperature Stability	±1.2°C variation	±0.08°C variation	93% reduction
Defect Rate	3.2% (640 defects/20k wafers)	0.3% (60 defects/20k wafers)	91% reduction
Annual Defect Cost	$960,000 ($1,500/defect)	$90,000	$870K saved
Energy Consumption	15.2 kWh/hr (over-cooling)	12.0 kWh/hr (optimized)	21% reduction
Response to Disturbance	12 minutes to stabilize	3 minutes to stabilize	75% faster
Operator Interventions	8 adjustments/shift	0.5 adjustments/shift	94% reduction

Key Insights:

1. Local Feedback is Critical: Edge PLC responds in <200ms. Cloud-only control would have 2-5 second latency—unacceptable for 5-second disturbances (door openings, equipment cycling).

2. System Decomposition: Clean separation of sensing (PT1000 RTDs), processing (PLC control algorithm), actuation (modulating dampers), and monitoring (Azure cloud) enabled modular upgrades without full replacement.

3. ROI Analysis: $12,220 total investment ($1,020 sensors + $1,800 actuators + $800 PLC + $600 cloud/3 months). Payback in 15 days from defect reduction alone ($870K annual savings ÷ 365 days = $2,384/day).

4. Feedback >> Open-Loop: Manual control (effectively open-loop with slow human feedback) caused ±1.2°C swings. Continuous sensor feedback reduced this to ±0.08°C—a 15× improvement.

Putting Numbers to It

The return on investment for closed-loop control systems is calculated by comparing the system implementation cost to the annual savings from improved performance. Payback period determines how quickly the investment recoups its cost.

\[\text{Payback Period (days)} = \frac{\text{Total Investment Cost}}{\text{Daily Savings from Improvement}}\]

Worked example: A semiconductor facility invests $12,220 in closed-loop temperature control ($1,020 sensors + $1,800 actuators + $800 PLC + $600 cloud). Defect reduction saves $870K/year. Payback period: \(\frac{\$12,220}{\$870,000/365 \text{ days}} = \frac{\$12,220}{\$2,384/\text{day}} = 5.1\) days. The system pays for itself in just 5 days, delivering 51x ROI in the first year alone.

61.7 Closed-Loop IoT System Diagram

Here’s a complete closed-loop control system architecture applied to an IoT smart thermostat:

Complete closed-loop IoT thermostat control system

Figure 61.4: Complete closed-loop control system for IoT smart thermostat showing setpoint, error calculation, PID controller, actuator (HVAC), process (room heating), sensor (thermometer), and feedback loop

System Operation:

1. Setpoint: User sets desired temperature (22°C) via app or thermostat dial
2. Comparison: Summing junction compares setpoint to measured temperature from sensor
3. Error Calculation: If room is 18°C, error = 22°C - 18°C = +4°C (too cold)
4. Controller: Calculates control signal based on error magnitude
5. Actuator: HVAC system adjusts heating/cooling power based on control signal
6. Process: Room temperature changes due to HVAC output
7. Sensor: Thermometer measures actual room temperature
8. Feedback: Measured temperature fed back to summing junction → loop repeats

Continuous Cycle: This loop executes continuously (e.g., every 30 seconds) to automatically maintain 22°C despite disturbances (door openings, sunlight, occupancy changes).

Figure 61.5: Alternative view: Side-by-side comparison of open-loop (no feedback, unpredictable results) versus closed-loop control (continuous feedback, maintains setpoint). This comparison highlights why closed-loop systems are essential for reliable IoT automation.

Common Misconception: “More Feedback = Better Control”

Misconception: Adding more feedback loops and faster sampling rates always improves system performance.

Reality: Excessive feedback can cause instability, noise amplification, and wasted energy:

Problem 1: Too-Fast Feedback (Oversampling)

• Example: Smart thermostat checking temperature every 100ms
• Issue: HVAC systems have thermal inertia (heating takes 5-10 minutes)
• Result: Controller sees no change → increases heating → overshoots → oscillates
• Fix: Match sampling rate to system dynamics (check every 30-60 seconds)

Problem 2: Sensor Noise Amplification

• Example: Soil moisture sensor with ±5% noise, sampled every second
• Issue: Small noise fluctuations trigger irrigation on/off rapidly
• Result: Valve wear, energy waste, inconsistent watering
• Fix: Add moving average filter (average last 10 readings) before control decision

Problem 3: Cascaded Loops Instability

• Example: Temperature controller → valve controller → motor controller
• Issue: Each loop tries to “correct” the others’ actions
• Result: System hunts/oscillates, never settles to stable state
• Fix: Use hierarchical control with proper time-scale separation (outer loop 10× slower than inner)

Best Practices:

• Match feedback rate to process dynamics: Slow systems (temperature) = slow feedback (10-60s), Fast systems (motor position) = fast feedback (1-10ms)
• Filter sensor noise: Use moving average, median filter, or low-pass filter before control logic
• Tune gains carefully: Start conservative (low gains), increase until system responds quickly without oscillation
• Consider energy: Each sensor reading costs power—sample only as often as needed

Real IoT Example - Smart Irrigation:

• Bad: Check soil moisture every 1 minute, water when <30%
• Good: Check every 30 minutes (averaged over 5 readings), water when <25% for 3 consecutive checks

Rule of Thumb: Feedback interval should be 5-10× faster than the system’s settling time, but no faster (diminishing returns + wasted energy).

Cross-Hub Connections

Interactive Learning Tools:

• Simulations Hub - Try control system simulations to see feedback loops in action
• Videos Hub - Watch process control demonstrations showing real controllers, feedback loops, and system stability concepts in action
• Quizzes Hub - Test your understanding of open-loop vs closed-loop systems and feedback mechanisms

Knowledge Resources:

• Knowledge Gaps Hub - Common misconceptions about control systems and feedback loops
• Knowledge Map - See how control types connect to sensors, actuators, and system architecture

Related Chapters

This Series:

• Processes & Systems: Core Definitions - What are processes and systems
• Processes & Systems: PID Control - PID controller theory and applications

Deep Dives:

• Process Control and PID - Advanced feedback control systems
• Processes Labs and Review - Hands-on implementations

Foundation:

• IoT Reference Models - System architecture layers
• Architectural Enablers - IoT technology stack

Sensing & Actuation:

• Sensor Fundamentals - Input devices
• Actuators - Output devices

Worked Example: Smart Greenhouse Temperature Control - Open vs Closed-Loop Economics

Scenario: A commercial greenhouse operation grows 10,000 tomato plants across 5,000 sq ft. Temperature must stay between 18-24°C for optimal growth. Compare open-loop timer-based heating versus closed-loop PID control.

System Components:

Component	Open-Loop System	Closed-Loop System
Heater	50 kW propane, $8,000	Same heater
Controller	$150 timer (on 6am-10pm daily)	$450 PID controller + $200 DHT22 sensors (×5)
Initial Cost	$8,150	$9,650

Operational Analysis (January - cold month):

Open-Loop Performance:

• Heater runs 16 hours/day regardless of actual temperature
• Energy consumption: 50 kW × 16 hr/day × 31 days = 24,800 kWh/month
• Propane cost: 24,800 kWh × $0.12/kWh = $2,976/month
• Temperature swings: 15-28°C (external temp drives actual temp)
• Crop stress events: 12 days below 18°C, 8 days above 24°C
• Estimated yield loss: 15% = $4,500 lost revenue (10,000 plants × $3/plant × 15%)

Closed-Loop Performance:

• PID maintains 21°C setpoint, heater runs only when needed
• Energy consumption: 50 kW × 9.2 hr/day × 31 days = 14,260 kWh/month
• Propane cost: 14,260 kWh × $0.12/kWh = $1,711/month
• Temperature stability: 20.5-21.5°C (±0.5°C)
• Crop stress events: 0
• Yield loss: 0%

Monthly Comparison:

Metric	Open-Loop	Closed-Loop	Savings
Energy Cost	$2,976	$1,711	$1,265
Yield Loss	$4,500	$0	$4,500
Total Monthly Cost	$7,476	$1,711	$5,765

ROI Calculation:

• Additional investment: $9,650 - $8,150 = $1,500
• Monthly savings: $5,765
• Payback period: 7.8 days
• Annual savings: $69,180

Key Insight: Closed-loop control paid for itself in less than 8 days through combined energy savings (42% reduction) and eliminated crop loss. The sensor feedback prevented both under-heating (crop stress) and over-heating (wasted energy).

Decision Framework: When to Use Open-Loop vs Closed-Loop Control

Decision Matrix:

Factor	Choose Open-Loop When…	Choose Closed-Loop When…
Environment Predictability	Highly predictable, controlled environment	Variable external conditions (weather, load, disturbances)
Accuracy Requirements	Tolerance ±20% acceptable	Tolerance ±2% or tighter required
System Dynamics	Slow changes, long time constants (hours/days)	Fast changes, short time constants (seconds/minutes)
Disturbance Frequency	Rare disturbances (<1 per day)	Frequent disturbances (>1 per hour)
Consequence of Error	Low cost: $10-100 per failure	High cost: >$1,000 per failure or safety-critical
Component Budget	<$50 per unit, cost-sensitive	Budget allows $100-500 for sensors and controllers
Maintenance Access	Easy manual adjustment by operators	Remote locations, minimal maintenance access
Energy Constraints	Constant energy available	Battery-powered, energy optimization required

Real-World Application Examples:

Open-Loop Acceptable:

• LED lighting on fixed schedule (sunrise/sunset timers)
• Irrigation in desert (predictable evaporation, no rain)
• Industrial ovens with consistent batch sizes
• Timed chemical dosing with stable pH buffering

Closed-Loop Required:

• HVAC systems (weather changes hourly)
• Chemical reactors (exothermic reactions, runaway risk)
• Drone flight control (wind gusts, rapid dynamics)
• Medical devices (patient variability, life-safety)

Quick Decision Test: If you can’t predict the output within ±10% for 95% of operating conditions, you need closed-loop control.

Common Mistake: “Adding More Sensors Makes Control Better”

The Mistake: Engineers assume that adding more sensors (e.g., 10 temperature sensors instead of 1) automatically improves closed-loop control accuracy.

Why It’s Wrong: More sensors create three problems:

1. Sensor Fusion Complexity: 10 sensors give 10 different readings (22.1°C, 22.3°C, 21.9°C…). Which one do you use for PID? Average? Max? Min? Incorrect fusion amplifies noise instead of reducing it.

2. Noise Amplification: Each sensor has ±0.5°C noise. Without proper filtering, the control loop sees rapid fluctuations and over-reacts. Example: Temperature “changes” from 22.0°C to 22.5°C in 1 second due to sensor noise → heater cycles on/off rapidly → wears actuator.

3. False Precision: 10 sensors with ±0.5°C accuracy don’t give you ±0.05°C precision. The fundamental sensor accuracy is still ±0.5°C. You need better sensors (±0.1°C) or proper calibration, not more sensors.

Real Example: Smart aquarium heater project added 5 thermometers “to be more accurate.” The PID controller used the average temperature. Result: when one sensor failed and read 10°C (stuck), the average dropped from 25°C to 22°C ((25+25+25+25+10)/5), triggering excessive heating. Fish tank hit 28°C before operator noticed.

The Fix:

• Use 2-3 sensors with outlier rejection: Median filter (use middle value) rejects 1 failed sensor
• Apply moving average filter: Average last 10 readings per sensor before feeding to PID
• Implement fault detection: Flag sensors that deviate >3°C from median for >30 seconds
• Start with one good sensor (±0.1°C accuracy, factory-calibrated) rather than 10 cheap sensors (±0.5°C, uncalibrated)

Rule of Thumb: For most IoT control systems, 1-2 high-quality sensors with proper filtering outperform 5-10 cheap sensors with no filtering. Add sensors for fault tolerance (redundancy), not for precision improvement.

Match Control System Features

Order the Closed-Loop Control Cycle

Key Concepts

• Open-Loop Control: A control strategy applying predetermined actuator output based on setpoint alone, without measuring the actual output — cannot compensate for disturbances but is simpler and always stable
• Closed-Loop Control: A control strategy continuously measuring process output and adjusting the actuator to minimize error — automatically compensates for disturbances and model uncertainty at the cost of potential instability
• On/Off Controller: The simplest closed-loop controller switching between fully-on and fully-off states when the process variable crosses hysteresis band boundaries — robust and simple but inherently oscillatory
• Proportional Controller: A controller producing output linearly proportional to current error: u(t) = Kp × e(t) — faster and smoother than on/off control but always has residual steady-state error
• PI Controller: A controller combining proportional and integral terms to eliminate steady-state error while maintaining acceptable transient response — the most commonly used configuration in process industry
• Derivative Kick Prevention: The technique of differentiating only the process variable (not the error) to prevent the large output spike caused by a step change in setpoint propagating through the derivative term

Common Pitfalls

1. Using Open-Loop Control for Variable-Load Applications

Applying fixed PWM duty cycle to control motor speed when load varies. Open-loop ignores speed measurement, so load changes cause speed deviation with no correction. Any application requiring consistent output under varying load needs closed-loop feedback.

2. Not Understanding Hysteresis in On/Off Control

Setting hysteresis band to zero in an on/off controller. Zero hysteresis causes infinite switching frequency at the setpoint, rapidly wearing the actuator. Always set hysteresis to the minimum practical value that prevents rapid cycling.

3. Jumping to PID When PI Suffices

Adding derivative action to a slow thermal process (furnace, HVAC system) where the derivative term only amplifies thermocouple measurement noise. For processes with large time constants and low noise, PI control is usually optimal. Add derivative only when process dynamics genuinely require it.

4. Neglecting Bumpless Transfer When Switching Control Modes

Switching from manual to automatic control without bumpless transfer — the integral term initializes to zero, causing the output to immediately jump to its proportional-only value, potentially creating a large step disturbance to the process.

Label the Diagram

💻 Code Challenge

61.8 Summary

This chapter covered the fundamentals of control types in IoT systems:

• Open-Loop vs Closed-Loop Control: Distinguishing between systems that operate without feedback (open-loop) and those that continuously measure and adjust based on output (closed-loop)
• Feedback Control Systems: Understanding how closed-loop systems use sensors to measure outputs, compare to desired setpoints, and adjust actuators to maintain target behavior
• Error Signal Calculation: Learning how error = setpoint - measured value determines control direction and magnitude
• System Stability: Recognizing that feedback must be matched to system dynamics to avoid oscillations and instability
• Noise and Filtering: Understanding that sensor noise requires filtering before use in control decisions
• Real-World Applications: Applying control concepts to smart factory temperature control, irrigation systems, and HVAC automation
• Design Trade-offs: Evaluating accuracy vs complexity, cost vs reliability, and speed vs energy consumption

For Kids: Meet the Sensor Squad!

Some systems are “dumb” and some are “smart” – the difference is whether they CHECK what’s happening!

61.8.1 The Sensor Squad Adventure: The Great Lemonade Stand

The Sensor Squad decided to run a lemonade stand at the school fair. They had two pitchers to keep full of lemonade for customers.

“I have a plan!” said Max the Microcontroller. “I’ll make a new batch of lemonade every 30 minutes, no matter what.”

For the first hour, everything was fine. But then a big group of kids came and drank ALL the lemonade in 5 minutes! Max’s plan said to wait another 25 minutes before making more. Customers left thirsty and sad.

“That’s an OPEN-LOOP system!” Sammy the Sensor pointed out. “You’re following a fixed schedule without CHECKING the actual lemonade level!”

Bella the Battery sighed. “But the opposite happened at 2 PM – nobody was buying lemonade because it got cold outside. Max made a fresh batch right on schedule, and it all went to waste!”

Lila the LED had the solution: “Let me watch the pitcher! When I see the lemonade drop below the halfway mark, I’ll flash my light. Max, you ONLY make more lemonade when you see my signal!”

They tried Lila’s plan. Now: - When the big group came, Lila flashed immediately, and Max made lemonade right away (no 30-minute wait!) - When nobody was buying, Lila stayed dark, and Max saved ingredients (no waste!)

“That’s a CLOSED-LOOP system!” Max explained. “We CHECK the result (lemonade level) and ADJUST our action (make more or wait). The FEEDBACK from Lila makes us SMART instead of just following a dumb timer!”

61.8.2 Key Words for Kids

Word	What It Means
Open-Loop	Following a plan WITHOUT checking if it’s working – like watering plants on a timer even when it’s raining
Closed-Loop	Checking the result and adjusting – like tasting soup while cooking and adding more salt if needed
Feedback	Information about what ACTUALLY happened, sent back to the decision-maker
Sensor	The “checker” that measures what’s really going on

61.8.3 Try This at Home!

The Blindfold Challenge:

1. Place a cup on a table across the room
2. Try to throw a ball into the cup WITH your eyes closed (open-loop – no feedback!)
3. Now try WITH your eyes open (closed-loop – you can see and adjust!)
4. Which way do you score more? That is the power of feedback!

61.9 What’s Next

If you want to…	Read this
Study PID control in depth	PID Control Fundamentals
Learn about feedback mechanisms	Feedback Mechanisms
Explore PID tuning methods	PID Tuning and Applications
Study core process fundamentals	Core Process System Definitions
Practice with the PID lab	PID Simulation Lab