This section provides a stable anchor for cross-references to duty cycling fundamentals across the curriculum.
By the end of this chapter, you will be able to:
If you only have 5 minutes, here’s what you need to know about duty cycling for IoT:
Bottom line: Duty cycling can extend battery life from days to years, but only if you first minimize sleep current. A 1% duty cycle with 5 mW sleep power still drains batteries quickly - the math doesn’t lie.
In one sentence: Duty cycling trades responsiveness for battery life by putting devices to sleep between sensing/transmitting events.
Remember this rule: Always minimize sleep current first (<10 µA), then optimize duty cycle. A device with 5 mW sleep power can NEVER achieve multi-year battery life, regardless of duty cycle.
Learning Resources:
Before diving into this chapter, you should be familiar with:
Imagine your smartphone being smart about when to check for new emails. Instead of checking every minute (draining battery), it learns: “My owner usually checks email at 9 AM, lunch, and 5 PM.” So it checks more often during those times and sleeps the rest of the day. That’s the essence of energy-aware operation.
For IoT devices, duty cycling means periodically waking up to perform tasks (sensing, transmitting) and then returning to a low-power sleep state. A device that wakes for 100ms every 60 seconds has a 0.167% duty cycle - meaning it spends 99.8% of its time in sleep mode!
| Term | Simple Explanation |
|---|---|
| Duty Cycle | The percentage of time a device is active vs sleeping |
| Sleep Current | The tiny amount of power used when the device is “off” (typically microamps) |
| Wake Latency | How long it takes to go from sleep to fully operational (milliseconds to seconds) |
| Deep Sleep | Very low power (nanoamps) but slow wake-up and lost RAM |
| Light Sleep | Higher power (microamps) but instant wake-up and RAM retained |
Why this matters for IoT: Battery life is the biggest complaint about IoT devices. A smart doorbell that dies every week is useless. Duty cycling extends battery life from days to years by being clever about when to use power.
Meet the Sleepy Sensor!
Imagine you have a pet hamster named “Sensor Sam” who needs to check the temperature in his cage. Instead of staying awake ALL day watching the thermometer (and getting tired and hungry!), Sensor Sam has a better plan:
The Nap-and-Check Strategy:
Why is this smart?
Real-World Example:
Your smart watch does the same thing! It doesn’t check your heart rate every single second - that would drain the battery in hours. Instead, it: - Sleeps most of the time (using almost no power) - Wakes up every few minutes to check your heart - Goes right back to sleep
Fun Math Challenge:
If Sensor Sam checks the temperature for 1 minute every hour, what percentage of the day is he awake?
Answer: 1 minute ÷ 60 minutes = 1.67% awake (that’s the duty cycle!)
The Sensor Squad says: “Sleep smart, not hard! The best sensors know when to take a power nap!”
One of the most fundamental energy-saving techniques in IoT is duty cycling - the practice of turning devices on and off in cycles rather than running continuously. The duty cycle percentage represents how much time a device spends in active mode versus sleep mode.
Calculate energy consumption and visualize wake/sleep patterns for duty-cycled IoT devices.
Legacy HTML Calculator (for browsers with JavaScript disabled):
<h4 style="margin-top: 0; color: #2C3E50;">Duty Cycle Parameters</h4>
<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(200px, 1fr)); gap: 15px; max-width: 800px;">
<label style="display: flex; flex-direction: column;">
<span style="font-weight: 600; margin-bottom: 5px;">Wake Interval (seconds):</span>
<input type="number" id="wake-interval" value="60" min="1" max="3600" style="padding: 8px; border: 1px solid #ccc; border-radius: 4px;">
</label>
<label style="display: flex; flex-direction: column;">
<span style="font-weight: 600; margin-bottom: 5px;">Active Time (milliseconds):</span>
<input type="number" id="active-time" value="100" min="1" max="10000" style="padding: 8px; border: 1px solid #ccc; border-radius: 4px;">
</label>
<label style="display: flex; flex-direction: column;">
<span style="font-weight: 600; margin-bottom: 5px;">Sleep Power (µW):</span>
<input type="number" id="sleep-power" value="10" min="0.1" max="1000" step="0.1" style="padding: 8px; border: 1px solid #ccc; border-radius: 4px;">
</label>
<label style="display: flex; flex-direction: column;">
<span style="font-weight: 600; margin-bottom: 5px;">Active Power (mW):</span>
<input type="number" id="active-power" value="50" min="1" max="1000" style="padding: 8px; border: 1px solid #ccc; border-radius: 4px;">
</label>
</div><h4 style="margin-top: 0; color: #1976D2;">Understanding Duty Cycle</h4>
<p style="margin: 10px 0;"><strong>Formula:</strong></p>
<p style="margin: 10px 0; padding: 10px; background: white; border-radius: 4px; font-family: monospace;">
Duty Cycle (%) = (Active Time / Total Period) × 100
</p>
<p style="margin: 10px 0;"><strong>Average Power:</strong></p>
<p style="margin: 10px 0; padding: 10px; background: white; border-radius: 4px; font-family: monospace;">
P<sub>avg</sub> = (P<sub>active</sub> × T<sub>active</sub> + P<sub>sleep</sub> × T<sub>sleep</sub>) / T<sub>total</sub>
</p>
<p style="margin: 10px 0; font-size: 14px; color: #555;">
<strong>Key Insight:</strong> Even small reductions in duty cycle can dramatically extend battery life.
A sensor that wakes for 100ms every 60 seconds has a 0.167% duty cycle - using 600× less energy than continuous operation!
</p>Real-World Example: A soil moisture sensor that wakes every 60 seconds for 100ms to read the sensor has a 0.167% duty cycle. If sleep power is 10µW and active power is 50mW, the average power is only 0.093mW instead of 50mW - extending battery life from 6 days to over 10 years!
This dramatic improvement comes from calculating average power across the full duty cycle.
\[P_{avg} = P_{active} \times D + P_{sleep} \times (1-D)\]
For our soil moisture sensor with duty cycle \(D = 0.00167\) (0.167%): \[P_{avg} = 50\text{mW} \times 0.00167 + 0.01\text{mW} \times 0.99833 = 0.093\text{mW}\]
This means the sensor draws only 0.093mW on average—a 538× reduction that transforms battery life from days to a decade.
Context-Aware Optimization: Smart duty cycling adjusts the wake interval based on context: - Normal conditions: Wake every 60 seconds - Rapid changes detected: Wake every 10 seconds (increased monitoring) - Nighttime/stable conditions: Wake every 300 seconds (reduced monitoring) - Critical battery (<15%): Wake every 600 seconds (emergency mode)
This adaptive approach combines duty cycling with context awareness for optimal energy management.
The Misconception: Many beginners assume that reducing duty cycle from 10% to 1% will automatically give 10x battery life improvement.
The Reality: Battery life depends on AVERAGE power consumption, not just duty cycle. The actual improvement depends on the ratio of active power to sleep power.
Quantified Example:
Scenario 1: High Active/Sleep Ratio (typical sensor)
Scenario 2: Low Active/Sleep Ratio (poorly designed device)
Key Insight: Sleep power dominates when duty cycle is very low. A device with 5 mW sleep power (common with poorly configured radios or leaky regulators) can NEVER achieve year-long battery life, regardless of duty cycle optimization.
Design Lesson: Before optimizing duty cycle, first minimize sleep current to <10 µA. Then duty cycling becomes highly effective. Context-aware systems help by identifying when to enter deep sleep (nanoamp leakage) versus light sleep (microamp leakage but faster wake).
Option A (Fixed Duty Cycling): Periodic wake-up at fixed intervals (e.g., every 60 seconds), predictable power consumption (P_avg = 0.1-1 mW for 1% duty cycle), guaranteed data freshness, simple firmware implementation, no external wake-up circuitry required
Option B (Event-Driven Wake-up): Sleep until external interrupt (PIR sensor, accelerometer threshold, comparator), near-zero sleep current (0.5-5 µA), immediate response to events, requires always-on low-power wake-up sensor (10-50 µA), potential missed events if wake sensor fails
Decision Factors: Choose fixed duty cycling for environmental monitoring where data must be logged at regular intervals regardless of changes (regulatory compliance, scientific studies), or when wake-up events are unpredictable (weather stations, air quality). Choose event-driven for motion-triggered applications (security cameras, asset tracking) where 99% of time nothing happens - sleeping at 2 µA vs periodic wake at 0.5 mW provides 250x better battery life during idle periods. Hybrid approach: Use accelerometer interrupt (20 µA always-on) to detect motion, then switch to 10-second duty cycling during activity periods, returning to pure event-driven when motion stops for 5 minutes. Real example: A door sensor using event-driven wake achieves 5-year CR2032 life vs 6-month life with 60-second polling.
Option A (Deep Sleep): 0.5-10 µA current draw, RTC and wake-up logic only, RAM contents lost (or retained in specific ultra-low-power RAM), wake-up latency 1-10 ms (oscillator startup + state restoration), requires full peripheral re-initialization
Option B (Light Sleep): 50-500 µA current draw, CPU halted but peripherals clock-gated, full RAM retained, wake-up latency 1-10 µs (immediate resume), peripherals remain configured and ready
Decision Factors: Choose deep sleep for long idle periods (>1 second between events) where 10-100x lower sleep current dominates battery life - a sensor reading every 60 seconds saves 90% battery using deep sleep despite 5ms wake penalty. Choose light sleep for rapid event response (<1ms latency required) or when re-initialization overhead exceeds idle savings - a real-time audio classifier sampling at 16kHz cannot tolerate 5ms wake latency every 62.5µs. Quantified example: At 60-second intervals, deep sleep (5 µA × 60s = 0.3 mAs) beats light sleep (200 µA × 60s = 12 mAs) by 40x, but for 100ms intervals, light sleep (200 µA × 0.1s + 50mA × 1ms wake) beats deep sleep (5 µA × 0.1s + 50mA × 5ms wake) due to amortized wake cost. Break-even point: Interval > 10x wake time favors deep sleep.
Scenario: Design a soil moisture sensor for 5-year deployment in remote vineyard using 2× AA lithium batteries (3,000 mAh @ 3V).
Given:
Calculations:
Energy budget: 3,000 mAh ÷ 1,825 days = 1.64 mAh/day = 0.068 mA average
Sampling requirement: Hourly readings (24/day)
Per-cycle energy:
- Wake + sensor: 10 mA × 2.2s = 22 mAs
- LoRa TX (12 bytes): 120 mA × 1.1s = 132 mAs
- Total active: 154 mAs per cycle
Daily active: 24 × 154 mAs = 3,696 mAs = 1.03 mAh
Sleep budget: 1.64 - 1.03 = 0.61 mAh = 2,196 mAs
Sleep time: 86,400s - (24 × 3.3s) = 86,321s
Required sleep current: 2,196 mAs ÷ 86,321s = 0.025 mA = 25 µA
Problem: Combined sleep (1 µA + 0.2 µA) = 1.2 µA is fine,
but voltage regulator quiescent current is 45 µA!
Solution: Add MOSFET power switch for voltage regulator (0.1 µA leakage)
New sleep: 1 + 0.2 + 0.1 = 1.3 µA → 0.112 mAs/day sleep
Daily total: 1.03 + 0.11 = 1.14 mAh → 2,632 days = 7.2 years ✓Result: By identifying and eliminating voltage regulator quiescent current, achieved 44% longer battery life than target.
| Factor | Deep Sleep | Light Sleep | When to Choose |
|---|---|---|---|
| Sleep Current | 0.5-10 µA | 50-500 µA | Budget <5 µA avg? → Deep |
| Wake Latency | 1-10 ms | 1-10 µs | Need <100 µs response? → Light |
| Wake Interval | >1 second | <100 ms | Long intervals favor deep |
| State Retention | RAM lost (RTC only) | Full RAM kept | Need fast context restore? → Light |
| Peripheral Init | Full re-init required | Instant resume | Complex peripherals? → Light |
Rule of thumb: If (sleep_time × sleep_current_diff) > (wake_latency_penalty × wake_frequency), choose deep sleep.
The Mistake: Optimizing MCU sleep current to 1 µA while leaving I2C pull-ups (4.7kΩ at 3.3V = 700 µA) powered continuously.
Impact: Pull-ups alone consume 700× more than MCU sleep! With 0.7 mA from pull-ups, battery life drops from 8 years to 119 days.
Fix: Use MCU GPIO with internal pull-ups (disabled in sleep), or add MOSFET switches to disconnect external pull-ups during sleep. Energy cost of switching: <1 µJ per cycle, negligible vs. always-on leakage.
Duty cycling is the foundation of low-power IoT design:
The key insight is that duty cycling effectiveness depends on achieving very low sleep current (<10 µA). Once sleep current is minimized, even small duty cycles (0.1-1%) provide dramatic battery life extensions.
| Strategy | Best For | Power Savings | Complexity |
|---|---|---|---|
| Fixed Duty Cycling | Regular sampling (weather, environment) | 10-100x | Low |
| Event-Driven | Sporadic events (motion, alerts) | 100-1000x | Medium |
| Hybrid | Variable activity patterns | 50-500x | High |
| Deep Sleep | Long idle periods (>1s) | Maximum | Medium |
| Light Sleep | Fast response (<1ms) | Moderate | Low |
Reducing duty cycle from 10% to 0.1% provides negligible benefit if sleep current is 1 mA. Always profile sleep current first and get it below 10 µA before tuning the duty cycle fraction.
Sensors, ADCs, and voltage regulators can draw significant current even when the MCU is in deep sleep. Always power down peripherals or use load switches to disconnect them during sleep, and measure total system sleep current rather than just MCU sleep current.
Deep sleep takes 100 ms to 1 s to wake up, which is unacceptable for applications requiring sub-10 ms response (alarm systems, interactive devices). Profile your wake latency requirements before choosing sleep mode.
Fixed duty cycling wakes the device even when nothing has changed, wasting energy on unnecessary sensing cycles. For applications with sporadic events (motion, button presses), event-driven wake-up can reduce active time by 10–100x compared to fixed duty cycling.
| If you want to… | Read this |
|---|---|
| Learn intelligent context-aware energy saving | ACE & Shared Context Sensing |
| Understand when to offload computation | Computation Offloading |
| Apply these strategies to real hardware | Hardware & Software Optimisation |
| Measure actual device power consumption | Energy-Aware Measurement |
| Explore energy harvesting for perpetual operation | Energy Harvesting |