10  Cellular Spectrum for IoT

In 60 Seconds

Cellular IoT technologies (NB-IoT, LTE-M, 5G RedCap) operate on licensed spectrum, providing carrier-grade reliability and wide-area coverage. Low-band frequencies (700-900 MHz) offer the best range and building penetration for IoT, while multiple access techniques have evolved from FDMA to OFDMA, improving spectral efficiency by over 100x.

Key Concepts

• Cellular Spectrum: Radio frequency bands allocated by government regulators to mobile network operators for cellular IoT communication
• Licensed Spectrum: Frequency bands requiring regulatory licenses (e.g., 700 MHz, 850 MHz, AWS bands); provides interference protection
• LTE Band: A specific frequency range and duplex arrangement used by LTE networks; IoT devices must support device-specific bands
• NB-IoT: Narrowband IoT operating in 180 kHz bandwidth within LTE spectrum; optimized for low-data-rate IoT with deep indoor penetration
• CAT-M1 (LTE-M): LTE Category M1, supporting up to 1 Mbps for IoT with mobility and voice support
• LPWAN Spectrum: Low-Power Wide-Area Network frequencies used by LoRaWAN (ISM bands) and Sigfox; unlicensed but shared
• Spectrum Refarming: Repurposing older 2G/3G spectrum for modern NB-IoT and LTE-M deployments
• Roaming and IoT: Cellular IoT device behavior when operating in a different operator’s coverage area

10.1 Introduction

⏱️ ~8 min | ⭐⭐ Intermediate | 📋 P08.C16C.U01

While unlicensed spectrum (Wi-Fi, Bluetooth, LoRa) dominates many IoT deployments, cellular IoT technologies like NB-IoT and LTE-M leverage licensed spectrum to provide wide-area coverage with carrier-grade reliability. This chapter explores how cellular spectrum has evolved and why specific bands matter for IoT applications.

Learning Objectives

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

• Outline the evolution of cellular spectrum from 1G to 5G and identify which generations support IoT
• Differentiate multiple access techniques (FDMA, TDMA, CDMA, OFDMA) by their spectral efficiency and design trade-offs
• Compare low-band, mid-band, and mmWave frequencies for IoT range, penetration, and bandwidth requirements
• Justify why NB-IoT and LTE-M operate in specific cellular bands based on link budget analysis
• Evaluate cellular vs unlicensed spectrum trade-offs when selecting connectivity for an IoT deployment

Sensor Squad: The Cellular Tower Adventure

“Have you ever wondered how your phone connects to the internet when there is no Wi-Fi?” asked Max the Microcontroller. “It uses cellular networks – the same ones that IoT devices like NB-IoT and LTE-M use!”

Sammy the Sensor looked up at a cell tower. “But those towers are for phones. Can tiny sensors use them too?” Max nodded enthusiastically. “Absolutely! Cellular companies set aside special slices of their radio spectrum just for IoT devices. Low-band frequencies around 700 to 900 MHz are perfect – they travel far and penetrate buildings, so a sensor in a basement can still reach the tower.”

Bella the Battery asked, “Does using cellular drain my energy like it does on a phone?” Lila the LED explained, “Not with IoT-optimized cellular! NB-IoT uses just 180 kHz of bandwidth – a tiny sliver compared to a phone’s 20 MHz. That means way less power needed. Bella, you could last ten years on a single charge while sending data through the cellular network!”

“The big advantage,” Max concluded, “is that licensed spectrum means no interference from neighbors. Unlike Wi-Fi where everyone shares the same channels, cellular gives you guaranteed quality of service. You pay for a subscription, but your data always gets through.”

10.2 Prerequisites

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

• Electromagnetic Waves and Spectrum Basics: Understanding of EM wave properties
• IoT Frequency Bands and Licensing: Licensed vs unlicensed spectrum concepts

Related Chapters

Series Navigation:

• Previous: IoT Frequency Bands and Licensing
• Next: Propagation and Design

Cellular IoT Deep Dives:

• Cellular IoT Fundamentals - NB-IoT, LTE-M architecture
• NB-IoT Fundamentals - Deep dive into NB-IoT spectrum and power modes
• LPWAN Fundamentals - Comparison with unlicensed alternatives

---

10.3 Cellular Spectrum Allocation for IoT

⏱️ ~15 min | ⭐⭐ Intermediate | 📋 P08.C16C.U02

10.3.1 Evolution of Cellular Spectrum

The cellular industry has evolved through multiple generations, each requiring new spectrum allocations and improving spectral efficiency. Understanding this evolution is critical for IoT designers choosing between cellular IoT technologies (NB-IoT, LTE-M, 5G) and unlicensed alternatives (LoRaWAN, Sigfox).

Figure 10.1: High-level evolution of cellular generations and spectrum usage (1G→5G), including where LTE‑M/NB‑IoT fit.

10.3.2 Cellular Frequency Band Allocations

The following table summarizes representative cellular bands and typical channel bandwidths by generation (actual allocations vary by region and operator):

Generation	Example bands (varies)	Typical channel bandwidth	Multiple access (typical)	IoT Relevance
1G	~800–900 MHz	~30 kHz channels (analog)	FDMA (AMPS)	Historical only
2G	900 / 1800 / 1900 MHz	200 kHz carriers (GSM)	TDMA (GSM)	Legacy M2M modules (sunset in many regions)
3G	850 / 900 / 1700 / 1900 / 2100 MHz	5 MHz carriers (WCDMA)	CDMA/WCDMA	Being phased out
4G/LTE	700–2600+ MHz	1.4–20 MHz carriers	OFDMA	NB-IoT, LTE-M (in LTE bands)
5G NR (FR1)	600 MHz–6 GHz	5–100 MHz carriers	OFDMA	RedCap, mMTC (coverage-focused)
5G NR (FR2)	~24–52 GHz	50–400 MHz carriers	OFDM + beamforming	High throughput, limited range

Key Insight: More Spectrum vs Better Efficiency

The wireless industry faces a fundamental choice: 1. Get more spectrum - Acquire new frequency bands (expensive, limited availability) 2. Improve spectral efficiency - Send more bits per Hz (technical innovation)

Each cellular generation has pursued BOTH strategies, but spectral efficiency improvements provide greater long-term value.

10.3.3 Spectral Efficiency Evolution: FDMA → TDMA → CDMA → OFDMA

The evolution of multiple access techniques shows how the industry has dramatically improved spectral efficiency—the number of bits transmitted per second per Hz of bandwidth.

Note: The spectral-efficiency figures below are illustrative; real-world efficiency depends heavily on modulation/coding, MIMO, and radio conditions.

Figure 10.2: Multiple access methods (FDMA/TDMA/CDMA/OFDMA) and how they share spectrum among users and devices.

How Each Technique Works:

1. FDMA (Frequency Division Multiple Access) - 1G
   • Divides available spectrum into fixed frequency channels (30 kHz each)
   • One user occupies one channel for entire call duration
   • Simple but highly inefficient use of spectrum
   • Spectral Efficiency: ~0.03 bits/s/Hz
2. TDMA (Time Division Multiple Access) - 2G GSM
   • Combines frequency and time division
   • Each 200 kHz channel divided into 8 time slots
   • 8 users share one frequency channel
   • Spectral Efficiency: ~0.5 bits/s/Hz (16× improvement over FDMA)
3. CDMA (Code Division Multiple Access) - 3G
   • All users transmit on same frequency simultaneously
   • Each user assigned unique spreading code
   • Receivers filter out other users using correlation
   • Spectral Efficiency: ~1-2 bits/s/Hz with soft capacity limits
4. OFDMA (Orthogonal Frequency Division Multiple Access) - 4G/5G
   • Thousands of orthogonal subcarriers (15 kHz each in LTE)
   • Dynamic allocation based on channel conditions
   • Enables massive MIMO and spatial multiplexing
   • Spectral Efficiency: ~5-10 bits/s/Hz (or higher with MIMO)

Putting Numbers to It

How many NB-IoT sensors can one 20 MHz LTE carrier support compared to legacy 2G GSM?

Scenario: Smart city deployment with 10,000 parking sensors sending 100-byte status updates every 5 minutes.

2G GSM (TDMA - 200 kHz carriers):

Each 200 kHz carrier supports 8 time slots. At ~13 kbps per slot:

\[\text{Data Rate per Carrier} = 8 \times 13 \text{ kbps} = 104 \text{ kbps}\]

Sensor data per hour: \((100 \text{ bytes} \times 8 \text{ bits}) \times (60/5) = 9,600 \text{ bits/hour} = 2.67 \text{ bps}\)

\[\text{Sensors per Carrier} = \frac{104,000 \text{ bps}}{2.67 \text{ bps}} = 38,950 \text{ sensors}\]

Total 20 MHz spectrum: \(\frac{20,000 \text{ kHz}}{200 \text{ kHz}} = 100 \text{ carriers}\)

NB-IoT (OFDMA - 180 kHz per NB-IoT carrier):

Each NB-IoT carrier supports ~250 kbps uplink (coverage extension mode):

\[\text{Sensors per NB-IoT Carrier} = \frac{250,000 \text{ bps}}{2.67 \text{ bps}} = 93,600 \text{ sensors}\]

Total NB-IoT carriers in 20 MHz: \(\frac{20,000 \text{ kHz}}{180 \text{ kHz}} = 111 \text{ carriers}\)

Capacity Comparison: NB-IoT supports \(111 \times 93,600 = 10.4 \text{ million sensors}\) vs GSM’s \(100 \times 38,950 = 3.9 \text{ million}\) — 2.7× improvement from better spectral efficiency plus dynamic resource allocation.

10.4 Trade-offs: Range vs Bandwidth vs Penetration

⏱️ ~12 min | ⭐⭐ Intermediate | 📋 P08.C16C.U03

The choice of cellular frequency band involves fundamental physics trade-offs that directly impact IoT deployment strategies:

Figure 10.3: Frequency trade-offs for IoT: lower bands improve range/penetration; higher bands improve bandwidth but increase path loss.

Quantitative Trade-offs:

Note: These are order-of-magnitude examples; real-world cells vary widely with spectrum, site density, terrain, and network load.

Parameter	Low Band (700-900 MHz)	Mid Band (2.5 GHz)	mmWave (28 GHz)
Cell Radius	10-30 km	1-5 km	100-300 m
Building Penetration	Excellent (through concrete)	Good (through drywall)	None (blocked by glass)
Bandwidth per Carrier	5-10 MHz	20-100 MHz	100-800 MHz
Typical Data Rate	5-20 Mbps	50-300 Mbps	1-10 Gbps
IoT Suitability	Excellent	Good	Limited

For Beginners: Why Cellular Frequencies Matter for IoT

Think of radio frequencies like different types of roads:

Low frequencies (700-900 MHz) are like country highways: - They travel far with few obstacles - Not many lanes (limited data capacity) - Great for reaching remote areas - IoT example: NB-IoT sensors in rural water treatment plants

Mid frequencies (2-3 GHz) are like city roads: - Good balance of reach and capacity - Can handle moderate traffic - Work well in urban environments - IoT example: LTE-M connected cars in cities

High frequencies (mmWave) are like high-speed express lanes: - Extremely fast but very short - Blocked by almost everything - Only work with direct line of sight - IoT example: Factory robots with guaranteed clear paths (rare for IoT)

The key insight: Most IoT devices benefit from lower frequencies because they need: - Wide coverage (sensors spread over large areas) - Building penetration (indoor sensors) - Low power (small batteries)

This is why NB-IoT and LTE-M typically operate in the 700-900 MHz bands when available, even though 5G mmWave offers much higher speeds.

10.4.1 Implications for IoT Deployments

The cellular spectrum landscape directly impacts IoT technology selection:

NB-IoT (Narrowband IoT)

• Uses just 180 kHz bandwidth (one LTE resource block)
• Typically deployed in 800-900 MHz bands for maximum coverage
• +20 dB link budget improvement over LTE
• Ideal for: Deep indoor sensors, smart meters, remote agriculture

LTE-M (LTE Cat-M1)

• Uses 1.4 MHz bandwidth
• Supports voice and higher data rates than NB-IoT
• Deployed across multiple LTE bands (700, 850, 1700, 1900 MHz)
• Ideal for: Asset trackers, wearables, connected vehicles

5G IoT (RedCap/NR-Light)

• Reduced capability 5G for IoT applications
• Lower complexity than full 5G but higher than LTE-M
• Expected to use FR1 bands (sub-6 GHz)
• Ideal for: Industrial IoT, video surveillance, AR glasses

Related: Cellular IoT Deep Dive

For detailed coverage of cellular IoT technologies, see:

• Cellular IoT Fundamentals - NB-IoT, LTE-M architecture and comparison
• NB-IoT Fundamentals - Deep dive into NB-IoT spectrum and power modes
• LPWAN Fundamentals - Comparison with unlicensed alternatives

10.5 Interactive: Cellular IoT Band Comparison

viewof cellTech = Inputs.select(
  ["NB-IoT (180 kHz)", "LTE-M (1.4 MHz)", "5G RedCap (20 MHz)"],
  {label: "Technology", value: "NB-IoT (180 kHz)"}
)

viewof cellBand = Inputs.select(
  ["700 MHz (Band 28)", "800 MHz (Band 20)", "900 MHz (Band 8)", "1800 MHz (Band 3)", "2600 MHz (Band 7)"],
  {label: "Frequency Band", value: "800 MHz (Band 20)"}
)

viewof cellRadius = Inputs.range([0.5, 30], {label: "Cell Radius (km)", step: 0.5, value: 5})

cellBandMap = ({
  "700 MHz (Band 28)": 700,
  "800 MHz (Band 20)": 800,
  "900 MHz (Band 8)": 900,
  "1800 MHz (Band 3)": 1800,
  "2600 MHz (Band 7)": 2600
})

cellTechSens = ({
  "NB-IoT (180 kHz)": -164,
  "LTE-M (1.4 MHz)": -156,
  "5G RedCap (20 MHz)": -110
})

cellFreq = cellBandMap[cellBand]
cellSens = cellTechSens[cellTech]
cellFSPL = 20 * Math.log10(cellRadius) + 20 * Math.log10(cellFreq) + 32.45
cellTxPower = 23
cellMargin = cellTxPower - cellSens - cellFSPL

html`<div style="background: linear-gradient(135deg, #2C3E50 0%, #3498DB 100%); color: white; padding: 1.2rem; border-radius: 8px; font-family: Arial, sans-serif; margin: 0.5rem 0;">
<div style="display: grid; grid-template-columns: repeat(auto-fit, minmax(160px, 1fr)); gap: 1rem;">
  <div style="text-align: center;">
    <div style="font-size: 0.85rem; opacity: 0.85;">RX Sensitivity</div>
    <div style="font-size: 1.3rem; font-weight: bold;">${cellSens} dBm</div>
  </div>
  <div style="text-align: center;">
    <div style="font-size: 0.85rem; opacity: 0.85;">FSPL at ${cellRadius} km</div>
    <div style="font-size: 1.3rem; font-weight: bold;">${cellFSPL.toFixed(1)} dB</div>
  </div>
  <div style="text-align: center;">
    <div style="font-size: 0.85rem; opacity: 0.85;">Max Coupling Loss</div>
    <div style="font-size: 1.3rem; font-weight: bold;">${(cellTxPower - cellSens).toFixed(0)} dB</div>
  </div>
  <div style="text-align: center;">
    <div style="font-size: 0.85rem; opacity: 0.85;">Link Margin</div>
    <div style="font-size: 1.3rem; font-weight: bold; color: ${cellMargin > 20 ? '#16A085' : cellMargin > 10 ? '#E67E22' : '#E74C3C'};">${cellMargin.toFixed(1)} dB</div>
  </div>
</div>
<div style="margin-top: 0.8rem; font-size: 0.8rem; opacity: 0.75; text-align: center;">TX power: ${cellTxPower} dBm (UE max) | Free-space only -- add 15-30 dB for real-world losses</div>
</div>`

10.6 Worked Example: Smart Water Utility — NB-IoT vs LTE-M Band Selection

Scenario: AquaMetrics, a water utility in southern Germany, deploys 25,000 smart water meters across a mixed urban/rural service area (1,200 km²). Meters are installed in underground vaults (1.5 m below ground) and basement meter rooms. The utility needs daily meter readings with 99.5% delivery reliability over a 15-year device lifetime.

The Decision: NB-IoT on Band 20 (800 MHz) vs LTE-M on Band 3 (1800 MHz) — both available from Deutsche Telekom.

10.6.1 Link Budget Analysis

The underground vault scenario is the critical constraint. Each wall/floor layer adds attenuation:

Path Component	Band 20 (800 MHz)	Band 3 (1800 MHz)
Free space path loss (2 km cell radius)	98.5 dB	105.5 dB
Ground penetration (1.5 m soil, wet)	15 dB	22 dB
Concrete vault wall (20 cm)	8 dB	12 dB
Total path loss	121.5 dB	139.5 dB
Max coupling loss (3GPP spec)	164 dB (NB-IoT)	156 dB (LTE-M)
Link margin	42.5 dB	16.5 dB

NB-IoT on 800 MHz provides 26 dB more link margin — the difference between 99.8% and 87% delivery reliability in underground installations.

10.6.2 Cost Comparison (25,000 meters, 15-year lifecycle)

Cost Component	NB-IoT (Band 20)	LTE-M (Band 3)
Module cost per meter	EUR 4.50	EUR 6.80
Subscription (per device/month)	EUR 0.35	EUR 0.75
Battery (15-year target)	1x D-cell lithium (EUR 3.20)	2x D-cell lithium (EUR 6.40)
External antenna (for vaults)	Not needed (sufficient margin)	EUR 12.00 per vault meter (8,000 units)
Hardware total	EUR 192,500	EUR 426,000
Subscription total (15 yr)	EUR 1,575,000	EUR 3,375,000
Lifecycle total	EUR 1,767,500	EUR 3,801,000

Why AquaMetrics Chose NB-IoT on 800 MHz

Three factors drove the decision:

1. Underground reach: 42.5 dB link margin eliminated the need for external antennas in 8,000 vault installations, saving EUR 96,000 in hardware and significant installation labor
2. Battery life: NB-IoT’s 180 kHz bandwidth and PSM/eDRX power modes achieve 18-year projected battery life vs 9 years for LTE-M, avoiding a costly mid-life battery replacement campaign
3. Subscription cost: At EUR 0.35/device/month vs EUR 0.75, the 15-year subscription savings alone (EUR 1.8M) exceeded the entire NB-IoT hardware budget

LTE-M would be preferred if the utility also needed firmware-over-the-air updates (higher throughput) or voice capability for emergency alerts — neither was required for daily meter reads of 50-byte payloads.

Common Pitfalls

1. Ignoring Regional Band Differences

A cellular IoT device certified for North American LTE bands (Band 4, 12, 13) will not work in Europe (Band 3, 7, 20). Always specify band support for target deployment regions. Multi-band modems cost more but are essential for global deployments.

2. Assuming NB-IoT and LTE-M Are Interchangeable

NB-IoT and LTE-M serve different use cases. NB-IoT (180 kHz) provides the deepest indoor penetration and lowest cost for static sensors. LTE-M (1.4 MHz) supports mobility and higher throughput. Choosing NB-IoT for a moving vehicle tracker or LTE-M for a static meter with 10-year battery life are common mismatches.

3. Not Accounting for Spectrum Refarming Impact

Operators regularly retire 2G/3G networks and repurpose their spectrum. Devices using 2G (GPRS) or 3G fallback may lose connectivity when the operator sunsets these networks. Design IoT deployments with LTE-only modems to avoid this obsolescence risk.

4. Confusing Carrier Aggregation With Dual-Band Support

Carrier aggregation combines multiple LTE bands simultaneously for higher throughput. Most IoT applications do not benefit from carrier aggregation but pay for the more expensive chipset. Specify single-band or dual-band IoT modems appropriate for your throughput requirements.

🏷️ Label the Diagram

💻 Code Challenge

10.7 Summary

This chapter covered cellular spectrum allocation for IoT:

• Cellular spectrum evolution from 1G (800-900 MHz) to 5G (sub-6 GHz and mmWave) reflects increasing bandwidth and spectral efficiency
• Multiple access techniques evolved from FDMA (0.03 bits/s/Hz) to OFDMA (5-10 bits/s/Hz)
• Low-band cellular (700-900 MHz) is optimal for IoT due to range and penetration
• NB-IoT and LTE-M leverage existing LTE infrastructure in IoT-friendly bands
• 5G mmWave offers high bandwidth but is unsuitable for most IoT applications due to limited range

10.8 What’s Next

Chapter	Focus	Why Read It Next
Propagation and Design	Path loss models, interference mitigation, band selection framework	Apply the spectrum trade-offs from this chapter to real-world propagation calculations and deployment planning
Cellular IoT Fundamentals	NB-IoT, LTE-M, and 5G RedCap architecture comparison	Examine the protocol-level details of the cellular IoT technologies introduced here
NB-IoT Fundamentals	NB-IoT spectrum usage, power saving modes, coverage classes	Analyse NB-IoT’s 180 kHz narrowband design and how PSM/eDRX achieve multi-year battery life
LPWAN Fundamentals	LoRaWAN, Sigfox, and cellular LPWAN comparison	Evaluate licensed cellular spectrum against unlicensed LPWAN alternatives for your IoT use case
EM Waves and Spectrum Basics	Electromagnetic wave properties and propagation fundamentals	Revisit the physics foundations if any frequency or path-loss concepts need reinforcement

---

10.9 Knowledge Check

Quick Check: Cellular Spectrum