Introduction

When analyzing an RF transmitter with a spectrum analyzer, engineers often encounter unexpected spectral peaks. Some immediately label them as harmonics, others call them intermodulation products, while some simply refer to everything as spurious emissions.

Although these terms are closely related, they describe fundamentally different phenomena.

Misidentifying a spectral peak can lead engineers to troubleshoot the wrong circuit, replace perfectly good components, or waste hours chasing nonexistent problems.

This guide explains the differences between harmonics, intermodulation distortion (IMD), and spurious emissions, then introduces a practical troubleshooting workflow that experienced RF engineers use to identify the real source of an unexpected signal.

Harmonics: Integer Multiples of a Single Signal

The easiest phenomenon to recognize is the harmonic.

A harmonic is generated when a single-frequency signal passes through a nonlinear device.

Mathematically,

F_n = n × F₀

where:

• F₀= fundamental frequency
• n = harmonic order (2, 3, 4…)

For example,

Fundamental:

1 GHz

Possible harmonics:

• 2 GHz
• 3 GHz
• 4 GHz

These frequencies are exact integer multiples of the original signal.

Why Harmonics Are Generated

An ideal amplifier follows

y = ax

Real RF devices do not.

Instead,

y = a₁x + a₂x² + a₃x³ + …

The x² term generates second harmonics.

The x³ term generates third harmonics.

Almost every nonlinear RF component can generate harmonics, including:

• Power Amplifiers (PA)
• Low Noise Amplifiers operating near compression
• Mixers
• RF Switches
• PIN Diodes
• Frequency Multipliers
• DACs / ADCs
• Saturated passive components

Therefore, harmonics are not unique to power amplifiers—they are a natural consequence of nonlinear behavior throughout an RF system.

Intermodulation: Multiple Signals Meet Nonlinearity

Intermodulation (IMD) also originates from nonlinearity, but unlike harmonics, it requires two or more signals interacting inside the same nonlinear device.

Suppose:

F₁ = 900 MHz

F₂ = 901 MHz

A nonlinear component produces frequencies such as

• F₁ + F₂
• F₁− F₂
• 2F₁− F₂
• 2F₂− F₁
• 3F₁− 2F₂

Among these products, the third-order intermodulation products

• 2F₁− F₂
• 2F₂− F₁

are usually the most problematic because they fall very close to the desired signals.

For the example above,

2F₁ − F₂ = 899 MHz

2F₂ − F₁ = 902 MHz

Unlike harmonics, these products cannot easily be removed with filters because they often lie inside or adjacent to the receiver bandwidth.

This is why IM3 is usually the most critical specification for communication systems.

Passive Intermodulation (PIM) Is Still Intermodulation

Many engineers confuse IMD with PIM.

Actually,

PIM (Passive Intermodulation) is simply a special type of intermodulation.

Instead of occurring inside active devices,

PIM originates from passive components such as

• Connectors
• Cables
• Antennas
• Duplexers
• Filters
• Loose mechanical joints
• Corroded metal contacts

Both IMD and PIM result from nonlinearity.

The only difference is where the nonlinearity exists.

Spurious Signals: The Umbrella Category

Spurious emissions are often misunderstood.

Unlike harmonics or intermodulation,

Spurious is not a generation mechanism.

It is simply any unwanted spectral component that should not exist.

A spurious signal may originate from

• Harmonics
• Intermodulation
• LO leakage
• PLL fractional spurs
• Reference spurs
• Clock coupling
• Switching regulators
• Digital noise
• PCB coupling
• Power supply ripple

Therefore,

Harmonics and intermodulation are both subsets of spurious emissions.

Spurious Emissions

│

├── Harmonics

├── Intermodulation

│     ├── Active IMD

│     └── Passive IMD (PIM)

├── PLL Spurs

├── LO Leakage

├── Clock Spurs

├── Power Supply Spurs

└── Digital Coupling

Understanding this hierarchy helps engineers avoid confusing cause with classification.

Practical Spectrum Analyzer Workflow: Identifying an Unknown Spectral Peak

Knowing the theory is one thing.

Identifying the source of a real spectral peak is another.

Experienced RF engineers usually follow a systematic elimination process.

Step 1 — Is the Frequency an Integer Multiple?

Suppose the transmitter operates at 2.4 GHz.

A spectral peak appears at 4.8 GHz.

Immediately ask: Is it exactly 2 × 2.4 GHz?

If yes, there is a very high probability that it is the second harmonic.

Likewise, 7.2 GHz suggests the third harmonic.

If the answer is yes, begin checking

• amplifier linearity
• output filtering
• compression
• matching network

before investigating anything else.

Step 2 — Does It Match an Intermodulation Formula?

If two transmitters are operating simultaneously, calculate mF₁ ± nF₂, Typical combinations include

• 2F1 − F2
• 2F2 − F1
• 3F1 − 2F2

If the observed peak matches one of these frequencies, the problem is likely IMD. The next inspection points are

• PA linearity
• Mixer compression
• Shared RF path
• Duplexer isolation
• Passive junction quality (possible PIM)

Step 3 — Does the Peak Move with LO or PLL Settings?

This is one of the quickest ways to identify PLL Spurs. Change the synthesizer frequency slightly. For example, increase the LO by 1 MHz. Observe the suspicious peak. Possible outcomes:

Peak moves together with the LO

→ likely PLL spur or LO leakage.

Peak stays fixed

→ external interference or clock-related emission.

Peak changes irregularly

→ fractional-N PLL spur.

Fractional synthesizers commonly generate

• Fractional Spurs
• Reference Spurs
• Sigma-Delta Modulation Spurs

These are frequently observed around modern RF transceivers.

Step 4 — Turn Digital Circuits On and Off

Many “mysterious” spectral peaks disappear when

• FPGA clocks stop
• Ethernet disconnects
• DDR memory idles
• MCU enters sleep mode

If the peak disappears, it is probably digital coupling instead of RF nonlinearity.

Step 5 — Change Output Power

This is one of the most powerful diagnostic techniques. Increase transmitter output power by 1 dB. Observe how the spectral peak changes.

Typical behavior:

Signal Type	Power Growth
Fundamental	+1 dB
Second Harmonic	≈ +2 dB
Third Harmonic	≈ +3 dB
Third-order IMD	≈ +3 dB

This simple test immediately distinguishes many nonlinear effects.

Step 6 — Remove Components One by One

If the source is still unknown, disconnect the RF chain progressively.

Typical sequence:

Transmitter

↓

PA

↓

Filter

↓

Cable

↓

Connector

↓

Antenna

↓

Load

Watch whether the suspicious peak disappears after removing each stage.

This isolation method is standard practice during RF system debugging.

A Practical Example

Imagine a 2.4 GHz Wi-Fi module under test. Three unexpected peaks appear.

Peak A 4.8 GHz, Exactly twice the carrier.

Conclusion: Second harmonic.

Peak B

Two nearby carriers generate a new signal at 2F1 − F2.

Conclusion: Third-order intermodulation.

Peak C

A narrow spectral line appears at 2.537 GHz. It is neither an integer multiple nor an IMD product. Changing the PLL channel shifts this peak.

Conclusion: PLL fractional spur.

Best Practices for RF Engineers

When an unexpected spectral peak appears, resist the temptation to label it immediately.

Instead, follow a structured process:

1. Check whether it is an integer multiple of the carrier (harmonic).
2. Calculate whether it matches an intermodulation product.
3. Observe whether it tracks the LO or PLL frequency.
4. Disable digital subsystems to identify coupling-related emissions.
5. Measure how the peak scales with output power.
6. Isolate each RF component to locate the source.

This disciplined workflow dramatically reduces debugging time and prevents unnecessary hardware changes.

Conclusion

Although harmonics, intermodulation, and spurious emissions often appear together on a spectrum analyzer, they represent different concepts.

• Harmonicsare integer multiples of a single signal generated by nonlinear behavior.
• Intermodulationresults from multiple signals interacting inside the same nonlinear element.
• Spurious emissionsare the umbrella category for all unintended spectral components, including harmonics, IMD, PLL spurs, LO leakage, clock coupling, and power supply noise.

For RF engineers, the real challenge is not seeing a spectral peak—it’s understanding why it is there.

A systematic troubleshooting approach transforms spectrum analysis from trial-and-error into a repeatable engineering process, enabling faster root-cause identification and more reliable RF system performance.