Error Vector Magnitude
Understanding EVM
EVM is the single most comprehensive metric for digital signal quality. It captures every impairment source in a single number, making it the standard figure of merit for transmitter and receiver evaluation.
EVM Contributions
- Noise: Adds random scatter to constellation points. EVM_noise = 1/sqrt(SNR_linear) x 100%.
- Phase noise: Rotates constellation points randomly. Dominates for closely spaced subcarriers.
- AM-AM and AM-PM: Compresses and rotates outer constellation points. PA nonlinearity.
- IQ imbalance: Gain/phase mismatch distorts the constellation shape.
- Frequency offset: Progressive rotation of constellation with time.
EVM Requirements
| Modulation | EVM Req | Approx SNR |
|---|---|---|
| QPSK | 17.5% | 15 dB |
| 16-QAM | 12.5% | 18 dB |
| 64-QAM | 8% | 22 dB |
| 256-QAM | 3.5% | 29 dB |
Frequently Asked Questions
What is EVM?
EVM measures the magnitude of the error vector between ideal and actual constellation points, expressed as a percentage. It captures all signal impairments (noise, distortion, imbalance) in one number. Lower EVM = better signal quality.
What is a good EVM?
Depends on the modulation: QPSK needs < 17.5%, 64-QAM needs < 8%, 256-QAM needs < 3.5%. For design margin, target 3-6 dB better than the requirement. A 256-QAM system should target 2% EVM to have margin.
How do you measure EVM?
EVM is measured with a vector signal analyzer (VSA). The instrument demodulates the signal, determines the ideal constellation points, and calculates the error vector magnitude. The measurement requires knowing the modulation format and symbol timing.