Wavemeter vs Optical Spectrum Analyzer: Measuring a Laser's Wavelength and Spectrum
When to use a wavemeter and when to use an optical spectrum analyzer to measure a laser: what each reports, accuracy and resolution compared in picometers and megahertz, single-mode and multimode sources, side-mode suppression, tuning and mode hops, with worked numbers at 1550 nm.
Scope
This article compares the two instruments used to measure a laser's wavelength: the wavemeter, which reports the wavelength of a narrowband source as one very accurate number, and the optical spectrum analyzer (OSA), which displays the power spectrum across a band. How each works is covered in their entries and in Optical Spectrum Analyzer Fundamentals.
In short: use a wavemeter when the question is "exactly what wavelength is this single-frequency laser at, and is it drifting?"; use an OSA when the question is "what does the spectrum look like?", meaning side modes, multiple lines, amplified spontaneous emission or noise. Many benches use both: the OSA to confirm the laser is single-mode, the wavemeter to read and lock its wavelength.
Side by side
| Wavemeter | Grating OSA | |
|---|---|---|
| Reports | One wavelength (or a few lines, on some models) | Power versus wavelength across a span |
| Typical absolute accuracy | 10⁻⁶ to 10⁻⁸ relative: 1.55 pm to 0.016 pm at 1550 nm | About ±10 pm (entry level) to ±1 pm (with internal reference) |
| Resolution of spectral features | None; it does not show a spectrum | Resolution bandwidth about 0.01 to 2 nm |
| Dynamic range | Not applicable | Tens of dB, for side modes and noise floors |
| Update rate | Several to thousands of readings per second | Sweeps of 0.1 s to tens of seconds |
| Input needed | A narrowband source with enough power for the interferometer | Works down to low powers, with sensitivity set by the resolution bandwidth and sweep time |
| Best for | Absolute wavelength, drift, tuning, locking | Side-mode suppression, multimode and broadband sources, OSNR, filter shapes |
Units: picometers and megahertz
Near 1550 nm, 1 pm corresponds to 124.8 MHz, and the 50 GHz DWDM grid spacing is 0.40 nm. A wavemeter with a relative accuracy of 10⁻⁶ is accurate to 193 MHz, or 1.55 pm; at 10⁻⁷, 19 MHz or 0.16 pm. An OSA with ±10 pm accuracy is uncertain by 1.25 GHz. The wavelength-frequency-wavenumber entry has the conversions; the wavelength calculator does them directly.
Worked example. An InGaAsP DFB laser near 1550 nm tunes at about 0.086 nm per kelvin, which is 10.7 GHz per kelvin. A wavemeter at 10⁻⁶ resolves a change of about 0.02 K; a ±10 pm OSA cannot reliably see a change smaller than about a tenth of a kelvin, and its absolute reading may be off by a similar amount. For setting a laser on a DWDM channel, or for tracking slow drift, the wavemeter is the right tool.
Single-mode and multimode sources
A wavemeter assumes a single narrow line. Given a laser with several longitudinal modes, a scanning Michelson instrument reports a power-weighted average or fails to settle, and a Fizeau instrument shows a washed-out or beating fringe pattern; some models flag this as multimode, but the number cannot be trusted. An OSA shows every mode and their relative powers. Before trusting a wavemeter reading on an unfamiliar laser, look at it once on an OSA.
The side-mode suppression ratio is an OSA measurement: the ratio of the main mode to the strongest side mode, 35 dB or more for a telecom DFB laser and 40 to 45 dB for a healthy one, measured at a resolution bandwidth fine enough to separate the modes. A wavemeter cannot measure it.
Tuning and mode hops
When a laser is swept in temperature or current, a mode hop appears as a sudden jump in wavelength. A wavemeter's fast update and fine resolution show each jump and its size directly, and its readings can feed a control loop that locks the laser to a set wavelength. An OSA sweep is usually too slow to follow a laser being tuned, but it shows whether the laser is single-mode on each side of the hop and how the side modes change.
Linewidth
Neither instrument measures the linewidth of a narrow laser. A grating OSA's resolution, at best about 0.01 nm (1.25 GHz at 1550 nm), is far wider than the 1 to 10 MHz linewidth of a typical DFB laser, and a wavemeter reports only the center. Linewidth needs a heterodyne or self-heterodyne measurement; see How to Measure Laser Linewidth.
Checking accuracy
Both instruments drift, and their readings are only as good as their last calibration. A gas absorption cell (acetylene or hydrogen cyanide lines near 1510 to 1560 nm) or a laser locked to such a line gives a known wavelength to check against. Wavemeters that work in air correct to vacuum wavelength using the index of air; make sure readings and specifications use the same convention, since the difference is about 2.7 × 10⁻⁴, or 0.42 nm at 1550 nm.
Choosing
- Absolute wavelength or frequency of a single-frequency laser: wavemeter.
- Drift, tuning curves and mode hops: wavemeter, with an OSA check that the laser stays single-mode.
- Laser locking: wavemeter feeding the controller, or a reference cell.
- Side-mode suppression, multimode lasers, LEDs, ASE: OSA.
- OSNR and DWDM channel power: OSA (see How to Measure OSNR).
- Filter and component transmission spectra: OSA with a broadband source, or a swept laser with a power meter.
- Linewidth: neither; use a heterodyne method.
References: J. L. Hall and S. A. Lee, "Interferometric real-time display of cw dye laser wavelength with sub-Doppler accuracy," Applied Physics Letters 29, 367 (1976); D. Derickson (ed.), Fiber Optic Test and Measurement (Prentice Hall, 1998).