Mach Zehnder Bias Calculator

A simplified Mach–Zehnder modulator transmission can be modeled by: Pn = (1 + m·cos(Φ))/2 where Pn is normalized output power, m is visibility, and Φ is total phase.

The total phase is the sum of offset and bias phase: Φ = Δφ + φb. With half-wave voltage Vπ, the bias phase relates to bias voltage by: φb = π·Vbias / Vπ.

The small-signal slope is: dPn/dV = -(m/2)·sin(Φ)·(π/Vπ), which is maximized at quadrature (Φ ≈ π/2).

1. Pick a mode: preset, target power, or target phase.
2. Enter Vπ, plus optional visibility and phase offset.
3. Use branch index k to explore equivalent solutions.
4. Press Calculate to see results above the form.
5. Export the result using the CSV or PDF buttons.

A Mach–Zehnder modulator converts a voltage-driven phase imbalance into optical intensity. Biasing selects where the transfer curve sits: maximum power for low loss, minimum power for high extinction, or quadrature for linear modulation around mid-power. This matters in links where the receiver expects a defined average power and a stable modulation index. A small bias error can move you onto a nonlinear region and increase distortion.

Many devices show Vπ in the few‑volt range, varying with wavelength and electrode design. Visibility m approaches 1 for well-matched arms, while smaller values reduce contrast and usable slope. Static phase offset Δφ can shift with temperature and packaging stress.

The calculator uses Pn = (1 + m·cos(Φ))/2, with Φ = Δφ + φb and φb = π·Vbias/Vπ. Because cosine is periodic, adding 2πk produces equivalent bias points. The tool also reports a wrapped bias in [0, 2Vπ) for convenient instrument settings.

Quadrature targets Φ ≈ π/2, where the small‑signal slope magnitude is near maximum. This supports linear modulation and predictable gain. The reported dPn/dV scales roughly with m/Vπ, so higher visibility and lower Vπ typically improve sensitivity.

For maximum transmission, aim for Φ ≈ 0. For minimum transmission, aim for Φ ≈ π. If a nonzero Δφ exists, the required Vbias shifts to compensate it. The predicted normalized power helps verify the intended operating point after applying the bias.

In power mode, the tool solves cos(Φ) = (2Pn − 1)/m. Two phase solutions exist: +acos and −acos. They yield the same power but opposite slope signs, which is useful when system polarity or control direction is fixed.

Bias drift often comes from temperature‑dependent Δφ changes. A steep slope improves modulation efficiency but can turn small bias errors into noticeable power wander. Many links add slow feedback to lock to quadrature or a power setpoint using monitor photodiodes. Logging the wrapped bias voltage is useful when instruments have limited output range. If your controller dithers the bias, the reported slope sign can help interpret the error signal direction.

Sweep the DC bias to estimate Vπ from adjacent maxima, then estimate m from observed max/min power. Enter measured Δφ if available, calculate Vbias, and confirm by monitoring output power. Export CSV or PDF to document the bias for maintenance and repeat setups.

1) What is Vπ and why does it matter?
Vπ is the voltage that produces a π phase shift. It sets the voltage-to-phase scaling, so it directly determines required bias voltage and modulation efficiency.

2) Why do I see multiple valid bias solutions?
The transfer function is periodic. Adding 2π to the phase produces the same optical power, so multiple bias voltages are equivalent after wrapping.

3) What does visibility m represent?
Visibility models interferometer contrast. Lower m reduces extinction and slope, meaning smaller modulation depth for the same drive.

4) When should I choose quadrature bias?
Use quadrature for linear intensity modulation around mid-power. It maximizes small-signal slope, which helps analog links and many transmitter chains.

5) How do I use target power mode correctly?
Enter normalized power from 0 to 1. Choose an arccos branch to select a phase family, then adjust k to explore periodic equivalents.

6) What does the slope dPn/dV tell me?
It shows sensitivity of normalized power to bias voltage. Large magnitude implies stronger response, but it can increase sensitivity to slow drift.

7) How can I reduce long-term bias drift?
Stabilize temperature, reduce back-reflections, and consider active bias control. Starting near the computed bias speeds lock acquisition and improves stability.