Calculator Inputs
Formula Used
Statistical definition
F = ⟨M²⟩ / ⟨M⟩²
F = ⟨M²⟩ / ⟨M⟩²
McIntyre local model
F = kM + (1 − k)(2 − 1/M)
F = kM + (1 − k)(2 − 1/M)
Empirical power law
F = Mˣ
F = Mˣ
Measured RMS current
F = iₙ² / (2qIₚM²B)
F = iₙ² / (2qIₚM²B)
Noise penalty
FdB = 10 log₁₀(F)
FdB = 10 log₁₀(F)
How to Use This Calculator
- Choose the model matching your available measurements.
- Enter gain, ionization data, or measured noise values.
- Select current and bandwidth units carefully.
- Add optional dark-current and amplifier-noise data.
- Set a gain sweep for model comparison.
- Press calculate and review warnings before exporting.
Example Data
| Example | Inputs | Purpose |
|---|---|---|
| Silicon APD | M = 10, k = 0.1 | McIntyre factor and penalty |
| Statistical sample | ⟨M⟩ = 10, ⟨M²⟩ = 220 | Gain-distribution calculation |
| Empirical estimate | M = 20, x = 0.3 | Power-law comparison |
| Measured system | Iₚ = 10 nA, M = 10, B = 10 MHz | Noise-current back-calculation |
Frequently Asked Questions
What is excess noise factor?
It measures additional avalanche noise caused by random multiplication gain.
Why is the factor usually above one?
Avalanche multiplication varies between carriers, increasing current fluctuations.
What does the ionization ratio mean?
It compares electron and hole ionization behavior under the selected convention.
Can I use manufacturer gain data?
Yes. Manufacturer values often improve accuracy for real device structures.
When should I use statistical mode?
Use it when mean gain and mean-square gain are known.
What causes a result below one?
Incorrect units, bandwidth, or background subtraction commonly cause that result.
Does thin-region behavior matter?
Yes. Dead-space effects can invalidate simple local-ionization assumptions.
How is decibel penalty calculated?
The calculator uses ten times the base-ten logarithm of F.
Can this replace detailed device simulation?
No. It supports estimates, comparisons, and measurement checks.