Formula used
Definition Noise-equivalent power (NEP) is the input signal power that produces a signal-to-noise ratio of 1 in a 1 Hz output bandwidth. It is commonly reported as W/√Hz.
- From noise spectral density and responsivity:
NEP = N / Rwhere N is noise spectral density (V/√Hz or A/√Hz) and R is responsivity (V/W or A/W).
- From specific detectivity (D*, Jones = cm·√Hz/W):
NEP = √A / D*with A in cm².
- Over a measurement bandwidth:
NEPRMS(BW) = NEP · √BWwhich gives RMS noise-equivalent power across bandwidth BW (Hz).
- Minimum detectable power at target SNR:
Pmin = SNR · NEP · √BWand the estimated incident power can include coupling efficiency: Pincident = Pmin/η.
Tip: If you average measurements for τ seconds, a simple approximation is BW ≈ 1/(2τ).
How to use this calculator
- Select a method based on your available specifications.
- Enter noise spectral density and matching responsivity, or enter D* and detector area.
- Set measurement bandwidth directly, or compute it from averaging time.
- Optionally set a target SNR and coupling efficiency to estimate required incident power.
- Press Calculate. Export results using CSV or PDF buttons.
Example data table
| Scenario | Inputs | Computed outputs |
|---|---|---|
| Noise/Responsivity | Noise = 5 nV/√Hz, Responsivity = 0.8 V/W, BW = 1 Hz, SNR = 1 | NEP ≈ 6.25×10-9 W/√Hz, Pmin ≈ 6.25×10-9 W |
| Detectivity | D* = 1×1011 Jones, Area = 1 mm², BW = 10 Hz, SNR = 1 | NEP ≈ 1.00×10-12 W/√Hz, Pmin ≈ 3.16×10-12 W |
| With coupling loss | Same as above, η = 0.5 | Incident power estimate doubles: Pincident ≈ 6.32×10-12 W |
NEP detector guide
1) What NEP represents
Noise-equivalent power (NEP) summarizes detector sensitivity by linking noise to an equivalent input power. By definition, NEP is the input power that gives a signal-to-noise ratio of 1 in a 1 Hz output bandwidth. Smaller NEP means weaker signals become measurable with the same readout.
2) Connecting NEP to your readout
If you know output noise spectral density N and responsivity R, the calculator applies NEP = N/R. Example: N = 10 nV/√Hz and R = 1 V/W gives NEP = 1.0 × 10-8 W/√Hz. Keep voltage-noise with V/W, or current-noise with A/W.
3) Using specific detectivity (D*)
Datasheets often provide specific detectivity D* in Jones (cm·√Hz/W). The tool converts your area to cm2 and uses NEP = √A / D*. This helps compare devices of different sizes because D* is normalized to area.
4) Bandwidth and averaging time
Minimum detectable power depends on the effective measurement bandwidth. The RMS noise-equivalent power across bandwidth follows NEPRMS = NEP × √BW. If NEP = 2 × 10-12 W/√Hz and BW = 100 Hz, then NEPRMS = 2 × 10-11 W. For a simple boxcar average, the calculator can estimate BW ≈ 1/(2τ).
5) Target SNR and detection threshold
NEP corresponds to SNR = 1. If your application needs higher certainty, the threshold scales linearly: Pmin = SNR × NEP × √BW. Moving from SNR 1 to SNR 5 increases the required detected power by 5× for the same bandwidth.
6) Coupling efficiency and optical losses
Real optical chains lose power in windows, filters, and imperfect alignment. Coupling efficiency η adjusts incident power as Pincident = Pmin/η. For η = 0.4, you need 2.5× more incident power to deliver the same power to the detector.
7) Method selection and consistency checks
Use noise/responsivity when you have measured readout noise and in-situ responsivity. Use the D* path for fast comparisons from datasheets. If both are available, agreement (within your uncertainty) is a strong unit and area sanity check.
8) Reporting results for experiments
For traceable documentation, report spectral NEP (W/√Hz), the bandwidth used, and the resulting Pmin at your selected SNR and η. Note whether bandwidth came from filtering or averaging time, and export CSV or PDF for repeatable reporting.
FAQs
1) What is the difference between spectral NEP and NEP over bandwidth?
Spectral NEP is normalized to 1 Hz. NEP over bandwidth multiplies spectral NEP by √BW to estimate RMS noise-equivalent power within your measurement bandwidth.
2) Which noise should I enter, voltage or current?
Enter the noise type produced by your readout output. Use voltage noise with V/W responsivity or current noise with A/W responsivity. Mixing types leads to an incorrect NEP.
3) Why does detector area matter when using D*?
D* is area-normalized. The calculator uses NEP = √A / D*, so changing area changes the NEP derived from the same D* value.
4) How do I choose the bandwidth value?
Use the effective post-detection bandwidth set by filters, lock-in settings, or digital processing. If you only average for τ seconds, you may approximate BW ≈ 1/(2τ).
5) What does the target SNR option change?
NEP corresponds to SNR = 1. Setting a higher SNR scales the required minimum detectable power directly: doubling SNR doubles Pmin for the same bandwidth.
6) What is coupling efficiency η used for?
η models optical and coupling losses between the source and the active area. The calculator estimates required incident power as Pincident = Pmin/η, which increases when coupling is poor.
7) My datasheet lists NEP directly. Is this tool still useful?
Yes. You can project minimum detectable power for your bandwidth and SNR, compare readout-limited versus datasheet-limited expectations, and export consistent results for lab notes and reports.