Noise Figure Calculator
Noise figure is a key parameter in RF and microwave engineering that quantifies the degradation of signal-to-noise ratio (SNR) caused by a device or system. It's defined as the ratio of input SNR to output SNR, expressed in decibels. A lower noise figure indicates better performance, especially for weak signals.
Noise figure determines the sensitivity of receivers, the performance of low-noise amplifiers (LNAs), and the overall quality of communication links. In applications like radio astronomy, satellite communications, and wireless systems, minimizing noise figure is essential to detect weak signals. It also helps in budgeting the noise contribution of each stage in a cascaded system.
Key noise concepts:
- Noise Figure (NF): SNR degradation in dB: NF = SNR_in(dB) - SNR_out(dB)
- Noise Factor (F): Linear ratio of input SNR to output SNR: F = SNR_in/SNR_out
- Equivalent Noise Temperature (Te): Temperature of a resistor producing same noise power
- Friis Formula: Calculates total noise factor of cascaded stages
- Thermal Noise Floor: kTB, with k = 1.38×10⁻²³ J/K, T in Kelvin, B in Hz
- -174 dBm/Hz: Thermal noise power density at 290K
This calculator solves four common noise figure problems for RF engineering:
- Noise Figure (NF): Convert between NF, F, Te, and T₀
- Noise Temperature: Calculate NF from Te and T₀
- Cascaded Stages: Compute overall NF using Friis formula
- SNR Degradation: Find output SNR from input SNR and NF
The calculator provides:
- Visual noise floor indicator with SNR display
- Noise performance quality indicator (Excellent/Good/Poor)
- Complete noise parameters (F, Te, output SNR)
- Component presets for common RF devices
- Dynamic stage addition for cascaded systems
- Reference temperature adjustment (T₀ can be changed)
- Complete unit conversions (dB, linear, K, °C)
- Friis formula implementation with up to 4 stages
Common noise figures for RF and microwave components:
| Component | Typical NF (dB) | Typical Gain (dB) | Equivalent Te (K) | Applications |
|---|---|---|---|---|
| High-End LNA | 0.3 - 0.8 | 15 - 25 | 21 - 60 | Radio astronomy, satellite |
| General Purpose LNA | 1 - 2 | 15 - 20 | 75 - 170 | Wireless base stations |
| Mixer | 6 - 10 | -6 to +6 | 870 - 2600 | Downconverters |
| Attenuator (3dB) | 3.0 | -3 | 290 | Passive attenuation |
| Attenuator (10dB) | 10.0 | -10 | 2610 | Passive attenuation |
| Power Amplifier | 5 - 10 | 10 - 30 | 627 - 2610 | Transmit chain |
| Filter (passive) | Insertion loss (dB) | -IL | IL dependent | Frequency selection |
| Receiver Front-End | 3 - 8 | 20 - 40 (total) | 290 - 1540 | Overall receiver |
| NF (dB) | F (linear) | Te (K) at T₀=290K | SNR degradation | Performance |
|---|---|---|---|---|
| 0.5 | 1.122 | 35 | 0.5 dB | Excellent |
| 1.0 | 1.259 | 75 | 1.0 dB | Excellent |
| 2.0 | 1.585 | 170 | 2.0 dB | Good |
| 3.0 | 2.0 | 290 | 3.0 dB | Good |
| 4.0 | 2.512 | 438 | 4.0 dB | Fair |
| 6.0 | 3.981 | 864 | 6.0 dB | Fair |
| 8.0 | 6.310 | 1540 | 8.0 dB | Poor |
| 10.0 | 10.0 | 2610 | 10.0 dB | Poor |
Radio Astronomy: NF < 0.5 dB (cryogenic cooling)
Satellite Communications: NF 0.5-1.5 dB
Cellular Base Station: NF 1-2 dB
WiFi Receivers: NF 3-5 dB
TV Tuners: NF 5-8 dB
Test Equipment: NF depends on measurement needs
Below are answers to frequently asked questions about noise figure calculations:
The standard Friis formula assumes matched conditions. For mismatched stages, you need to consider available gain and noise parameters. Use:
F_total = F₁ + (F₂ - 1)/G_avail₁ + (F₃ - 1)/(G_avail₁·G_avail₂) + ...
Where G_avail is the available gain of the preceding stage into the actual load impedance. For exact calculations, use noise parameters (F_min, Γ_opt, rn) and source reflection coefficient. Our calculator assumes conjugate match for simplicity.
Practical approach: Most RF designs aim for good matching (VSWR < 2:1) to minimize gain uncertainty. Use S-parameters and noise parameters from datasheets for precise cascaded analysis.
Noise figure is defined at T₀=290K. If you need Te at a different physical temperature T:
Given NF at T₀: F = 10^(NF/10), Te = T₀(F-1)
At temperature T: Noise factor F(T) = 1 + (Te/T) = 1 + (T₀/T)(F-1)
NF(T) = 10·log₁₀(F(T))
Example: NF=3dB (F=2) at T₀=290K → Te=290K. At T=77K (liquid nitrogen), F(77)=1+290/77=4.77 → NF=6.8dB. So cooling increases effective NF if referenced to 290K? Actually, noise power decreases, but NF definition uses T₀. Care needed.
Note: For receiver calculations at physical temperature T, the available noise power is kTB, not related to NF. NF is a figure of merit independent of operating temperature, but the actual SNR degradation depends on T. Our calculator uses T₀=290K for standard definitions.
LNA design involves choosing a transistor with low F_min, biasing for minimum noise, and matching the input for optimum noise impedance (Γ_opt) rather than maximum gain.
| Design Step | Action | Considerations | Example |
|---|---|---|---|
| 1. Transistor selection | Choose device with low F_min at operating frequency | GaAs HEMT, SiGe, CMOS | ATF-54143 at 2GHz, F_min=0.5dB |
| 2. Bias point | Select Vds, Id for minimum noise | Trade-off with gain | Vds=3V, Id=10mA |
| 3. Input matching | Match to Γ_opt (not 50Ω) | Noise circles on Smith chart | Γ_opt=0.5∠120°, design network |
| 4. Output matching | Match for gain or stability | Usually 50Ω | Output matching for max gain |
| 5. Stability analysis | Check K > 1 | Add resistors if needed | Source/load stability circles |
| 6. Simulate/measure | Use CAD (ADS, Microwave Office) | Validate NF, gain, stability | NF < 0.8dB, Gain > 15dB |
Trade-offs: Minimum noise match often gives less than maximum gain. Some designs use feedback to simultaneously optimize noise and input match (noise measure).
The Y-factor method uses a calibrated noise source (ENR) and measures output power ratio when noise source is ON vs OFF.
- ENR (Excess Noise Ratio) known: ENR(dB) = 10·log₁₀((T_hot - T_cold)/T₀). Usually T_hot ≈ 10000K (diode), T_cold = 290K.
- Measure Y-factor: Y = P_hot / P_cold (linear) = 10^((P_hot_dBm - P_cold_dBm)/10).
- Compute noise figure: F = ENR_linear / (Y - 1), NF(dB) = ENR(dB) - 10·log₁₀(Y - 1).
- Correct for second-stage noise: If measuring system including preamp, use Friis to extract DUT NF.
Example: ENR=15dB (ENR_linear=31.62), measure P_hot=-80dBm, P_cold=-85dBm → Y_dB=5dB → Y_linear=3.162 → NF(dB)=15 - 10·log₁₀(3.162-1)=15 - 10·log₁₀(2.162)=15 - 3.35=11.65dB. Spectrum analyzers often have built-in NF measurement.
Several physical phenomena generate noise in electronic components:
| Noise Type | Physical Mechanism | Frequency Dependence | Examples | Mitigation |
|---|---|---|---|---|
| Thermal Noise (Johnson-Nyquist) | Random motion of charge carriers due to temperature | White (constant with f) | Resistors, lossy materials | Cooling, low-loss materials |
| Shot Noise | Discrete nature of charge carriers crossing potential barrier | White | Diodes, BJTs, vacuum tubes | Smoothing, high current |
| Flicker Noise (1/f) | Trapping/release of carriers in defects | 1/f | MOSFETs, carbon resistors | Large device area, chopping |
| Burst Noise (Popcorn) | Defect capture/emission | Lorentzian | Some BJTs, ICs | Clean processing |
| Avalanche Noise | Avalanche multiplication | White | Zener diodes | Avoid breakdown region |
Quantum limit: At very low temperatures and high frequencies, quantum noise becomes dominant, with minimum noise figure approaching 0dB (but not zero due to Heisenberg uncertainty).
The minimum detectable signal (MDS) of a receiver is determined by noise figure, bandwidth, and required SNR:
MDS (dBm) = -174 + 10·log₁₀(B) + NF + SNR_min
where B = bandwidth in Hz, NF in dB, SNR_min in dB.
-174 dBm/Hz is kT at 290K (k = 1.38×10⁻²³, T=290).
Example: B=1MHz (60dBHz), NF=3dB, SNR_min=10dB → MDS = -174 + 60 + 3 + 10 = -101dBm.
Noise floor: -174 dBm/Hz + NF is the input-referred noise floor. For B=1MHz, noise floor = -174+60+3 = -111dBm. To detect a signal, it must be above this floor by SNR_min.