Formula Used
The calculator starts with thermal noise density adjusted for temperature:
N density = -174 + 10 log10(T / 290)
Effective channel noise is estimated as:
N channel = N density + 10 log10(BW × ENBW) + NF + Loss
FFT bin noise uses resolution bandwidth:
RBW = Sample rate / FFT bins × ENBW
N bin = N density + 10 log10(RBW) + NF + Loss
Displayed floor is referenced to digital full scale:
Displayed dBFS = Input noise dBm + RF gain - Full scale dBm
Dynamic measured floor is found by trimming extreme bin values. The remaining dBFS samples are averaged as power values.
How to Use This Calculator
Enter the SDR bandwidth you want to observe. Add the receiver noise figure, cable loss, RF gain, and temperature. Enter the sample rate and FFT size used by your spectrum display. Set the ENBW factor for the selected window. Rectangular windows are near 1.0. Hann windows are often near 1.5. Add a detection margin for weak signal planning.
Paste measured quiet FFT bin levels when available. The tool trims extreme samples, then calculates a dynamic floor. Press calculate. The result appears below the header and above the form. Use CSV or PDF buttons to save the output.
Example Data Table
| Use case |
Bandwidth |
Noise figure |
RF gain |
FFT bins |
Measured floor |
Approx input floor |
| Wide FM survey |
200000 Hz |
4 dB |
25 dB |
4096 |
-92 dBFS |
-120.4 dBm |
| Narrow signal hunt |
12500 Hz |
3 dB |
30 dB |
8192 |
-108 dBFS |
-135.9 dBm |
| HF weak signal |
2700 Hz |
6 dB |
35 dB |
16384 |
-118 dBFS |
-147.1 dBm |
| Airband scan |
25000 Hz |
5 dB |
22 dB |
4096 |
-101 dBFS |
-128.0 dBm |
SDR Noise Floor Planning
A software defined radio sees weak signals through the limits of bandwidth, temperature, analog gain, and digital processing. The noise floor is the level where random energy hides low power carriers. A dynamic calculator helps because an SDR rarely works under one fixed condition. Span, FFT length, window type, and averaging can change the visible floor by many decibels.
Why the Floor Moves
Thermal noise starts from kTB. At room temperature, it is about minus 174 dBm per hertz. Wider bandwidth collects more noise. A receiver noise figure adds internal noise. Cable loss and filters also raise the required input level. RF gain then moves the displayed level closer to full scale. Too little gain can bury signals in converter noise. Too much gain can cause overload and false peaks.
Using FFT Data
Spectrum displays show noise in bins. The bin width is the sample rate divided by FFT bins. Window functions widen the effective noise bandwidth. This is why the same antenna can show a different floor after changing FFT size or window settings. Averaging reduces random variation. It does not remove real noise. It only makes the estimate steadier.
Dynamic Measurement
Measured bins can improve the model. Paste quiet bin levels from a spectrum trace. The calculator trims extreme values to reduce the effect of carriers, spurs, and clicks. It then averages the remaining power values. This gives a practical dynamic floor in dBFS. The value can be converted back to input dBm when full scale power and gain are known.
Practical Interpretation
Use the displayed floor for screen reading. Use input dBm for antenna and preamp planning. Add a detection margin when estimating minimum visible signal level. A three to ten decibel margin is common for quick checks. For precision work, calibrate the SDR with a known signal source. Keep notes for gain, temperature, bandwidth, and antenna state. That makes later comparisons much more reliable. The goal is not one perfect number. The goal is a repeatable estimate that changes correctly when receiver settings change. During field tests, save exports beside screenshots. Small records reveal drift, interference, and setup errors before they become confusing. They also help tune filters for changing band conditions.
FAQs
What is SDR noise floor?
It is the visible or calculated noise level in a software defined radio. Signals below it are harder to detect without filtering, averaging, or better hardware.
Why does bandwidth change the floor?
Wider bandwidth collects more random noise power. Each ten times increase in bandwidth raises the theoretical noise by about ten decibels.
What does noise figure mean?
Noise figure shows how much extra noise the receiver adds. A lower value usually improves weak signal reception.
What is ENBW?
ENBW means equivalent noise bandwidth. Window functions spread noise across bins, so the effective bin bandwidth becomes wider than the simple FFT spacing.
Why paste measured FFT bins?
Measured bins help estimate the live display floor. They include local receiver state, gain settings, interference, and real processing effects.
Why trim measured samples?
Trimming removes extreme low and high values. This reduces bias from carriers, spurs, clicks, and occasional display spikes.
What is full scale reference?
It is the input power level that corresponds to zero dBFS after the selected gain chain. Use calibration data when possible.
Is this a calibrated lab result?
No. It is a practical estimator. For lab accuracy, calibrate the SDR, cables, gain stages, and signal source.