RF Utilities
Receiver sensitivity and minimum detectable signal for RF systems integration. Compute the MDS from bandwidth, noise figure, and required SNR, audit a datasheet sensitivity, add spread-spectrum processing gain, and turn the threshold into detection margin and range.
Free to use, no sign-up needed.
Walkthrough
Sensitivity mode with the thermal floor, receiver floor, and MDS from the bandwidth, noise figure, and required SNR.
Overview
The minimum detectable signal is the single most important number about a receiver: the weakest signal at its input that still produces a usable output. It is the receiver noise floor plus the signal to noise ratio the demodulator needs, and the noise floor is the thermal noise in the channel raised by the receiver noise figure. Written out, MDS equals −174 dBm/Hz plus 10·log₁₀ of the bandwidth plus the noise figure plus the required SNR. Every term matters, and the mistakes come from dropping one of them or mixing up the bandwidth.
The noIM₃ Minimum Detectable Signal Calculator is the sensitivity utility for that work, written for a systems integrator rather than a circuit designer. The noise figure and required SNR are inputs taken from datasheets and modulation tables, not synthesised, and the output is framed as the numbers that appear in the field. The Sensitivity mode reports the MDS as a power in dBm, as an rms voltage and a dBµV level across 50 Ω that can be dialled straight into a signal generator, and as a minimum detectable field strength in dBµV/m through the antenna factor, so the same threshold reads against a bench source, a spectrum analyser marker, or a field strength limit.
Beyond the forward calculation the tool does the three jobs that surround it. Spec Audit reverse-solves a datasheet sensitivity into the noise figure and SNR budget it implies, flags a figure that is physically impossible for its bandwidth, and compares receivers on a like for like basis. Processing Gain covers spread spectrum, where a direct sequence, GPS, or LoRa waveform is recovered below the thermal noise floor by the processing gain Gp = 10·log₁₀(BW ÷ R). Range turns the threshold into an operational answer: the received signal against the MDS gives a detection margin, and the maximum free space range is solved from the path loss the link can tolerate.
Compute the minimum detectable signal as MDS = −174 dBm/Hz + 10·log₁₀(B) + NF + SNR, with the thermal floor, the receiver floor, and the sensitivity reported separately. The noise bandwidth takes a Hz, kHz, or MHz unit selector and the temperature defaults to the 290 K reference on which −174 dBm/Hz and the noise figure are defined.
The MDS is expressed as an rms voltage and a dBµV level across 50 Ω, so it can be dialled into a signal generator, and as a minimum detectable field strength in dBµV/m through the antenna factor AF = 20·log₁₀(f) − G − 29.79. That is the field a monitoring or EMC receiver can just detect at the stated sensitivity.
Enter a quoted sensitivity and its bandwidth and the tool reports the combined noise figure and SNR budget above the thermal floor, then splits it into the implied noise figure for a given SNR and the implied SNR for a given noise figure. A sensitivity below the thermal floor is flagged as impossible, which catches a misread bandwidth or a transcription error.
Add several radios to the comparison list and read their implied budgets side by side on the same bandwidth. It is the fast way to sanity check competing sensitivity specifications, which are only meaningful once the bandwidth and the assumed SNR are made explicit.
A spread spectrum waveform occupies far more bandwidth than its information rate, and the receiver recovers a processing gain Gp = 10·log₁₀(BW ÷ R). That gain lowers the effective required SNR, so the sensitivity drops by Gp and can sit tens of dB below the thermal noise floor, exactly as a GPS or LoRa receiver does. A LoRa spreading factor table reports SF7 to SF12.
Build the received signal from the EIRP, free space path loss, receive antenna gain, and feeder loss, and compare it against the MDS for a colour coded detection margin. The maximum range is solved from the path loss the link can tolerate before the signal reaches the threshold, with and without a required margin.
The range figures use free space path loss alone. Real paths lose more to clutter, terrain, diffraction, and fade, so the free space range is an optimistic upper bound. The tool states this plainly and leaves the excess loss and fade margin to be added knowingly rather than buried in a single number.
Runs entirely in your browser. No bandwidths, sensitivities, or link parameters are submitted to a server. Useful for commercially confidential work, classified projects, or any environment where information security policy prohibits sending engineering data to third party services.
Inputs and outputs
Inputs 9
Outputs 9
Use cases
Add the thermal noise floor, the receiver noise figure, and the required SNR. MDS equals −174 dBm/Hz plus 10·log₁₀ of the bandwidth in Hz, plus the noise figure in dB, plus the required SNR in dB. For a 12.5 kHz channel with a 6 dB noise figure needing 12 dB of SNR, the thermal floor is −174 + 41 = −133 dBm, the receiver floor is −127 dBm, and the MDS is −115 dBm. The calculator does this live and also expresses the result as a voltage and a field strength.
The noise floor is the noise power at the receiver input, kTB plus the noise figure. The minimum detectable signal is the noise floor plus the SNR the demodulator needs to actually recover the signal. So the MDS is always above the noise floor by the required SNR. A signal exactly at the noise floor has a 0 dB SNR, which is the classic definition of the minimum detectable signal for a simple energy detector, but a real demodulator needs several dB more, and that margin is the required SNR term.
Because thermal noise is spread evenly across frequency, so a wider channel collects more of it. The noise power rises as 10·log₁₀ of the bandwidth, which is 3 dB for every doubling. A 25 kHz channel has a 3 dB higher floor than a 12.5 kHz channel and therefore a 3 dB worse sensitivity, all else equal. This is why narrowband systems are more sensitive and why reducing the receiver bandwidth to match the signal improves the minimum detectable level.
Through processing gain. A spread spectrum waveform occupies far more bandwidth than its information rate, and the receiver correlator concentrates the signal energy while averaging the noise, recovering a processing gain of 10·log₁₀ of the bandwidth divided by the information rate. That gain is subtracted from the demodulator SNR to give the effective SNR needed at the input, so the sensitivity drops by the processing gain and can sit tens of dB below the thermal noise floor. A GPS C/A signal spreads 50 bit/s across about 2 MHz for roughly 46 dB of gain, which is why it works far under the noise.
Compare it against the thermal floor for its bandwidth. The gap between the quoted sensitivity and the thermal floor is the sum of the noise figure and the required SNR, so it should be a positive number of a sensible size, typically 10 to 25 dB. If the quoted sensitivity is below the thermal floor for the stated bandwidth, it is physically impossible and the bandwidth or the figure has been misread. The Spec Audit mode does this automatically, flags the impossible case, and splits the budget into the implied noise figure and SNR.
It depends on the modulation and the target error rate. A simple detection threshold is 0 dB, analogue SSB or AM voice needs about 10 dB, analogue FM to 12 dB SINAD needs about 12 dB, digital voice like P25 or DMR reaches its reference error rate around 5 dB, QPSK needs about 10 dB, and 64-QAM needs around 22 dB. The Reference tab lists these as indicative ranges for orientation. They are not a specification, and the real figure is whatever the demodulator in your radio needs for the error rate you are designing to.
No. The range figures use free space path loss alone, so they are an optimistic upper bound rather than a prediction. Real paths lose more to clutter, terrain, diffraction, building penetration, and rain or fog fade, and gain nothing back. The tool states the free space assumption explicitly and leaves the excess path loss and the fade margin to be added separately, so you can see the free space bound and then apply the site specific losses knowingly. For a full propagation prediction use a path loss or coverage tool.
No. The calculator runs entirely in your browser. No bandwidths, sensitivities, or link parameters are submitted to a server. Useful for commercially confidential work, classified projects, or environments where information security policy prohibits sending engineering data to third party services.
