2.45 GHz superheterodyne receiver: 3.36 dB NF and −85.4 dBm sensitivity
Preselector, limiter, LNA, mixer, IF filter and VGA in front of the ADC cascade to 3.36 dB NF and 41.6 dB gain, so a 20 MHz channel meets the 12 dB SNR target at −85.4 dBm. The −65 dBm test carrier reaches the ADC 22.4 dB below clipping.
Change this: the LNA's NF from 0.9 to 1.5 dB. System NF reads 3.91 dB, so 0.54 dB of the 0.6 dB reaches the output.
NF 3.36 dB: 2.45 GHz Preselector sets it, 0.84 dB better if ideal, 33% of the excess noise.
IIP3 −13.64 dBm: IF VGA sets it, 20.60 dB better if ideal, 99% of 1/IIP3.
Block diagram
Level diagram
Budget at IF ADC
| Row | Value |
|---|---|
| Cumulative Gain | 41.60 dB |
| Source Contribution | −23.40 dBm |
| Cumulative NF | 3.36 dB |
| SNR | 32.39 dB |
| Noise Density | −128.80 dBm/Hz |
| Total Noise Power | −55.79 dBm |
| Noise Floor (Input) | −97.39 dBm |
| Sensitivity at Required SNR (Input) | −85.39 dBm |
| P1dB Headroom | 42.34 dB |
| Cumulative OIP3 | 27.96 dBm |
| Cumulative IIP3 | −13.64 dBm |
| Cumulative OP1dB | 17.94 dBm |
| Cumulative IP1dB | −22.66 dBm |
| First Stage to Compress OP1dB | 18.00 dBm |
| SFDR | 59.60 dB |
| IM3 | −126.12 dBm |
| Cumulative Noise + Distortion Ratio | 32.38 dB |
| Fully Linear Power | −23.40 dBm |
| Local Compression | 0.00 dB |
| Cumulative Compression | 0.00 dB |
| ADC Input Level | −22.40 dBFS |
| ADC Clipping Margin | 22.40 dB |
| ADC Effective ENOB | 5.09 bits |
Every number here is the analyzer's own, solved when the site was built from the same template the button opens. How the solver is checked.
Model limits
- The solver is unidirectional and matched: no mismatch interaction, return-loss cascade or stability.
- Noise density, NF, compression, gain and IM3 are read at each block's stage frequency, not solved across the operating band.
- A filter or any part with a frequency curve shapes noise inside the selected channel only: integrated noise, SNR and the noise floor integrate the curve across the channel, and Noise Density and NF stay spot values at its centre. A part with no curve is a scalar loss and NF. Noise outside the channel, such as an ADC's alias zones, is not counted.
- Mixer image noise in the NF is one point per band: the image band runs through every block upstream and is read at its centre as a spot density. Behind a curve the integrated noise rows carry the image channel point by point, so a filter skirt across it counts there.
- LO phase noise sets a scalar LO noise density. It is not integrated through the mixer. A phase noise mask gives integrated phase error, jitter and an EVM floor as scalar rows and a reciprocal-mixing estimate in the blocker panel; no noise row reads it.
- Distortion is IM3 only, and IM3 is not frequency resolved.
- Compression is a memoryless AM/AM curve: no AM/PM, bias or memory effects. Noise keeps the small-signal gain under a blocker.
Other solved chains
- Direct-sampling SDR receiver at 1.2 GHz: 4.45 dB NF with 18.8 dB of ADC headroom
- LoRa 915 MHz receiver front end: 4.66 dB NF with an LNA after the filter
- ADS-B 1090 MHz receiver, filter first: 4.99 dB NF
- ADS-B 1090 MHz receiver, LNA first: 2.65 dB NF
- 2.4 GHz chip-antenna receiver: 8.00 dB NF and −94.0 dBm sensitivity
- 28 GHz 5G transmitter lineup: 44.5 dBm EIRP with 8.7 dB P1dB headroom
- Ka-band 28.5 GHz transmitter lineup: 57.1 dBm EIRP with 6.2 dB P1dB headroom
- 2-way Wilkinson power-combining PA at 2.2 GHz: 36.5 dBm out, OIP3 51.0 dBm