LNA for SDR: Best Setup for ADS-B, GNSS and Weak-Signal Receivers

An LNA for an SDR is a low-noise preamplifier placed ahead of the receiver to set the system noise figure before cable loss and the SDR’s own front end can degrade it. Whether you need one depends on three things: how much loss sits between antenna and radio, how noisy the SDR is compared with the antenna’s noise, and how strong the off-channel signals are. ADS-B, GNSS and L-band satellite reception usually benefit; HF reception usually does not.

This page is a decision guide for weak-signal receive setups built around a software-defined radio. It sits under RF amplifiers and complements the general LNA selection guide.

Does an SDR need an LNA?

The answer comes from the Friis formula for cascaded noise factor, with every quantity in linear units:

Ftotal = F1 + (F2 − 1)/G1 + (F3 − 1)/(G1G2) + …

A passive cable of loss L has a noise factor equal to L and a gain of 1/L. Put a lossy cable first and its loss adds directly to the noise figure. Put an LNA with enough gain first and the cable and SDR contributions are divided by that gain. Low-cost SDRs are wideband designs whose noise figure is often several decibels, so a good LNA at the antenna can buy a large improvement. A well-designed receiver with a short cable gains little.

Worked example: ADS-B at 1090 MHz with 5 dB of cable

Assume an SDR with a noise figure of 6 dB (F = 3.98), a coax run with 5 dB of loss at 1090 MHz (L = 3.16), and an LNA with a noise figure of 0.8 dB (F = 1.20) and 20 dB of gain (G = 100). These are illustrative values, not measurements of a particular product.

  • No LNA: F = 3.16 × 3.98 = 12.6, so NF = 11.0 dB. Cable loss and SDR noise figure simply add.
  • LNA at the SDR end of the cable: F = 3.16 × (1.20 + 2.98/100) = 3.90, so NF = 5.9 dB. The LNA hides the SDR but cannot recover the 5 dB already lost in the cable.
  • LNA at the antenna: F = 1.20 + 2.16/100 + 2.98/(100/3.16) = 1.32, so NF = 1.2 dB.

Noise figure is not the whole story, because the antenna also delivers noise. Converting to noise temperature, Te = 290(F − 1), gives 3361 K, 840 K and 92 K for the three cases. With an antenna temperature of 290 K, typical of a low-elevation antenna that sees the warm ground and horizon, system temperatures are 3651 K, 1130 K and 382 K. The mast LNA therefore improves signal-to-noise ratio by 10 log(3651/382) = 9.8 dB, and the receiver-end LNA by 5.1 dB. For a sky-facing GNSS or satellite antenna with an antenna temperature nearer 100 K, the mast LNA gains about 12.6 dB.

Nearly 10 dB of SNR is roughly a factor of three in free-space range, which is why the mast LNA is the standard upgrade for ADS-B feeders.

When an LNA does not help

Below about 30 MHz the antenna noise is dominated by atmospheric, galactic and man-made noise, which is typically tens of decibels above kTB. Suppose the external noise is 30 dB above kTB (Fa = 1000) and the SDR has a 10 dB noise figure. The SNR penalty from the receiver is 10 log((1000 + 9)/1000) = 0.04 dB. Even a 15 dB noise figure costs only 0.13 dB. An LNA at HF adds gain, and therefore overload risk, without improving sensitivity. A simple test: if the noise floor rises by more than a few decibels when the antenna is connected, the receiver is already sensitive enough.

Use-case matrix

The table summarises typical choices. Noise figure and gain targets are typical ranges, not requirements, and the right answer depends on cable length and the local RF environment.

ApplicationFrequencyLNA worthwhile?Placement and filteringMain threat
ADS-B1090 MHzUsually yes, especially with long cable runsAt the antenna; filtered LNA with a 1090 MHz SAW or cavity filter; NF below about 1 dB, gain 15 to 25 dBCellular, paging and FM broadcast overload; nearby DME/TACAN in the 960 to 1215 MHz band
GNSS L1/E11575.42 MHzUsually built into active antennasAt the antenna; typical active antennas have a filtered LNA with 25 to 40 dB gain, powered over the coaxStrong nearby cellular and L-band transmitters; too much gain into a receiver that expects a passive antenna
GNSS L5/E5a1176.45 MHzYes, same reasoning as L1At the antenna; filter must pass L5 and reject aeronautical pulses in the same bandDME/TACAN pulse interference
Satellite L-band (weather, Inmarsat, Iridium)Roughly 1.5 to 1.7 GHzYes for small dishes and patch antennasAt the feed; filtered LNA with NF well below 1 dBAdjacent cellular bands, 4G/5G base stations
VHF weather satellitesAround 137 MHzSometimesAt the antenna, with an FM broadcast notch or bandpass before the LNAFM broadcast at 88 to 108 MHz
HF3 to 30 MHzRarelyAttenuation and filtering usually help more than gainAM broadcast and strong amateur signals

Where the LNA goes

Figure 1 shows the three configurations from the worked example across cable losses from 0 to 10 dB. The no-LNA case rises one decibel per decibel of cable loss. The receiver-end LNA removes the SDR’s contribution but still rises one for one. The mast-mounted LNA stays near its own noise figure until cable loss approaches the LNA gain. With 20 dB of LNA gain and 10 dB of cable, the system noise figure is still only about 2 dB.

System noise figure versus cable loss with no LNA, an LNA at the SDR and an LNA at the antenna
Figure 1: With 5 dB of cable, system noise figure is 11.0 dB with no LNA, 5.9 dB with the LNA at the SDR and 1.2 dB at the antenna.

A filter on the antenna side of the LNA adds its loss directly to the noise figure: 1 dB of filter loss ahead of a 0.8 dB LNA gives at least 1.8 dB. Filtered LNAs therefore either use a very low-loss first filter or split the filtering around the first gain stage. The LNA design article covers the first-stage trade-offs.

Bias-tee powering

A mast LNA needs DC, and running a separate power cable up the mast is awkward. A bias tee puts DC on the centre conductor of the same coax that carries the RF. At the receiver end, an inductor (the RF choke) feeds DC onto the line while presenting a high impedance to RF, and a series capacitor blocks DC from reaching the SDR input. At the LNA end, a matching choke takes the DC off and a capacitor keeps it out of the antenna.

Bias tee feeding DC up the coax to a mast-mounted LNA while blocking DC from the SDR input
Figure 2: Bias-tee powering of a mast LNA (schematic only). The choke feeds DC onto the coax and the capacitor keeps it out of the SDR.

The component values follow from reactance. The choke should present an impedance much higher than 50 Ω at the lowest operating frequency. For 500 Ω at 1090 MHz, L = X/(2Ï€f) = 500/(2Ï€ × 1.09 × 109) = 73 nH. The blocking capacitor should present much less than 50 Ω; for 5 Ω at the same frequency, C = 1/(2Ï€fX) = 29 pF. The choke’s self-resonant frequency must be above the band of interest, or it will look capacitive and load the line.

Many SDRs and GNSS receivers include a bias tee that can be switched on from software or a jumper. Before relying on one, check four points:

  • Voltage and current. The LNA’s supply range must include the bias-tee voltage, and its current draw must sit within what the SDR can supply. Built-in bias tees are usually low-current supplies.
  • Voltage drop. The drop along the coax is V = I × Rloop. An LNA drawing 60 mA through 1.5 Ω of loop resistance loses 0.09 V, which matters only for marginal supplies or very long runs.
  • DC paths. A DC-shorted antenna, such as a folded dipole or some loop and ground-plane designs, will short the supply unless the LNA blocks DC at its input. Splitters, filters and some attenuators also block or short DC.
  • Other equipment. Never connect a powered line to test equipment or another receiver that is not DC-blocked.

Overload risk near FM and cellular transmitters

An unfiltered wideband LNA amplifies everything the antenna delivers. Near an FM broadcast site or a cellular base station, the total power at the LNA can push the LNA, or the SDR behind it, into compression. The result is a raised noise floor, spurious signals across the spectrum and lost weak signals: the receiver looks deaf even though the LNA has a good noise figure. The mechanisms are covered in detail on receiver blocking and desensitization.

Worked example: an FM transmitter 5 km away

Take a 10 kW ERP FM transmitter at 98 MHz. ERP is referenced to a half-wave dipole, so EIRP = 70 dBm + 2.15 dB = 72.15 dBm. The free-space path loss at 5 km is 20 log(5) + 20 log(98) + 32.44 = 86.4 dB. With a 0 dBi antenna response at 98 MHz the received level is about −14 dBm. Free space is an upper bound; terrain and clutter usually reduce it, but a line-of-sight rooftop can approach it.

Free-space FM broadcast power versus distance for 10 kW and 100 kW transmitters with and without filtering
Figure 3: Free-space FM level at 98 MHz versus distance, unfiltered and with 40 dB of pre-filter rejection, against example overload levels.

Consider an example LNA with 20 dB gain and an output P1dB of +15 dBm, so its input P1dB is about −4 dBm. The LNA itself is not compressed at −14 dBm, but its output is −14 + 20 = +6 dBm of FM power. If the SDR, with its gain set high, overloads at an illustrative −30 dBm, that point referred to the LNA input is −50 dBm, and the FM signal is 36 dB above it. Turning down the SDR gain helps only partly.

Figure 3 plots the free-space level against distance for 10 kW and 100 kW transmitters. Without filtering, both exceed the illustrative SDR overload level everywhere out to 50 km. With 40 dB of rejection at 98 MHz from a filter before the LNA, the 10 kW case falls below that level beyond about 3.2 km and the 100 kW case beyond about 10 km. Bandpass filters for ADS-B, GNSS and L-band, often SAW or BAW types, typically reject distant bands such as FM by much more than this; the harder cases are cellular signals only tens of megahertz away from the wanted band.

In practice, choose a filtered LNA for any single-band setup, add an FM notch for VHF reception near broadcast sites, and set LNA gain roughly 10 to 15 dB above the cable loss; beyond that, gain only reduces dynamic range. The broadband amplifier page discusses when an unfiltered wideband amplifier is still the right choice. After installation, the noise floor should rise when the LNA is powered, but the spectrum should not fill with spurious signals.

Frequently asked questions

Do I need an LNA for my SDR?

Above roughly 100 MHz with more than a couple of decibels of cable loss, an LNA at the antenna usually improves SNR noticeably. With a short cable and a quiet receiver, the gain is small. At HF, external noise dominates and an LNA rarely helps.

What is the best LNA for ADS-B at 1090 MHz?

A filtered LNA with a 1090 MHz bandpass filter, a noise figure below about 1 dB and moderate gain, mounted at the antenna. The filter matters as much as the noise figure, because cellular and broadcast signals are the most common cause of poor ADS-B range near towns.

How do I power an LNA via bias tee?

Enable the SDR’s built-in bias tee or insert an external bias tee between the coax and the SDR, then feed DC onto the coax. The LNA must accept that voltage on its RF input or output port. Check current limits and make sure nothing else on the line shorts or blocks the DC.

Related topics

For selection criteria in general, see how to choose an LNA. GNSS receivers are covered in GNSS front-end design, satellite links in satellite RF design, and wideband SDR receivers in Radio Architectures Part 13: direct RF sampling. Terms used here are defined in the wireless terms glossary.

Scroll to Top