This guide covers the T-shaped dipoles and single-wire antennas that come with most ExpressLRS receivers, plus the ceramic antennas soldered onto the smallest boards. The physics is the same for other 2.4 GHz and 900 MHz links.
How a dipole sends and receives
A dipole radiates in a doughnut shape. It is strongest broadside, from the sides of the element, and weakest off its ends. Oscar Liang describes the worst case as the tips of the transmitter and receiver antennas pointing at each other: each one then looks at the other through its weakest direction.
Receiver antennas are also linearly polarized. The signal is strongest when the antenna in the radio and the one on the drone are oriented the same way. A vertical antenna on one end and a horizontal one on the other is cross polarization, and Oscar Liang calls the loss from it considerable. The small single-wire antennas on whoops behave much the same way.
That gives the two rules everything else follows from: keep the tip of the antenna from pointing at the pilot, and keep its orientation close to that of the antenna on the radio.
What blocks the signal
Carbon fibre conducts electricity. It can block, attenuate and detune an antenna that sits too close to it. The same goes for the battery, the ESC and flight controller, large metal parts and the thick battery leads. Keep the receiver antenna away from the video transmitter and its antenna too.
Only the active element needs that clearance: the exposed end of the antenna, beyond the coax. The coax itself can run along the frame. The FPV Bible build guide puts numbers on it:
- at least 1-2 cm of the element should stick out past the carbon,
- don't run the element parallel to carbon,
- don't tape it flat against the frame,
- don't coil the excess cable,
- keep it out of the prop path.
One antenna
Oscar Liang prefers a single receiver antenna mounted vertically. The FPV Bible suggests leading it away from the frame, rearward and upward. Both point to the same layout: the element at the back of the drone, standing up and clear of the carbon, with the props unable to reach it.
For long range, Oscar Liang mounts it vertically and tilts it slightly backwards. In forward flight the drone pitches nose down, and the tilt brings the antenna back towards vertical.
Mounting options, from the quickest:
- two zip ties on a rear arm or standoff, pointing in opposite directions, with the elements secured in heatshrink,
- TPU tubes zip-tied to the rear standoffs,
- a 3D-printed mount made for your frame.
In the comments under his article, Oscar Liang answers two common questions: arms are a popular place for the antennas and the motors shouldn't interfere, and it doesn't matter which half of the dipole points up.
Two antennas
Receivers with two antenna connectors come in two kinds, and the ExpressLRS docs name them clearly:
- antenna diversity: two antennas and a switch, one radio chip (RP3, Namimno Flash Diversity, Matek R24-D),
- true diversity: two complete radio chips, each with its own antenna (EP1 Dual, BetaFPV SuperD).
According to the ExpressLRS hardware guide, diversity matters most at medium and long range. It keeps an antenna null from pointing at the transmitter in certain orientations of the drone, and it covers the moment when carbon or concrete blocks one of the antennas.
How to place the pair depends on the flying:
- freestyle: about 90 degrees apart, one vertical and one horizontal. That way at least one antenna is roughly aligned with the radio whatever the drone is doing. Oscar Liang prefers this to two antennas at plus and minus 45 degrees, and the FPV Bible suggests the same 90 degrees with tubes on the rear standoffs,
- long-range cruising: in Oscar Liang's tests, both antennas vertical, parallel and on opposite sides of the drone gave stronger RSSI and more consistent LQ. A cruising drone doesn't flip much, so polarization stays predictable, and the other side of the drone is covered when the frame shadows one antenna.
On antenna diversity receivers the ExpressLRS Lua script has an "Antenna Mode" setting. "Diversity" keeps both antennas active and switches to the one with better RSSI. "Ant1" or "Ant2" uses only one antenna, for both control data and telemetry. Leave it on "Diversity" unless you are testing one antenna at a time. On true diversity receivers, ExpressLRS 4.x negotiates the mode between the transmitter module and the receiver during sync, so there's no separate receiver setting.
Ceramic antennas on the smallest receivers
Receivers such as the EP2, RP2 and R24-S have the antenna on the board. The ExpressLRS docs rank the range like this: ceramic antenna < mini dipole ("Minimortal-T") < sleeved dipole < half-wave dipole.
That doesn't make them useless. For racing, the docs say the ceramic antenna is fine because flights are short, and they note that 5 km is achievable on a ceramic antenna receiver with no LNA and no PA, at 250 Hz and 100 mW, with a clear line of sight. A clear line of sight is the key part. The docs warn that tucking the ceramic antenna deep inside the stack behind carbon reduces its performance. Heatshrink, a whoop canopy or carbon parts around it also weaken the signal.
If the receiver has a ceramic antenna, mount it as far from the carbon and electronics as the frame allows, not between the plates.
The antenna on the radio
The other end of the link matters too. Oscar Liang keeps the antenna on the radio or module vertical as well, so it matches a vertical receiver antenna, and never points its tip at the drone. If you fly almost directly overhead, for example hovering above yourself, a horizontal antenna on the radio may work better, because the drone is then off the tip of a vertical one.
A bench test that checks the whole installation
The ExpressLRS "Before First Flight" guide describes a quick test that catches a bad antenna before it costs you a drone:
- Take the props off and power the drone.
- In the Lua script, set the transmitter module to its lowest power.
- Put the radio about 1 m from the receiver, with both antennas oriented the same way.
- Turn off the video transmitter, put it in pit mode or cool it with a fan.
- Read the RSSI in dBm, or 1RSS in the telemetry screen.
For 2.4 GHz, the docs give values from around -40 dBm up to -25 dBm as good. For 900 MHz, about -20 dBm.
If you get much lower numbers, the guide suggests checking:
- whether the antenna on the module is loose,
- whether the pigtail is damaged or badly connected to the board,
- whether the receiver uses an SMD antenna, which naturally gives lower values, more so with heatshrink, a canopy or carbon around it.
If nothing helps, swap the antennas. Most ExpressLRS receivers use IPEX1 (U.FL) connectors, and any antenna tuned for the right frequency will do. The docs also allow old 2.4 GHz Wi-Fi router antennas for 2.4 GHz modules, but not dual-band ones.
On a two-antenna receiver, telemetry shows both: 1RSS for the first antenna, 2RSS for the second, and ANT tells you which one is active. If one value is clearly worse than the other at the same orientation, that antenna or its pigtail is the suspect.
Look after the U.FL connector
The U.FL connector is tiny and not built for frequent swapping. Hirose, which makes it, rates the standard series for 30 mating cycles. Plug the antenna in once, then secure the joint, for example with a dab of hot glue as the FPV Bible suggests, and add heatshrink as strain relief where the cable leaves the frame.
The Yuke Aero guide to its ExpressLRS receiver adds a warning: don't power the receiver without the antenna fitted. When it sends telemetry, the output of its amplifier is then left without a load. The same caution makes sense for any receiver with a telemetry amplifier.
After a crash, look at the antennas before the next pack: whether the elements are still there, whether the heatshrink is intact and whether the cable is cut. If anything looks off, repeat the bench test before you fly.