Antenna¶

CONNECTOR · SWR · RF SWITCH · PRACTICAL CHOICE

The antenna is the only part of a node that knows no firmware and still determines most of your range. What the firmware does know about it is the switch in between: one antenna has to both transmit and receive, and the chip has to know how that changeover is wired. This chapter describes that switch, the connector, and the relation between standing wave ratio and reflected power.

[!NOTE] Source. This page has been verified against the firmware itself: MeshCore v1.16.0, commit 03b6ef4, 28 July 2026 — files src/helpers/radiolib/CustomSX1262.h and the RF flags in variants/. Connectors, SWR and antenna types do not come from the firmware; that is general radio engineering and is marked as such.

Diagram of the RF path: from the chip through the RF switch to the
connector and the antenna, with DIO2 switching the transmit path and RXEN
switching the receive path

One antenna, two paths¶

Transmitting and receiving happen over the same antenna but not over the same path: the transmit path runs through the power amplifier, the receive path through the LNA. So there is a switch between chip and connector. MeshCore knows two ways of driving it, and they are not mutually exclusive.

Drive	Flag	Variant directories
DIO2 on the chip drives the switch	`SX126X_DIO2_AS_RF_SWITCH`	60 of 79
separate GPIO pins drive the switch	`SX126X_RXEN` · `SX126X_TXEN`	24 of 79

Counted per variant directory across variants/, including both -D flags in platformio.ini and #define lines in a header inside that directory; commented-out lines do not count.

The first is the simpler one: the chip pulls DIO2 high as soon as it transmits and low as soon as it listens. The firmware only has to pass the flag on to RadioLib, which happens in std_init() — see The LoRa Transceiver.

The second appears on boards with an external power amplifier or a separate LNA, where one line is not enough. If one of the two pins is missing, the wrapper fills it in with RADIOLIB_NC.

A board that uses both¶

The T-Beam 1W combines them, and the comment in the variant says exactly how:

variants/lilygo_tbeam_1w/platformio.ini r.22-28

  ; RF switch configuration:
  ;   DIO2 controls TX path (PA enable) via SX126X_DIO2_AS_RF_SWITCH
  ;   GPIO21 controls RX path (LNA enable) via SX126X_RXEN
  ; Truth table: DIO2=1,RXEN=0 → TX | DIO2=0,RXEN=1 → RX
  -D SX126X_DIO2_AS_RF_SWITCH=true
  -D SX126X_RXEN=21
  -D SX126X_DIO3_TCXO_VOLTAGE=3.0

DIO2 switches the transmit path, GPIO21 the receive path. SX126X_TXEN is not set and therefore becomes RADIOLIB_NC. The same file also explains what sits behind that switch:

variants/lilygo_tbeam_1w/platformio.ini r.33-34

  ; TX power: 22dBm to SX1262, PA module adds ~10dB for 32dBm total
  -D LORA_TX_POWER=22

The chip runs at 22 dBm, the external amplifier turns that into roughly 32 dBm. That is why LORA_TX_POWER can never be read on its own: it is the power the chip delivers, not the power leaving the antenna. What you are allowed to radiate is in Regulations & Duty Cycle.

[!WARNING] Transmitting without an antenna, or with an antenna for the wrong band, sends the power back into the power amplifier. At 22 dBm that is already unhealthy; on a board with an external amplifier it is fatal. Connect the antenna before you apply power.

Connector and cable¶

Not from the firmware — general radio engineering.

Connector	Where	Note
SMA	on most development boards	pin in the centre on the cable side
RP-SMA	much WiFi hardware	reversed pin; fits mechanically, does not work
IPEX/U.FL	on modules and compact nodes	fragile, meant for a pigtail

SMA and RP-SMA screw onto each other without making contact in the centre. That is the most common fault behind a node that transmits but reaches nothing.

Cable introduces loss. At 868 MHz thin coax easily causes several dB of loss per ten metres. A node close to the antenna with a short cable beats a node indoors with a long one. The effect of a dB on distance is discussed in Link Budget.

SWR: what comes back¶

Not from the firmware — this is the standard conversion from standing wave ratio to reflected power.

SWR	Reflected	In practice
1.0 : 1	0 %	theoretically perfect
1.5 : 1	4 %	fine
2.0 : 1	11 %	acceptable
3.0 : 1	25 %	too high, needs matching
∞	100 %	open or shorted

The SX1262 has no SWR measurement on board and the firmware therefore reads nothing about it. You notice a bad match only through disappointing range and a power amplifier that runs hot.

Three practical rules¶

More gain is direction, not power. An antenna with higher gain looks flatter and loses above and below. Why more gain can even work against you is in Higher and stronger isn't always better; the radiation pattern and the dead zone belong there and are not repeated here.
A half-wave dipole is nearly always enough. 2.15 dBi, no ground plane needed, and exactly the reference the regulations use.
Height beats power. One metre higher more often gains you more than one dB extra, because it lifts obstacles out of the first Fresnel zone.

Sources¶

Firmware, commit 03b6ef4 (v1.16.0, 28 July 2026):

src/helpers/radiolib/CustomSX1262.h — setDio2AsRfSwitch() and setRfSwitchPins()
variants/lilygo_tbeam_1w/platformio.ini — the RF switch truth table and the external amplifier

Not from the firmware repo: connectors, SWR and antenna types. That is general radio engineering; the SWR table is the standard conversion from standing wave ratio to reflected power.