RAK10724 (RAK3401 + RAK13302 + RAK19007): PA not engaging — worse RSSI than T-Beam at 160mW

What product do you wish to discuss? RAK10724 WisMesh 1W Booster Starter Kit (RAK3401 + RAK13302 + RAK19007)

Computer OS? Windows (using Meshtastic app for configuration)

How often does the problem happen? Consistently, across all three kits

How can we replicate the problem? Assemble the RAK10724 kit per the quick start guide, flash the rak3401-1watt firmware, set TX power to 30 dBm, and compare RSSI at short range against a T-Beam.

Background

I purchased three RAK10724 WisMesh 1W Booster Starter Kits for use as Meshtastic ROUTER nodes. All three exhibit the same behavior: the SKY66122 PA does not appear to be engaging. Transmit performance is significantly worse than LILYGO T-Beams running at stock power (~160mW / 22 dBm).

Hardware Configuration

  • Core: RAK3401 (nRF52840)
  • LoRa module: RAK13302 1W (SX1262 + SKY66122)
  • Base board: RAK19007 (2nd gen)
  • Firmware: rak3401-1watt, version 2.7.26 (flashed via https://flasher.meshtastic.org, selected “RAK3401 1W”)
  • Firmware self-identifies as the 1W variant in the Meshtastic app
  • TX power: Set to 30 dBm in Radio Configuration → LoRa
  • Region: US (915 MHz)
  • Antenna: Connected to the RAK13302 IPEX port (NOT the RAK19007 base board IPEX port — visually confirmed)
  • LiPo: 3.7V 2000mAh connected to RAK19007 JST PH 2.0mm battery port
  • External power: Voltaic V75 battery bank via USB-C to RAK19007

Jumper Positions Tested

Both jumper positions on the RAK13302 3-pin header have been tested:

  • IN_5V (bridging IN_5V to center): LiPo connected, USB power via base board USB-C. Same weak result.
  • EX_5V (bridging EX_5V to center): USB power via base board USB-C. Same weak result.

Neither jumper position changes the transmit behavior.

The Problem — Comparative RSSI Data

All devices tested in the same room, approximately 5 feet apart, same antenna position, same receiving device (T-Beam “Alpha” paired to phone via Meshtastic app):

Device Hardware TX Power (configured) RSSI SNR
Gamma LILYGO T-Beam (stock) ~22 dBm (160mW) -45 dBm Strong
Delta LILYGO T-Beam (stock) ~22 dBm (160mW) -30 dBm Strong
Zeta RAK10724 (RAK3401 + RAK13302) 30 dBm (1W) -79 dBm 10 dB

The RAK13302 at “1W” is 34-49 dB weaker than T-Beams at 160mW in the same location.

If the PA were simply not amplifying (passthrough), I would expect roughly equivalent performance to the T-Beams (~22 dBm from the bare SX1262). The fact that the RAK13302 is dramatically worse suggests the unpowered PA circuit is acting as an attenuator — the signal passes through the inactive SKY66122 and loses significant power in the process.

What I Have Tried

  1. Flashed firmware via web flasher — confirmed rak3401-1watt variant
  2. Verified antenna is on the RAK13302 IPEX port, not the base board
  3. Tested both IN_5V and EX_5V jumper positions
  4. Tested with and without LiPo battery
  5. Tested with USB power connected before and after LiPo (different boot-up detection: “PWR” vs “BATT” modes)
  6. Confirmed TX power reads 30 dBm in the Meshtastic app
  7. Tested across three separate RAK10724 kits — all three show the same behavior

Observations

  • RX performance is improved over the T-Beams. The SAW filter and LNA on the receive path appear to be working — the RAK nodes hear distant nodes that the T-Beams cannot. This suggests the receive-side components are powered and functional; only the TX PA is not engaging.
  • The Meshtastic app shows different power source detection depending on boot order: USB first shows “PWR 4.19V”, LiPo first shows “BATT 98% 4.16V”. Neither mode produces correct TX output.
  • All three kits exhibit identical behavior, ruling out a single defective unit.

Related Threads

Those threads are on the Raspberry Pi / meshtasticd platform. This report confirms the same PA issue exists on the WisBlock (RAK3401 + RAK19007) platform with the native UF2 firmware.

Questions

  1. Is there a known-good firmware version where the PA is confirmed working on the RAK3401 + RAK13302 + RAK19007 combo?
  2. Does the rak3401-1watt firmware variant correctly assert the PA enable pin (ANT_SW / PA_CSD on the SKY66122) during transmit?
  3. Is there a hardware configuration step or additional jumper/connection that is not documented in the quick start guide?
  4. What RSSI values should I expect at short range (5 feet) if the PA is functioning correctly at 1W?

Thank you.

Welcome to the forum @RCS

We are not developing or maintaining the Meshtastic firmware.

Your questions (1) and (2) are better addressed directly to Meshtastic.

For (3), as long as you set the jumpers matching with your power supply selection there is no other hardware setup required.

For (4) depends on the used antenna. there is no “expected” RSSI value.

Since my post above is already linked in this thread (open about a month now with no response from RAK staff), seeing this one get a reply within hours is what prompted me to follow up here.

Have a T1000-E (credit-card sized unit with internal antenna) sitting a few feet away on my kitchen table, and a third party across town (miles away) can hear it consistently, day after day.

My RAK8623 WisMesh Station HP, by comparison, can’t be heard reliably even across the room. It’s putting out something, since nearby nodes can see it sometimes, but the signal is dramatically weaker than I’d expect from a 1W unit - or even a handheld card.

At this point there are three of us across different threads (myself, RCS above, and the RAK6421 + pyMC_repeater report) all independently landing on the same symptom: the PA doesn’t appear to be engaging, and in at least one case (RCS’s numbers here) it looks like it may actually be attenuating the signal rather than just failing to boost it. My T1000-E card test provides additional support for this attenuation hypothesis.

Given the pattern across three separate hardware/firmware paths, any insight from RAK’s side would help, whether this is a known hardware quirk, a config.d/GPIO issue, or something that needs to go upstream to Meshtastic?

For (4) depends on the used antenna. there is no “expected” RSSI value.

the Tbeams and the RAK have the same antenna at the same appox distance in the same room as each other and the receiving device.