RAK7289V2 (WisGateOS 2.2.14, Basics Station): one of two concentrators ~18–20 dB less sensitive — and gateway won't reconnect after reboot/power-cycle

Setup

  • Gateway: RAK7289V2 (16-channel, 2× SX1302 concentrators exposed as USB CDC-ACM: usb:/dev/ttyACM0 = slave-0, usb:/dev/ttyACM1 = slave-1).
  • Firmware: WisGateOS_2.2.14_RAK.
  • Mode: Basics Station → external ChirpStack v4 LNS over ws:// (plain), region AU915, 16 uplink channels enabled (ch 0–15).
  • The unit has been stable for ~12 days; load average ~0.5 (not overloaded).
    Problem 1 — One concentrator has ~18–20 dB worse RX sensitivity than the other

Verified with the same node (same TX power), whose frames hop across both concentrators:

Concentrator Channels / freq Same-node RSSI SNR
slave-0 / ttyACM0 (rctx0) ch0–7, ~915.4–916.0 MHz −80…−87 dBm +11…+13
slave-1 / ttyACM1 (rctx1) ch8–15, ~917.6–917.8 MHz −98…−103 dBm +9…+11

Aggregate over ~700 recent frames: rctx0 avg RSSI −97 (best −63); rctx1 avg RSSI −115 (best only −87). rctx1 never hears a strong signal, even from near nodes — pointing to a degraded RX path on that concentrator, not node distribution.

Supporting data:

[RAL:DEBU] [read_slave_pipe]|slave_index=0, crc_ok=4830, crc_err=412, crc_non=12   (~8% CRC err)[RAL:DEBU] [read_slave_pipe]|slave_index=1, crc_ok=9250, crc_err=1320, crc_non=5   (~12.5% CRC err)

dmesg shows the other port (ttyACM0, usb 1-1.2) re-enumerated once, while the degraded one (ttyACM1, usb 1-1.3) has been stable:

[972522.0] usb 1-1.2: USB disconnect, device number 6[972522.4] usb 1-1.2: new full-speed USB device number 7 using ehci-platform

Also seeing frequent GPS/PPS time-sync rejections (~37%, 402/1083) and MCU/SX130x drift up to ±20 ppm:

[SYN:INFO] MCU/SX130X drift stats: ... max: +20.5ppm - threshold q90: -16.2ppm[SYN:VERB] Time sync rejected: quality=933 threshold=850

Q1: Is a ~18–20 dB RX asymmetry between the two SX1302 concentrators a known hardware/firmware issue? How should we isolate it (antenna cable/connector/pigtail on that concentrator, USB, or the SX1302 module)? Is any per-concentrator RX calibration difference expected?

Problem 2 — After a UI reboot then a full power-cycle, the gateway is up but won’t reconnect (no LNS, no admin access)

To try to recover concentrator 2, we rebooted from the WebUI. The gateway reconnected to the LNS for ~2 minutes, then dropped. We then did a full power-cycle. Since then (>40 min):

  • No Basics Station connection to the LNS (gateway last_seen frozen; zero uplinks).
  • No remote admin access (WireGuard tunnel down).
  • On site: network LED green, unit booted, stable (no reboot loop).
  • During the brief reconnect window, our LNS logged the gateway reporting an unexpected region_id="au915_cds" alongside the normal au915_1 — which makes us suspect a config reset on reboot.

So the unit is alive and link-up locally, but has no working path to the internet/servers.

Q2: Does WisGateOS 2.2.14 reset network/Work-Mode/region config on reboot or power-cycle? Is there a known config-loss issue where a Basics Station gateway comes back link-up but no longer connects (e.g., reverting out of Basics Station mode, or losing WAN/DNS/WireGuard config)? How do we recover a remotely-managed gateway without on-site reconfiguration?

Q3: We’re on 2.2.14. Would upgrading to 2.2.16 or the 2.3.1 major release address either the per-concentrator RX asymmetry or the reboot config-reset behavior? Recommended upgrade path for a 16-channel Basics Station deployment managed remotely?

Thanks in advance.

Hello @fervio ,

First things first, the RX sensitivity is not a problem at all. There may be some noise on some of the channels, resulting in what you observed.

You should update the firmware of the gateway; it is recommended that you use the latest 2.3.1 firmware: https://downloads.rakwireless.com/LoRa/WisGateOS2/WisGateOS2_Latest_Firmware.zip
There are many fixes over the previous releases between 2.2.14 and 2.3.1, so update it and check if the reboot behaviour persists. Also, what is your backhole interface (WiFi, LTE, Ethernet)?

As for the GPS, where is the gateway deployed in the environment? Does it have a clear sky view, or are there some obstacles?

Regards
Nikola Semov