RSPduo dual-tuner: running two rtl_airband processes simultaneously (Master/Slave approach)

[ukbodypilot]

I've been running an RSPduo with two independent rtl_airband processes — one per tuner — and documented the working approach after spending a lot of time figuring out why the obvious method doesn't work.

The problem with Mode 2

The RSPduo's "Dual Tuner Independent RX" (rspduo_mode=2) sounds perfect for this use case but doesn't work with rtl_airband. Mode 2 opens the entire device and holds an exclusive lock — a second rtl_airband process cannot open it at all.

What actually works: Master/Slave (modes 4 and 8)

• Process 1: device_string = "driver=sdrplay,rspduo_mode=4" → Master, Tuner 1
• Process 2: device_string = "driver=sdrplay,rspduo_mode=8" → Slave, Tuner 2

One more catch: the Slave device only appears in SoapySDR's device list after the Master is already running and streaming. You must wait ~3 seconds after the Master is confirmed alive before starting the Slave.

Key requirements

1. fventuri's SoapySDRPlay3 fork — the standard package does not include rspduo_mode=4/8 support. You need the dual-tuner-submodes branch.
2. Sample rate ≤ 2.0 MSps per tuner — both tuners share the ADC clock in Master/Slave mode.
3. Start Master first, wait 3s, then start Slave — non-negotiable.
4. sdrplay_apiService ignores SIGTERM — use SIGKILL to restart it reliably.

Tool

I've packaged this as a standalone Python script + shell scripts:

https://github.com/ukbodypilot/rspduo-airband

It handles config generation, the start sequence, audio sink setup (PipeWire/PulseAudio null sinks), and settings persistence. No pip dependencies — pure stdlib.

Hope this saves someone the hours of trial and error I went through. Happy to answer questions.