SDR & Sub‑GHz: Signal Present but Wrong Decoder Block (URH vs SDR++)

Seeing energy in the spectrum or a “fat” trace in an analyzer does not mean your decoder chain matches the signal. Universal Radio Hacker (URH) and SDR++ are both excellent—but they organize demodulation, synchronization, and protocol interpretation differently. “It works in one app but not the other” is usually parameter mismatch, not a broken capture.

1) Same RF, different pipeline

FFT power proves something is there. Decoding requires the right sequence: filter → demod → clock recovery (if needed) → slicer/framer → protocol rules. If any stage uses the wrong bandwidth, wrong modulation family, or wrong symbol rate, the next block receives garbage—while the waterfall still looks “busy.”

2) URH: project model vs raw flowgraph

URH is built around signals as projects: you attach interpretation (bits, encoding, message types) after you’ve characterized the physical layer. It excels when you can nudge bit rate, tolerance, and message boundaries interactively.

• Autodetect is a hint: treat suggested parameters as starting points; verify against multiple bursts.
• “No bits” with a visible signal: often wrong modulation family (ASK vs FSK) or wrong samples-per-symbol.
• Import/export: use consistent sample rate and center when round-tripping IQ between tools.

3) SDR++: plugins, sinks, and manual wiring

SDR++ is a receiver workbench: you choose source, demodulator module, and downstream processing. If the demodulator expects FM voice and your signal is OOK pulses, you will get noise-shaped garbage that still “lights up” meters.

• Match demod to physics: narrowband digital modes need the correct digital mode path, not generic WFM.
• Bandwidth first: set channel filter width to roughly match the occupied bandwidth before chasing decoder plugins.
• Recorder sanity: confirm recorded baseband/IQ matches what URH expects (rate, format, endianness if raw).

4) Sub‑GHz-specific: OOK/FSK and timing

Garage-door and sensor remotes often use simple OOK or 2-FSK with tight timing. A decoder block tuned to the wrong baud or wrong encoding (NRZ vs Manchester-like) will fail even when the spectrum looks obvious.

• Measure symbol time: use a cursor on the demodulated trace or a period estimate before locking a slicer.
• Preamble hunting: some decoders need a stable preamble to lock; random mid-burst starts fail.
• Rolling / multi-frame: compare multiple captures—identical framing across presses suggests deterministic structure; changing frames suggest rolling security.

5) Parameter parity checklist

• Center frequency and sample rate identical (or properly shifted) between tools.
• Modulation family matches (ASK/OOK vs FSK vs FM).
• Bandwidth / channel filter wide enough for sidebands, narrow enough to reject neighbors.
• Symbol rate / bit length within a few percent of truth.
• Gain staging not clipping; overload mimics “no valid bits.”

6) A repeatable A/B workflow

1. Record 10–20 seconds of IQ from a known lab source at fixed gain.
2. Open the same file in URH and in SDR++ (or play through a file source in SDR++).
3. Establish the simplest physical-layer success: stable eye or clean OOK envelope in one tool, then copy parameters.
4. Only then add protocol interpretation (framing, checksums).
5. Document the winning numbers (rate, bandwidth, modulation) in your lab notes.

Validation criteria

1. The same IQ file decodes to consistent bits in one tool after parameter lock.
2. You can explain mismatch between tools as pipeline differences, not “mystery RF.”
3. Results reproduce on a second capture from the same lab setup.