Rebuilt From the Air is a 14-part postmortem series about the GopherTrunk decoders that passed every test they had and never matched a real radio. In each one the encoder in the test suite and the decoder under test were written from the same layout — a sync word, a bit order, an interleaver, a line code — and the layout was wrong. Round trips stayed green for months. The first real capture decoded nothing.

From the Issue Tracker named that failure the self-consistent trap, and From Spec to Shipping told the first full rebuild, SmartNet. This series is everything that came after: MDC1200, whose tones were sliced as the data; the CTCSS and DCS gates; FLEX, LoRa and System Fusion; three parts on NXDN — the frame, the control channel on real air, and the voice scrambler; D-STAR; and two parts on P25 Phase 2. Each part gives the symptom as it was reported, the theories that were wrong, the reference or capture that settled it, and the test that now fails without the fix.

It is also a correction. Several earlier posts on this blog described the formats these decoders used to implement. Part 1 lists them, each part says what it overturns, and the finale is the errata — together with the checklist for telling a real decoder from an invented one, and a ledger of what is on-air verified, what is capture-verified, and what still waits for a recording.

Every post reads three ways: a TL;DR + cheat-sheet for skimmers, bold headers and tables for the medium read, and the full investigation narrative for the deep read.

New here? Start with Troubleshooting a decode in the Digital Trunking module, then come back for the rebuilds.<p class="post-list__empty">No posts in this series yet — check back soon.</p><p class="blog-feed-link"> See all solution postmortems or subscribe via RSS. </p>