hamradiotools

Nexus 0.17.12

Dual-radio setup, honest rig mode, FT exchange fields. Released 25 July 2026. The current release is 1.24.0. Download it · Every version

  • Setting up a second radio no longer overwrites the first one's COM port.

    Pressing Test CAT or Auto-test while editing a radio you are not operating on used to save that radio's port, model and audio devices onto your active radio's profile, silently and permanently, leaving both radios pointing at one set of ports. Every write from the rig form now goes to the radio the form is actually describing. On the air: your two rigs stay two rigs.

  • Auto-test now probes for the radio you are configuring.

    It seeded every port with the active radio's Hamlib model, and an Icom only ever answers at its own CI-V address — so with two radios set up, the second one's port could never answer and Auto-test kept handing back the first radio's port. It also no longer claims a CAT test passed when the test it ran was on the other radio.

  • The top bar tells the truth about your rig's mode.

    Its USB/FM buttons stopped reaching the radio back in June, when the transmit path moved to per-section modes. Clicking FM could not command FM; all it did was force a retune that re-asserted the section's own mode, which is what dragged a rig sitting in FM into USB/USB-D. The dead buttons are gone, and when your radio is actually in a different mode than Nexus thinks, the top bar now says so (rig: FM) instead of confidently printing the wrong one.

  • FT8/FT4: the DX call and grid fill in however the QSO started.

    They were only ever populated by a single click on a decode row, so working a caller any other way — the Work/Call buttons, the roster, Shift+Enter, JTAlert/GridTracker, or a station simply answering your CQ while the sequencer handled it — left the exchange panel blank, with Tx1–Tx4 showing "—" and the Tx buttons dead, even though the QSO ran and logged correctly. They now track the live QSO, and the grid resolves exactly the way the logged GRIDSQUARE does. This also removes a real hazard: pressing a Tx row while a stale call was showing could retarget the contact to the wrong station.

  • RST_SENT no longer goes missing when you work a station that answered your CQ.

    The report the sequencer had already armed was being discarded at the moment you clicked, and the only other place that captured it does not run during your own transmit slot — so the contact logged with a blank sent report. This is the "the log has it right in almost every case" case.

  • CI runs in minutes again.

    The 15 SSTV transmit/receive loopback cases were built unoptimized and each took over a minute, pushing the test job past an hour and starving the gates queued behind it. The DSP crates are now optimized under cargo test: the same suite runs in 13 seconds with every case and every assertion intact.

Every version