hamradiotools

Nexus 1.10.3

Released 4 September 2026. The current release is 1.24.0. Download it · Every version

Added

  • Opt-in beta updates.

    Settings ▸ Appearance ▸ App updates has a new Receive beta (pre-release) updates switch. Turn it on and Nexus follows the beta channel — it offers pre-release builds as they land (newer features, less tested); leave it off and you stay on stable releases only, exactly as before. Beta builds carry a -beta.N version, are published as GitHub pre-releases, and never reach anyone who hasn't opted in. When a stable release supersedes the betas, opted-in stations roll onto it automatically.

  • The WSJT-X UDP feed can target several destinations at once.

    Settings ▸ Logging & Connectors ▸ WSJT-X UDP API now accepts a comma-separated list of addresses, not just one, so a single Nexus can feed a local tool (GridTracker, JTAlert) and a remote service (an FT8 contest scorer such as FT8 Battle Royale) at the same time — something WSJT-X's own single-sink UDP cannot do. A single address still works exactly as before, and a bad entry in the list is skipped rather than taking the whole feed down.

Fixed

  • A rig could get stuck transmitting after a CAT dropout (Yaesu, Enhanced USB port).

    On some Windows setups — reported on a Yaesu's Enhanced CAT port — a CAT interface hiccup while the radio was keyed (an over or a tune) could leave the radio transmitting until it was powered off. The 1.10.2 automatic CAT-reconnect rebuilt the link but forgot the radio was keyed and, on the external-Hamlib control path, never sent the key-up over the reopened channel — so the safety retry that recovers a stuck key (present in 1.10.1) was switched off exactly when it was needed. Nexus now always sends an unkey through a freshly reopened channel and keeps that recovery armed across a reconnect, so a transient CAT dropout no longer strands the transmitter. If the USB device itself disappears entirely, no software can unkey it — a hardware PTT that fails safe, or the rig's Standard port, is the sure guard there.

  • FT8 no longer sends 73 without receiving RR73.

    In an ordinary FT8 QSO, a DXpedition Fox's multiplexed confirmation addressed to you — from a Fox you'd been chasing on the same band — could be misread as your current partner's roger, and Nexus would send its closing 73 to a station that never rogered. The code that stitches a Fox's split confirmation back together was running on every QSO, not only in Hound mode; it now only does so when you're actually working a Fox, so an ordinary contact is never closed by a bystander's signal. (#236)

  • Frequency presets no longer collapse to a single band.

    After 1.10.2 the frequency/band dropdown could shrink to one band — an IC-7300 on 14.074 reading "custom", or only 40 m showing — for stations that had set per-radio band coverage. The band selector was trimming its list (and with it the frequency presets, which are the same list) by each radio's band-coverage setting, but that setting is a dual-radio routing choice, not a limit on what a rig can tune, so a radio with a partial coverage list wrongly hid every other band's presets. The presets now always show, independent of coverage. (#231, #232)

  • The generated-message pane now clears after an auto-logged contact.

    With "Clear DX call after logging" on, the DX Call/Grid fields cleared after a manual log but not when the FT8/RTTY sequencer logged a contact on its own — the app only noticed its own log actions, not the engine's. It now clears on every logged contact. (#210)

  • FTDX-101D power no longer dips on transmit.

    Nexus could send a power-level command to the radio mid-transmission, which WSJT-X never does; on the FTDX-101D that showed as a power foldback during the over. Nexus now sets power only before keying, never during a transmission. (#126)

  • The 3-D globe now remembers its map layers.

    The 2-D map remembered which layers you had on, but the 3-D globe reset to defaults every time you opened it, so layer choices made on the globe didn't stick. The globe now saves them the same way the flat map does. (#211)

  • The 60 m FT calling frequency is now a preset.

    Added 5.357 MHz for FT8 and FT4 on 60 m to the digital band presets. (#175)

  • The callsign lookup shows the beam heading.

    A QRZ/callbook lookup in the log entry now appends the short-path bearing from your grid, the way the station cards already do. (#222)

  • The log editor's time field no longer clips on macOS.

    The minutes part of the date/time field in the log editor could be cut off on the Mac; the field is now wide enough for it. (#233)

  • The radio engine no longer dies at launch on some Linux systems.

    On certain Linux audio devices — reported inside Debian 13 (trixie) containers, before any radio was configured — the capture stream was killed at startup with "RADIO ENGINE STOPPED — Buffer size 4800 is not in the supported range 10..=3840", leaving TX and RX dead until a restart. The 1.10.1 change that gave the capture buffer more scheduling slack sized it against the range the device declared it supported, and on these devices that declared range is looser than what ALSA actually enforces when the stream is opened, so the request was refused. Nexus now falls back to the system's default buffer whenever a device refuses the sized one, so the engine always starts. From a field report.

  • SSTV no longer loads some images upside down on macOS.

    An image with a 180° EXIF orientation tag came into the SSTV transmit stage upside down on the Mac (and would have transmitted that way). The app rotates images upright itself, and its safeguard against rotating one that the system already rotated only recognised the 90° cases — a 180° flip changes no dimensions, so it slipped through, and macOS's image decoder rotates on its own where Windows does not. Nexus now probes, once, whether the decoder applies orientation and acts on the answer, so a picture lands upright on every platform. From a field report.

Every version