hamradiotools

Nexus 0.21.0

Released 29 July 2026. The current release is 1.24.0. Download it · Every version

APRS gets a map

APRS had no map. Everything sat in a small area at the top left of the screen with the rest of the window empty. Stations, their tracks and their paths now plot geographically, with the controls and lists moved to a rail beside it. On a narrow window the map comes first.

Nothing new is decoded for this — position, course and speed were already in the packets, with nowhere to draw them. Clicking a station on the map highlights its row in the list, and the reverse.

SSTV shows you the band, then shows you the picture

The SSTV screen had no waterfall at all, so there was no way to see what was on the frequency before an image arrived. That space is now a live waterfall — and when a signal starts decoding, the same space becomes the picture, building downward as it comes in.

Because the picture stands where the spectrum was, you cannot see whether the radio is off frequency while an image is arriving. So the mistuning is now stated outright: a "tuning +12 Hz" readout beside the line count, whenever it drifts past 10 Hz. The decoder already worked this out from the header and had simply never shown it to you.

A DXpedition calendar you can actually read at a glance

The DXpedition view now opens on a traditional month calendar with today clearly marked and each operation drawn across the days it runs. Clicking one opens its detail.

Above it, a plain-language summary of what to chase: which are on the air now, which start soonest, the best band and time for each, and the best day or two to try. All of that was already being calculated and simply spread across the page for you to assemble yourself.

The dense band-by-hour heatmaps move behind a "Details" tab and are toned down when shown, so the page is no longer a wall of yellow, orange and red when you scroll it.

Satellites: one pass at a time, on a bigger globe

Clicking a satellite drew every OTHER satellite's ground track too, so the pass you had just chosen was buried under a dozen unrelated lines. Now only the selected bird is drawn.

The globe was also locked to a fixed width no matter how large the window was. It now grows with the space available.

QRZ confirmations arrive on their own

Nexus could already pull your QRZ logbook down — QSOs logged elsewhere and their confirmations — but only when you pressed Sync. Turn on Settings ▸ Logbook & QSL ▸ QRZ ▸ "Pull confirmations automatically" and it happens hourly instead, so confirmations appear as people post them.

Only what CHANGED is fetched after the first run, so an hourly check is a small request rather than your whole logbook twenty-four times a day. It is off by default, a failed check never skips the span it missed, and the schedule survives a restart.

As before, a QRZ confirmation shows the contact as confirmed but never counts toward DXCC or WAS — those need LoTW or a paper card, and counting QRZ would inflate them.

Fixed: alerts repeating on every cycle

A new-DXCC alert would fire again and again for the same station, once per transmission, instead of once when it appeared. Plain CQ alerts did the same.

Two causes, and they compounded. An alert was identified partly by the station's measured audio frequency — which drifts a few hertz between transmissions — so the same station saying the same thing looked like a brand new event each time. And because every one of those counted as a separate remembered alert, a busy band filled the "already alerted" memory in a minute or two; the oldest entries were then discarded first, which included the record saying the new one had already been announced. So it announced it again.

Alerts are now identified by who transmitted and what they said. The things that should only ever alert once — a new entity, a new grid, a watch-list hit — are remembered separately from the ones that legitimately repeat, so no amount of band traffic can push them out.

Fixed: one internal error could leave the radio deaf until you restarted

A safety lock guards the shared decoder, and if anything ever failed while holding it, that lock stayed broken for the rest of the session. Every decode and every transmit after it failed too — silently. The app kept running and the waterfall kept painting while nothing was being heard, and the only sign was a line in a log file you would never see. It now recovers and carries on.

Not something that was reported on the air. It was found while tracking down the JT65 crash, and it is exactly the failure that crash would have triggered.

Fixed: the window could stop responding while a decode was running

Transmitting and decoding both need the same audio engine, and the transmit side used to wait its turn while holding a lock the interface also needed. If a decode was still running when the next transmit came due, the whole window froze until it finished — under a second on a fast PC, several seconds on a Raspberry Pi.

The transmission is now prepared without holding that lock. Nothing changes on the air: the same work happens at the same moment, the interface just stays alive through it.

Every version