hamradiotools

Nexus 1.14.0

Released 20 September 2026. The current release is 1.24.0. Download it · Every version

Added

  • During a contest, the Operate cockpit now tells you whether you have worked a station in this contest.

    The callsign card's Dupe 20m badge has always meant "worked on this band, ever", which is what you want day to day but not mid-contest — a contact from years ago lit it even though the station was a fresh one to work. A second badge now appears beside it while a contest is running: Contest dupe if this contest's own log already has them, and an amber Club dupe if another position at your club has. The lifetime badge is unchanged and still means what it always did, so the two can be read at a glance without either one changing under you.

  • RTTY: double-click a callsign to grab it, and F1–F8 you can edit.

    Double-click a call in RTTY's Decoded text and it fills the Their call box and the log strip's callsign together — a portable or compound call such as VE3/K1ABC comes across whole, and a call garbled by a lost figures shift is filled as printed for you to correct. Anything that is not a call (599, CQ, TEST, a grid square) is ignored, and the caret stays where you were typing. The two call fields are now one: type in either and the other follows, and logging the contact clears both. The dock now has eight macro keys, and F1–F8 work from the keyboard while the RTTY cockpit is on screen. Two sets are built in and switched from the dock: Everyday, the four you had plus four empty keys, and Contest — CQ, exchange, TU, my call, his call, S&P exchange, AGN and B4. Hover a key and click ✎, or click an empty key, to change its title and message; it is saved straight away, and you can put one key or a whole set back to the built-in messages. Every macro now goes out on a line of its own and ends with a space, the way RTTY contest messages are written, so your call starts a line on the other station's screen. The Contest set's exchange keys send the exchange of the contest you are running — the zone and QTH, state or section you are sending this weekend, not last June's Field Day class — and with no contest running they tell you so instead of sending. In a contest the double-click also fills the exchange: a zone lands in the zone box, a section or state in the QTH box, by that contest's own rules rather than a fixed list, and Enter in one of those boxes logs the contact. A word that could be two things, or one Nexus has no list of values for, is left for you to type.

  • CQ World-Wide RTTY DX Contest.

    Pick it under Settings › Contesting › Contest and the log strip takes its exchange: RST, your CQ zone and, for stations in the continental USA and Canada, a state or Canadian call area, using the sponsor's own codes (NWT, NF, LB and PEI among them). Contacts score the way the sponsor counts them: 3 points between continents, 2 within a continent, 1 within your own country, with zones, countries and W/VE QTHs each counted once per band. The strip tells you when the rig is on a band the contest does not use (it runs on 80, 40, 20, 15 and 10 m) and still logs the contact. The Cabrillo export follows the sponsor's template: a two-digit zone, DX where a station sent no QTH, the LOCATION spelling the sponsor's list uses, and the category, claimed-score, name and email headers. The email comes from a new optional Email for contest logs setting; leave it blank and the line is left out. If your callsign is in the US or Canada but your contest state is blank or not on the sponsor's list, you would send the DX exchange with no QTH, so Nexus warns you in Settings, on the log strip and when the contest starts, and suggests the code a section means (EMA is MA). It never stops you, because operating from outside the US and Canada really is DX. Typing NT or PE is read as NWT or PEI.

  • Illinois QSO Party.

    Pick it under Settings › Contesting › Contest and the workspace runs the Western Illinois ARC's own rules: 1700Z Sunday of the third full weekend of October for eight hours, 160 through 2 m without the WARC bands, phone 1 point and CW or digital 2. Illinois stations send RST and their county; everyone else sends RST and their state, province or country — this party asks a DX station for its country rather than the word DX. The county box takes the county name as well as the code: type Cook and it offers COOK, st clair offers SCLA, and space turns a full name into the code that goes in the log. A half-typed name is not completed for you and neither is one that could be several counties — pick from the list or finish typing, because nothing is ever guessed onto the air. CW and digital count as one mode for dupes here, so a station worked on CW shows as a dupe on RTTY on that band — the strip says so before you call them. FT8 and FT4 earn no credit (the sponsor's own rule); other digital modes are encouraged. Illinois entrants multiply by counties, states, provinces and up to five DXCC entities; everyone else by the Illinois counties worked. The club's two calls, W9AWE and W9OAB, are worth 100 bonus points each and Nexus adds them to your score and your claimed score as soon as they are in the log — there is no box to tick. The Cabrillo export writes CONTEST: ILLINOIS QSO PARTY and an IL-COUNTY: header for an Illinois entry, exactly as the sponsor's sample log does.

  • The contest strip suggests a CQ zone and flags a missing QTH.

    Type a call in a contest that exchanges CQ zones and the zone box shows the zone the country file gives that call as a faint hint — it never fills the box, because a station outside its prefix's zone sends its own. And when a USA or Canada station is about to be logged without a QTH in a contest where they send one, the strip says so; you can still log it.

  • Nexus now picks working DX-cluster nodes for you, and shows how each node is doing.

    Phone spots on the Needed board and in the Spots panel come from human-run DX-cluster nodes, and a node can stop working without warning: both nodes Nexus 1.13.0 shipped with did, on the same day. Nexus now keeps two nodes connected, chosen from eight nodes built into this release, each checked before it went in. A node that stops answering for five minutes while your other spot feeds keep working is skipped for a day and another node takes its place, and the connection log says which node and why. A quiet band never counts against a node: once a node has logged you in, it stays in use however long the band is silent. Operators are spread across the nodes by callsign, so everyone does not end up on the same two. Settings › Logging & Connectors › Integrations & Feeds › Spot Sources shows each node as In use, Standby, or Not answering and why. If your install still had the node list Nexus shipped, it switches to this by itself. If you edited the list yourself, it stays exactly as you left it and nothing is switched for you: Spot Sources shows how each of your nodes is doing, and Pick working nodes automatically is there when you want it. A node you remove from your own list now disconnects when you save, instead of at the next restart.

  • Cluster login SSID — stop two stations on one callsign knocking each other off.

    A cluster node allows one session per callsign and disconnects the older one, so running a second Nexus, or Nexus beside another cluster program on the same call, made the two bump each other indefinitely: a cluster connection that flapped with no explanation on screen. Under Settings › Logging & Connectors › Integrations & Feeds › Spot Sources › Cluster login SSID you can now give each one its own SSID (2 logs in as W9XYZ-2), which the node treats as a separate user, and both stay connected. Nexus does not choose one for you — it is empty by default and logs in exactly as before, because a node that requires registration treats the suffixed call as a different, unregistered user and will not let it post spots, and because an SSID picked for you could collide with the one you already use in the other program. Your spots still reach the network under your plain callsign either way; nodes relaying them strip the suffix.

  • Stations you have already worked now step aside on the POTA/SOTA board and the Spots panel.

    Both are on by default, and each says how many rows it is hiding.

    • POTA/SOTA board — Hide worked today. An activator you logged at their park since 0000Z (UTC) is hidden, so the board lists the activations you still need today. It comes back at 0000Z, or as soon as that activator is spotted at a different park. Only a contact logged with the park counts: HUNT, double-clicking a park on the Connect map and Work on the Needed board all add it for you, but a contact logged without the park reference hides nothing.
    • Spots panel — Hide worked. A station you logged is hidden on every band, for the rest of the UTC day. The picker beside the chip changes that window to 1 hour, 4 hours, 24 hours or 7 days — for 13 Colonies, Route 66 and other events where one callsign is on the air all week. A station you still need on the band and mode it was spotted on always stays.
    • To see everything again, click the chip — Hide worked today · 3 on the board, Hide worked · 12 on the Spots panel. Every row comes back marked WORKED TODAY or worked 2h ago. The board remembers your choice per window; the Spots panel until you close Nexus.
  • Remote: a responsiveness check you can run on your own link.

    In the browser's Awards view, Responsiveness → Run check makes ten small dial steps, changes band and back, and sends one Stop while the radio is idle, then reports how long the screen took to respond, how long the radio took to confirm, how far the readout trailed, and how often controls went off by themselves — and whether each confirmation came by the station's own push or by polling — with a copy-as-text button so a result can be pasted into a report. It changes nothing while it is not running. It is the yardstick every later Remote responsiveness change is measured against; the same check runs in CI against a scripted station on 100 ms and 400 ms links.

Changed

  • Duplicate contacts are now kept in the log for the contests that cross-check, and still refused in Field Day.

    Work a station you have already worked in CQ WW (CW, SSB or RTTY), CQ WPX (CW or SSB), ARRL Sweepstakes (CW or SSB) or the ARRL VHF contests (January, June and September) and the contact is logged instead of turned away. It is marked as a duplicate, it is worth no points and no multiplier, and your QSO count and claimed score do not include it. It goes into your Cabrillo as an ordinary QSO line. This is what both sponsors ask for: their log checkers remove duplicates with no penalty, and CQ asks entrants not to delete them, because a contact missing from your log becomes a Not-In-Log penalty for the station that worked you — worth twice the contact at CQ. ARRL and Winter Field Day are unchanged: a duplicate there is still refused and nothing is written, because Field Day logs are not checked at all, so there is no penalty to spare anyone and its summary sheet asks for raw non-duplicate counts. The QSO parties are also unchanged — they are run by five different sponsors and we have not read a rule from any of them that settles it. The duplicate warning you get while typing a callsign is the same as it always was; what changed is only what happens if you log the contact anyway.

  • Large logbooks are much faster.

    With a 150,000-contact log, Nexus had become slow enough to stop working: every status update re-scanned the whole log, and logging one contact sent the whole log (about 100 MB) to every window. Now an unchanged log costs the status update nothing, a logged contact sends just that contact to each window, each window keeps one copy of the log instead of three or four, and awards, needs and Journey are counted once per change to the log instead of on every refresh. Reading the log no longer holds up the radio while it converts, and a contact WSJT-X logs in companion mode is added to the end of the log instead of rewriting the whole file.

  • Contest Cabrillo logs record the frequency you were on.

    Outside Field Day, each QSO line now carries the dial the contact was logged on instead of the band's lower edge, which sponsors such as CQ WW ask award entrants for. Contacts logged before this update keep the band edge. Field Day logs are unchanged.

  • Remote: the dial follows the wheel, and it stops swallowing your corrections.

    Spinning the readout digits over a remote link used to ignore every notch made while a command was still out — on a 100 ms link four of ten notches reached the radio, on a 400 ms link three — and the digits themselves did not move until the station read the new frequency back. Every notch now counts: a step made while the radio is still working joins the next command and is sent from the dial the station read back, so ten notches move the dial ten notches. The digits move on the gesture itself rather than when the radio answers, shown dimmed with a trailing "…" until the station's own reading confirms them. A value the station never confirms is never left standing — after two and a half seconds the digits go back to the station's dial and say the tune was not confirmed. Only the dial digits are shown ahead of the radio; the band, the mode, the sideband, the privilege shading, the S-meter and every transmit control keep reading the station and nothing else.

  • Remote: a tune, band or mode change confirms the moment the radio does it.

    Until now every rig-touching control was answered "pending" and the browser found out it had landed by asking — a state read, then a result read, each a round trip through the relay, with a one-second wait between them — so a control felt half a second to two seconds slow, and the tuning wheel ignored input for all of it. The station now tells the browser the outcome itself (operation protocol v5), together with the fresh control state, the instant its radio loop reads the new dial back; the browser installs both and is ready for the next gesture with nothing in between. Polling stays as the safety net, so a pushed message that is lost costs the old delay and never the outcome. A desktop or site on the previous version keeps polling exactly as before; the push needs both updated.

  • Remote: the station controls stop blinking off after every command, and a call pressed in that moment is held instead of refused.

    Every held control used to go grey for about a second after each command while the browser re-read the station, and a call, a Resend or an exchange pressed while the browser was still re-reading came back "Not sent. Nothing reached the station" — most often right after turning TX on. The controls now stay lit throughout, and once the station has answered, a gesture made while the browser is still re-reading waits for the fresh state and goes out on it, a second and a half at the outside. It is never sent late or behind your back: if the re-read does not come, or you press Stop or turn TX off while it waits, or the station's transmit permission moves under it, the gesture is refused and says so, with nothing reaching the radio. While a command is still on its way to the station a second one is refused the same way, "Not sent", with nothing reaching the radio — where before, the control simply went dead until the first had landed.

Fixed

  • Your CW and RTTY macro keys now send a serial number in the contests that use one — the text they transmit has changed.

    In CQ WPX (CW and SSB), ARRL November Sweepstakes (CW and SSB) and the California QSO Party, an {EXCH} macro used to leave the serial out entirely: an F-key set to {CALL} 599 {EXCH} {EXCH} keyed K1ABC 599 and stopped there, and Sweepstakes sent A W9XYZ 74 WI with no number in front. The same keys now send K1ABC 599 1 1 and 1 A W9XYZ 74 WI. The serial is counted for you: the log strip shows the number for the contact you are working, and the contact is logged and exported with the number that went out. Until now there was no number anywhere, and every contact was logged and exported as serial 0, which is a log a sponsor cannot check.

  • The log strip and the phone cockpit show the number you are actually sending.

    In a serial contest they read the exchange straight off the session, whose serial slot is a permanent placeholder, so the "Sent:" line said 599 0 while the contact was logged with the real number. On SSB that is what you read aloud — and it is what the voice-keyer hint told you to record into a slot — so your log would have recorded serials that never went on the air. Both now show the issued number. Contests with no serial are unchanged.

  • The number belongs to the station you are working, from the moment you enter their call.

    Search and pounce: call one station, get no answer, work somebody else and come back — the first station is given the number they already copied, not a later one, and the number you actually sent is the number that reaches your log. Correcting a busted call keeps the number too, so fixing K1ABC to K1ABD does not hand them a second one. Reading the strip, previewing an F-key, or sending your exchange twice never moves the run; logging the contact does, so the next station gets the next number. Contests with no serial in their exchange, Field Day among them, are unchanged.

  • The contest dupe warning now follows the contest's own rule instead of Field Day's, and it was wrong in both directions.

    While you type a callsign, the log strip and the Operate callsign card tell you whether this contest has already worked that station. That check asked Field Day's question — same call, same band, same mode class — whatever contest was running, and only the two Field Days actually work that way. The other fifteen were wrong, seven of them one way and eight the other. In Sweepstakes you work a station once, on any band and any mode, so someone you had already worked on 40m CW showed as new on 20m phone and you called them and were refused at the log, after the over. CQ WW and CQ WPX count one contact per band in either mode, so the same thing happened across a mode change. The other eight went wrong the opposite way: in the five QSO parties and the three ARRL VHF contests, working a mobile again from a new county — or a rover from a new grid — is a fresh contact worth points, and the warning said "already worked", so you passed over a station you should have called. Where a contest's rule depends on an exchange you have not copied yet, the warning now says nothing rather than guessing: being told nothing costs you a duplicate that scores zero, while a wrong "already worked" costs you the contact.

  • A duplicate in the contest log table is marked as the duplicate — and the real contact beside it no longer is.

    Now that a duplicate is kept in the log rather than refused (see above), the table has rows that are worth nothing, and you need to see which when you check your score against the sponsor's. It had been working its own dupes out by looking for a callsign that turned up twice on the same band and mode, which went wrong two ways: it marked both rows, so your first, real, scoring contact was flagged as a duplicate for the crime of being worked again later; and it used Field Day's rule everywhere, so a Sweepstakes duplicate on another band was not marked at all, while two perfectly legal QSO-party contacts with a mobile in two different counties were both marked. The table now shows what the log itself recorded.

  • Remote: the Awards list no longer gets squeezed off a phone screen.

    In the hosted browser the Awards view carries a status line above the summary and, since the responsiveness check was added, a strip below it. On a phone-sized window with the text enlarged there was no height left for the summary between the two: it collapsed to nothing, and the strip drew over the award cards, so a tap landed on the strip instead of the card under it. The summary now keeps a minimum height of its own whatever the window does — a card scroller worth using — and anything that will not fit scrolls with the column instead of being cut off at the bottom. Windows at 1024×768 and above look exactly as they did.

  • A Field Day exchange copied by the RTTY sequencer now reaches the log as fields, not just as a note.

    If Auto was running in Field Day and you switched Field Day off while a contact was still on the air, that contact finished under the Field Day exchange but no longer had a contest log to go to. Its class and section landed in the record's comment as plain text — 2A EMA — and nowhere else, so they were not in the ADIF you exported and no other program could read them back as an exchange. They are now written to their proper ADIF fields as well, and the comment still reads the same as before. Contacts made the ordinary way were never affected.

  • Editing a contest contact no longer wipes its exchange.

    Correcting anything on a contest QSO — a busted call, a report, a grid — silently dropped the contest data from the record: the contest name, the serials you sent and received, and both exchanges, none of which the edit form shows you in the first place. The contact stayed in your log, but its contest fields were gone from the log and from every ADIF export made afterwards, so a corrected contact would not score. Editing now keeps all of it, including on a callsign correction — the exchange is what went over the air, and fixing the call does not change what was sent.

  • The CW and Phone band-activity strips no longer rebuild themselves every time the spots refresh.

    On a busy band — a contest evening can put over 600 stations on the 20 m CW strip — one new spot made Nexus throw away and redraw every flag above it, which on a slower PC was enough to stall the waterfall for a moment before the new flag appeared. The flags of stations still on the air stay put now; only the spots that arrived or aged out change. (#320)

  • Prompt before logging now queues contacts instead of replacing them.

    If a second contact finished while the popup was still open, the popup kept showing the first station while Nexus had moved on to the second — and logging then filed the first station's call on the second contact's time and frequency. The popup now keeps the contact it is showing, tells you how many are waiting, and brings up the next one as soon as you log or discard it. Nothing is mixed, and no contact is lost.

  • If the queue ever fills, Nexus says so rather than logging quietly.

    Leaving an FT8 run to itself with Prompt before logging on can stack up contacts; past 64 waiting, the oldest is logged as it stands rather than lost. That contact is now named in the Connections log with its call, band, mode and time, and the popup says how many went in without your confirmation since you last answered it — because those contacts upload to your services and join the LoTW batch like any other.

  • Correcting a callsign in that popup corrects what was looked up from the wrong one.

    The country, state and a name that came from the busted call are re-derived for the call you actually worked, and a grid that had only been looked up is dropped — a grid the station itself sent, or one you typed, is kept. Correcting a call from Nexus Remote now does the same. An edited report no longer leaves the old report behind in the comment.

  • The ATU button says when your radio's tuner cannot be started over CAT.

    On Icom and Kenwood radios connected through Hamlib the command only switches the tuner in or out, so the button is greyed out and points at the TUNER button on the radio — instead of quietly switching the tuner in, tuning nothing, and, in the digital section, ending your QSO for a tune-up that never happened. Where a tune-up can be started, it now ends your QSO only once the radio has actually accepted it. (#322)

  • Dates you type are UTC everywhere.

    The Logbook's edit form, "Log a contact from another radio" and the export date range took dates through the Windows date control, which follows the PC's own calendar and silently discarded a date it could not read — so an impossible date looked like a cleared one, and on the export that quietly widened the range to your whole log. All three now take the date as plain UTC text, refuse a date that cannot be, and hold the button rather than exporting something other than what you asked for. (#280)

  • The waterfall no longer freezes a few minutes into a session.

    While the radio was slow to answer (a CAT read on a slow link, a log save), each Nexus window kept asking for its next update several times a second without waiting for the last one, and those waiting requests could take every thread the waterfall and meters are drawn from. Each screen now waits for its answer before asking again, and the waiting happens where it cannot hold up the waterfall. (#335)

  • Contact times in the Logbook's edit form and in "Log a contact from another radio" are always 24-hour UTC.

    Those time boxes followed the Windows clock format, so on a PC set to a 12-hour clock a contact at 00:58 UTC showed as "12:58 AM", and changing it to 00 put it straight back to 12. Contacts that were right got "corrected" by 12 hours. You now type the time as HH:MM or HH:MM:SS. A time that can't be right, such as 25:00, is refused with a message rather than saved or replaced with the current time. An edit that doesn't touch the time now keeps it to the second. The times Nexus stored were always correct. (#280)

  • Confirming a contact with Prompt before logging keeps its end time

    and, on a split contact, its receive frequency. Ham Radio Deluxe no longer shows 00:00 as the end time. (#329)

  • Digital section: running the radio's ATU now ends like Tune.

    TX switches off, and Nexus no longer goes back to calling the station of an unfinished QSO. (#322)

  • Uploads are never dropped silently.

    When the upload queue fills, the Connections log names each dropped contact and the service it never reached, and a ClubLog catch-up can no longer push out contacts you just made. Every upload service now says so in the Connections log when it gives up on a contact after its retries — ClubLog, World Radio League, N3FJP and Cloudlog/Wavelog as well as QRZ and eQSL — naming the contact and how to send it again. (#290)

  • The band menu outlines the band you are on.

    It was a thin bar on the left edge that looked like a "(". (#323)

  • With two radios, switching back no longer shows a band as "custom".

    A radio running FT8 reports its data mode, and the band menu didn't recognise that as its FT8 channel, so after a switch 80 m read "80m (custom)" although the radio hadn't moved. (#334)

  • Remote: revoking a browser's logging or station-control permission always takes effect

    , even while the station is busy. (#318)

  • Remote: the Release button no longer flickers.

    It greyed out on every heartbeat — about once a second — so a click could land while it was disabled and do nothing. It now stays lit, and a Release clicked while the station is answering waits for that answer and then releases the station, once.

  • Two radios: a default radio no longer takes over the one you are using.

    With a "default radio" set for everything else, changing frequency on your active radio could hand control to the default one — even on a band the active radio covers. Tuning, a spot click, or another program setting the frequency through Nexus's CAT sharing (WSJT-X or JTDX split, for example) could all do it, so a radio could switch away mid-session. The default radio now only takes bands the active radio does not cover.

  • Two radios: switching radios never leaves the old one transmitting.

    When you switch radios while one is transmitting, Nexus unkeys it first. If that unkey did not go through, the switch went ahead anyway and left the old radio keyed in the background, where nothing could unkey it. The switch now waits until the old radio has really stopped transmitting, and keeps trying to unkey it until it does. Nexus also now writes to its log which audio device it opened after a switch, or why it could not, so a radio that loses its audio can be diagnosed.

  • CW with the soundcard keyer: clicking a spot now puts you on the spot.

    The soundcard keyer sends CW as an audio tone in the radio's data mode, where the tone goes out a pitch above the dial (below it on 80 and 40 m). Clicking a spot tuned the dial straight to the spot, so you transmitted about 600 Hz off the station you clicked, and heard them at a pitch too low to hear or decode. The dial now sits a pitch away from the spot so your signal lands on it, from every place a spot can be clicked, and a pile-up split moves with it. On 80 and 40 m a click on the CW scope also went the wrong way by twice the pitch; it now follows the side the radio is really on. The CAT, WinKeyer and serial keyers were not affected: they put the radio in CW, which is already on the spot.

  • A satellite contact you log after the bird sets now keeps its satellite tag.

    The guide says you can log a contact once your hands are free, but a contact logged after the pass ended — or after choosing "None" or stopping the track — lost PROP_MODE and SAT_NAME, so it earned no LoTW satellite credit and a 2 m contact counted toward terrestrial VUCC instead. Nexus now remembers the last satellite you worked until you pick another transponder or restart, and still tags a contact only when its frequency is on that satellite's downlink, so an ordinary contact is never tagged.

  • Icom: a refused scope span now tells you why, instead of guessing.

    When an Icom refused a span change, Nexus always said the scope was in Fixed mode and told you to switch it to Center — even when it was already in Center, so you were sent to check a setting that was already right. It now asks the radio. If the scope really is in Fixed mode it says so; if it is in Center it says that plainly, and suggests the other likely cause: a span change while the radio is transmitting or tuning. Nothing about how Nexus sets the span has changed.

  • macOS: Prove TX now works, and SSTV can be sent on 145.800.

    Two confirmation prompts asked their question through the browser's own confirm() box, which this app's macOS webview does not draw. No dialog appeared and the answer came back as "no", so on macOS Prove TX did nothing at all — the button looked live and never keyed the radio — and SSTV refused to transmit on 145.800 MHz, the frequency that asks you to confirm first because it is the ISS downlink. Both now ask in the app's own dialog, with the same question and the same warning. Nothing was ever transmitted without asking: the prompts failed in the safe direction, so what was lost was the feature, not the protection. Windows and Linux were unaffected.

  • Work on the Needed board now tags the park.

    Working a POTA or SOTA row from the Needed board moved the radio but, unlike HUNT and the map, never told the logbook which park it was — so the contact was logged without the park reference: it earned no hunter credit, exported without SIG/SIG_INFO, and pota.app could not match it. The next contact with that activator is now tagged with the park, exactly as HUNT does it, and a reference Nexus cannot read is reported instead of dropped. A row with more than one activation live — a summit that is also a park, spotted to both programmes — tags nothing and names them, 2 activations live (POTA US-0001, SOTA W7A/MN-001), so you can pick the one you worked on the POTA/SOTA board; a SOTA spot from hours ago does not count as live.

  • A contact logged at two parks at once counts for both of them.

    A two-fer — one QSO at a site where two park boundaries overlap, logged as US-0001,US-0002 — matched neither park, so both kept their NEW PARK badge and the Needed board went on offering both activations after you had worked them.

  • CW macros no longer call "CQ FD" in another contest.

    The cockpit had two built-in macro sets, casual and Field Day, and used the Field Day one in every contest — so F1 in CQ WW CW or a state QSO party called CQ FD DE …, Field Day's own call, on the air. There is now a contest set with the same cadence and CQ TEST; Field Day keeps CQ FD, and a macro profile of your own still wins over both.

  • A mistyped Sweepstakes section no longer counts as a multiplier.

    A received section that is not on the list counted as a multiplier of its own. The contact is still logged; the section just counts for nothing.

  • Remote shows the contest screen for every contest.

    Through Remote, the contest screen was blank for any contest except ARRL Field Day and Winter Field Day, because the browser refused the contest's rules as unknown. It now shows for all of them — CQ WW, Sweepstakes and the state QSO parties included.

  • A contest other than Field Day comes back after a restart.

    If you left a QSO party or another contest running and restarted Nexus, it came back in Chat unless a Field Day class and section happened to be filled in, and the contacts you had logged stayed out of view. Nexus now reopens the contest you left running. A contest whose own exchange is incomplete still stays closed, and entering it tells you what to fill in.

  • {EXCH} in a CW macro sends the exchange of the contest you are running.

    It always sent your Field Day class and section, so in any other contest a macro with {EXCH} sent the wrong exchange, or nothing at all. It now sends that contest's exchange without the signal report — your zone in CQ WW CW, for example. Field Day sends exactly what it did before. {CLASS} and {SECTION} are empty outside Field Day.

  • RTTY Auto no longer sends a Field Day exchange in other contests.

    The auto-sequencer only knows the Field Day exchange, yet it turned on in any contest and would have sent your Field Day class and section to every station it worked. In any contest other than ARRL Field Day or Winter Field Day it now refuses to turn on and says why: send your exchange with the macros and log each contact yourself. Outside a contest, and in both Field Days, Auto works as before.

  • The contest log strip works for contests other than Field Day.

    In a contest whose exchange includes a signal report, the report box came back blank after every contact; it now goes back to 599 (59 on phone). A CQ zone has to be a number from 1 to 40, and the strip says so while you type. The strip's label, hint and button now say "contest log" instead of Field Day, and the zone box is captioned Zone. Field Day's strip is unchanged.

  • The contest screen names the contest you are running.

    Outside Field Day the banner still read "ARRL Field Day", the header showed an empty class and section, the log table's columns were Class and Section (blank on every row), the Score Summary printed a station class, a power multiplier and a bonus list, and the Field Day bonus checklist sat under the score. The banner and header now name the contest and show what you are sending, the log table has one column per field that contest exchanges, the Score Summary shows QSO points × multipliers, and the bonus checklist appears only in Field Day. Field Day's screen is unchanged.

  • A cluster node that never asks for your callsign no longer kills that feed slot for the rest of the session.

    Some nodes accept the connection and never send a login prompt — one of the nodes Nexus ships with was doing exactly that — and some send a prompt Nexus could not read (next entry). Nexus waited on them forever: it never logged in, never gave up, and never retried, so that node stayed dead until you restarted the app, and nothing on screen said why. Nexus now gives a node 30 seconds to ask for your callsign, then disconnects and tries again later, waiting longer after each try, up to ten minutes, so a node that never answers is not called over and over. Settings › Connections records which it was: the node said nothing at all, or it spoke and never prompted. Once a node has asked for your callsign and Nexus has answered, it is never dropped for being quiet afterwards — a cluster on a dead band can be silent for a long time and that is not a fault.

  • Nexus now logs in to nodes whose prompt ends with a new line.

    A node that greets with "Please enter your call:" and then a line ending — AR-Cluster nodes do, and k1ttt.net:7373 was measured doing it — was never answered, because Nexus replied only to a prompt left waiting at the end of what it had received. It now answers a prompt that is the last thing the node sent, in either shape. A prompt-looking line with more text behind it is still read as part of the greeting, so your callsign is never sent into a welcome message.

  • A cluster node that was working could be reported as dead.

    Nexus asked the system for the node's address and then only ever tried the first one it was handed. Many nodes have both an IPv6 and an IPv4 address, and on a connection with no working IPv6 route the IPv6 one is commonly offered first — so the attempt failed and the IPv4 address sitting right behind it, which would have connected, was never tried. The node read as down while it was answering perfectly well, and no phone spots arrived from it. Nexus now tries every address the node has, in turn, and only reports it unreachable once they have all failed. When one does fail, Settings › Connections now says so and names what was tried, so "this node is down" can be told from "this connection cannot reach it over IPv6".

  • The Phone source could name a cluster node that was down.

    The Now-Bar's Phone pill and the Phone source line on the Needed board named the first node Nexus had started, whether or not it was connected, so "live" could sit beside the name of a dead node. They now name only the nodes you are logged in to, or, while none is, the nodes being tried.

  • Remote: the bottom of the Awards view is no longer cut off on a phone screen.

    On a phone-sized window with the text enlarged, the last part of the hosted Awards view was cut off with no way to reach it; anything that does not fit now scrolls with the column instead. Windows at 1024×768 and above look exactly as they did.

  • SSB/phone spots stopped arriving, and the Spots panel showed no Phone chip.

    Both of the human DX-cluster nodes Nexus shipped with, ve7cc.net:23 and dxc.wa9pie.net:8000, went down on the same day, and every install that still had the default list had no phone source left — RBN kept CW and the digital modes flowing, which is why only Phone (and the mode chip the Spots panel builds from what it is actually receiving) vanished. Two more nodes are now in the default list — dx.w1nr.net:23 (DXSpider) and dxspots.com:7300 (CC Cluster, on a high port for networks that block telnet port 23) — and an install still on the nodes Nexus shipped now picks working nodes for itself on the next launch (see the DX-cluster entry under Added); a list you edited yourself is left alone. Until you upgrade, add either one under Settings › Logging & Connectors › Integrations & Feeds › Spot Sources: dx.w1nr.net:23 is in the + Add a known node… presets, and dxspots.com:7300 goes in with + Custom. Phone spots return at once.

  • The Listen button in the Remote browser did nothing.

    Opening the workspace replaced the connecting page with the operating one, and that swap permanently closed the audio channel before the button you could see was ever wired to it — so it looked enabled and sent nothing, every session. It now hands the channel back instead of closing it, and Listen works. If you were already listening while the page was still connecting, audio stops once at that moment and you press Listen again.

  • One slow read no longer ends a Remote session.

    A background read that took longer than three seconds — the needed board, a rare-DX lookup, a parks fetch — tore down the whole connection. Every control then greyed out, commands were refused, and audio dropped. Those reads now fail on their own and the session carries on.

  • A busy station no longer costs you control.

    A momentary refusal while the radio was busy read as "control was lost", dropping audio with it. A busy answer to a routine check now keeps your control and your audio.

  • Remote no longer turns itself off permanently after a hiccup.

    A single message the station could not read switched Remote off and remembered it off, so the only fix was to walk to the radio and turn it back on — which is the one thing you cannot do from somewhere else. The station now reconnects and retries; only a genuine refusal of access turns Remote off.

  • A slow moment at the station no longer stops the instrument feed for good.

    A late credit on the live feed dropped the station's whole connection, taking every browser with it. The feed now waits for the credit and picks up where it left off.

  • Remote: the amplifier, DSP and decode-depth controls no longer blink off when a clock is corrected.

    When a wall clock stepped at either end, the next station reading looked as though it had arrived before it was sent, and the browser threw away the good reading it already held along with it, so every control that depends on the radio's live readings (the amplifier buttons, the Phone DSP buttons, the FT decode-depth chips, the RX offset box) went blank until the next one came in. The odd reading is still never shown and the clock is still re-synced, but the reading already held now stays until it is three seconds old. A reading that arrives late still blanks them at once, because a late reading means the link has stalled.

  • A delete or edit from a Remote browser could make the shack's Logbook delete or overwrite the wrong contact.

    The shack's list loaded once and remembered each row by its position; when the browser deleted a contact above it, every later row moved up one, and the next Delete or Edit at the shack went to the contact now sitting at that position — deleted, or rewritten with another contact's details — while the message on screen named the one you meant. The shack now identifies a contact by what it is, as the browser already did, not where it sits: an action from the shack is refused with a "reload the log" message if that exact contact is no longer there, and the list refreshes itself whenever the log changes, so that refusal is rare.

  • Editing a WWFF contact from a Remote browser no longer turns its park into a POTA one.

    The hosted edit form only knew POTA and SOTA, so saving any correction — a grid, a name — rewrote the park's program to POTA. The program now rides through exactly as the log holds it, as it already did at the shack.

  • The Remote edit form no longer offers four fields it could not save.

    My grid, Rig, QSL sent and Card received were shown in the hosted form, took your entry, and dropped it under an "updated" message. They are shack-only for now; QSL sent and Card received are still on the hosted row menu, where they do save.

  • A Remote tab left in the background now lets the station rest, and there is a switch for when you don't want that.

    A Remote browser tab sitting behind your other windows kept the whole station session running, hour after hour, with nobody looking at it. Sending the tab to the background now pauses the feed the same way it already stops the audio and hands back station control, and bringing the tab back to the front picks it up again straight away. If you are watching a frequency on a second screen, Keep watching — next to Listen in the Remote header, or in that header's settings panel on a narrow screen — keeps it running; the choice is remembered for that station. The feed never pauses while your browser can still stop a transmission, and coming back from a pause says so.

  • A Remote browser that lost its station stopped hammering at it once a second.

    When the connection dropped after the service had already answered, the browser retried exactly once every second, forever — a station that was off, or a shack PC mid-reboot, was called thousands of times an hour for nothing. It now waits longer between tries the longer the trouble lasts, up to half a minute, and goes straight back to trying the moment you look at the tab. That wait only builds up while the station has really gone quiet: one that is still sending you data reconnects straight away. This matters because the browser itself sometimes drops the link after a slow reply, and a station that was answering the whole time should not be the one made to wait for it.

  • A hunt that could not be set now says so

    — on the Connect map, in the main window and in a torn-off one. When the feed's spelling of a reference was one Nexus could not read, Work went ahead with the QSY in silence and the contact was logged with no park at all — the only sign was its absence, hours later. It now tells you, and still works the station so you can add the reference by hand.

Every version