Repository navigation
v0.3.9
A 3D band scope and a Matrix rain screensaver, both on by default, TX only inside the amateur bands, the WR12 2 m receiver on the X6200's HOST port, the X6200's own PA bias, XARA as a third Winlink modem (experimental), Contestia, the Amber Trey theme, CapsLock shortcuts on a keyboard, Hide buttons, Pat 1.0.1, and fixes for FT8 QSOs lost when you moved on, a PTT that stayed on, LoTW duplicates and the web Memories page. No new card image: update in the radio, and read Good to know below. By default the radio now transmits only inside the amateur bands.
Which file. The X6100 takes OpenX6X00-x6100-v0.3.9-sdcard.img.zip and updates itself with OpenX6X00-x6100-v0.3.9-update.tar.gz. The X6200 takes OpenX6X00-x6200-v0.3.9-sdcard.img.zip and OpenX6X00-x6200-v0.3.9-update.tar.gz.
Card image or update. On a radio running v0.2.x or v0.3.x, Settings > Info > Update (Check, then Install) or the web configurator's System > Update brings everything in this release. The X6100 card image of v0.3.6 and the X6200 card image of v0.3.4 are still the newest these radios need:
- Nothing since v0.3.8 changes the kernel, its modules or U-Boot. The X6100 image is built from the same buildroot tree as v0.3.8, at the same commit. The X6200 tree changed one package, its FT8 library: ft8_lib 73b0db8, the X6100's, where it had 117022b. The card image level the radio asks for is unchanged (2 on both radios).
- Every new or changed program and file comes in the update bundle: the GUI, the web server, the start loop and its init scripts (S00adata, S50gpsd, S95gui), x6100-winlink, x6100-wr12-bypass, ardopcf, Pat 1.0.1, and XARA with its licences. The new DATA partition script runs at the next start after the update.
- The FT8 library (
/usr/lib/libft8.so, ft8_lib 73b0db8) now comes in the update bundle too, on both radios. Up to v0.3.8 an X6200 updated in place kept its card image's 117022b (the X6200 images up to v0.3.8 have it), with which a CQ with a compound or long call went out as free text or not at all (see Fixed). The X6100 has had 73b0db8 since its first image: nothing changes for it. Checked on the bench radios.
Who needs the card image. As before, Settings > Info > Update and the web Update page say when a card is older than the image the radio wants: an X6100 written before v0.3.6 (it lacks the new boot loader and the WiFi driver fix), or an X6200 written before v0.3.4 (it lacks the WiFi driver fix). A radio that no longer starts far enough to update can take this release from its card image. A new image keeps the settings and logs on the DATA partition, your own boot picture too; download the Backup from the web configurator first all the same. On a card whose DATA partition was never formatted (see A new card's DATA partition below), the settings are still on the system partition, which a new image overwrites: update in the radio first, and the next start moves them onto DATA. New radios: write the card image with balenaEtcher, cancel Windows' offer to format the card, and start the first time on external power.
Good to know.
- Coming from v0.3.7 or older: read the Good to know of v0.3.8 too.
- Reload any Panel or Screen page that was open during the update.
- Ports 50003 and 50004 (#58). Since v0.3.8 the https Panel and Screen page open their sockets on port 443, the page's own, and v0.3.8's notes said the two old ports would go in this release. They are closed now: reload a Panel or Screen page left open from v0.3.7, which says "Connection lost" until then. Checked on the bench radios.
- TX only inside the amateur bands (#121). By default the radio now transmits only inside the amateur bands, on the range of a transverter you turned on, and never on the X6200's receive-only ranges. Up to v0.3.8 it transmitted anywhere from 0.1 to 55 MHz.
- One table for every country: ITU Regions 1, 2 and 3 together, with the wider national bands (60 m, 8 m 40-45 MHz, 2200 m, 630 m, the Alaska emergency channel). Your licence still decides inside it.
- Outside it the screen says "No TX on 27.185 MHz", then "Out of band: Settings > TX Cal". Refused: PTT (the key, the walkie screen, CAT answers NG, TCI, the browser Panel), the digital apps, the voice messages, Winlink, CW, the tuner, TUNE and the SWR scan. A transmission that moves out of the bands is stopped.
- What the base keys by itself (the hand mic's PTT, VOX, the KEY jack) the radio cannot refuse. Outside the bands it holds the base at 0 W, with VOX off and KEY Train on. On the X6200 it also shuts the base's own band table (next item).
- The SWR scan's sweep stays inside the band: 2500 kHz on 40 m sweeps 7.000 to 7.300 MHz. The CW Trainer still sounds its sidetone outside the bands, at 0 W.
- Licensed outside the bands (MARS, CAP and the like)? Settings > TX Cal > TX PROTECTION > TX outside ham bands, also in the web Display page's Transmitter card. It is off on every radio after this update, whatever it was before. A backup restored afterwards brings its own.
- The X6100's two transverters are off by default now, as the X6200's already were. Up to v0.3.8 they were on from the start (2 m and 70 cm through a 28 MHz IF): with nothing connected, TX on 2 m or 70 cm sent the 28 to 34 MHz IF into the antenna. A transverter you set up is kept. One you used as it came: set its shift again in Settings > General > TRANSVERTERS (116 MHz for 2 m, 404 MHz for 70 cm).
- Tested on the bench X6200 into a dummy load (4 and 5 Nov): TCI refused, the side PTT and the ATU refused with "No TX on 27.185 MHz", the hand mic at 0 W; with the switch on TUNE gave 0.9 W, off it was refused again; in the band 4.9 W. On 6 Nov with a wattmeter, see the next item. Not on the air, and not on an X6100: what its base sends at 0 W is not measured.
- X6200: the base itself keys nothing outside its own band table where the radio refuses TX (PR #176 by ea4iln-debug, with our follow-up; #134).
- Where: outside the amateur bands, on the receive-only ranges, while the WR12 receives, and during Birdie Cal; not while the CW Trainer is open outside the bands, for its sidetone. The base's table is 160 to 6 m, 60 m only 5.332 to 5.405 MHz, as Xiegu's firmware always has it. So the hand mic's PTT, VOX and the KEY jack key nothing there. Everywhere else it stays open, also with TX outside ham bands on, and on a transverter's IF past 29.7 MHz.
- ea4iln-debug found where it is: byte 280 of the base's calibration block. It is kept in the base's memory only, and a normal power-off gives the base its own block back. The radio writes the block only when two reads agree and the rest is as at the start, never during TX, and half a second after the dial stops. Each change is a line in /mnt/base_i2c.log ("open TX 0: shut", "open TX 1: opened"). If a change does not go through, the screen says "Base's band table not set: TX held at 0 W only".
- One gap: the hand mic's PTT, VOX or the KEY jack within half a second of tuning from where TX is refused onto 2200 m, 630 m, the Alaska channel, 60 m outside 5.332-5.405 MHz or 8 m sends nothing until let go.
- Tested on 6 Nov on the bench X6200 (base 1.0.8), with an external wattmeter on a dummy load. The hand mic sent nothing on 145.000 MHz NFM (WR12 receiving, switched off on HOST, unplugged), nor on 5.060 MHz. The front PTT and the hand mic sent 1 W on 7.100 and 5.300 MHz, 0.8 W on 40.680 MHz, and very low power on 136 kHz. A TCI key 0.1 s after 7.350 to 5.300 MHz: 1 W at once. The knob ran smoothly over 14.350 MHz, one shut and one open each way. Birdie Cal shut the table and opened it after. The CW Trainer sounded its sidetone on 9.500 MHz with no RF. A GUI killed with the WR12 out of its bypass started again with the table shut. The calibration block was the same before and after but byte 280.
- No TX without data from the base (#152). What the radio keys itself is refused while its base sends no data: none for 3 s, or none yet since the start. The screen says "No TX: no data from the base", then "Baseband: Settings > Info". The SWR protection and the power reading come from that data.
- After 30 s without data (60 s after a start) the screen says "No data from the base" and three lines more.
- A tune or an SWR scan the data stops in is given up after 3 s. A transmission under way is not cut: the radio goes back to receive once it is let go. Up to v0.3.8 a tune without data stayed stuck and blocked the digital apps until a restart.
- An X6200 on base 1.0.8 sent its first data 16 s after a start (#47): TX asked before that is refused. Ask again once the waterfall moves.
- The hand mic's PTT, VOX and the KEY jack still go out without data. Don't transmit that way into an antenna you don't know.
- The X6100's POWER key now switches the radio off without data too. An X6200 POWER key the data stopped on counts as let go, so a tap cannot turn into a power-off.
- Host test (tools/base_guard) and the PC build only.
- Keyboard (X6100 and X6200, USB or Bluetooth): the F keys no longer do anything, nor do PgUp and PgDn (v0.3.8's manual had them, but they never worked on the radio). The shortcuts are CapsLock held + a key: L logs a QSO (in CW the typed line), K dial lock, P screenshot, S Settings, F frequency box, Up/Down band, 1-5 CW messages, 1-8 JTTY macros, Space PTT on/off (lets go after 3 minutes, when the base stops sending data, or when the keyboard goes away). CapsLock tapped alone is still Caps Lock. Scroll Lock (the dial lock) and Print Screen (a screenshot) now work on the main screen. (#175, F4ESO) Tried on the bench X6200 on 5 Nov with a USB keyboard: the CapsLock chords and the PTT toggle worked. Not tried on an X6100 yet, nor the Caps Lock light and a Bluetooth keyboard's link lost.
- Field Ops Log QSO takes the reports in the order call, sent, received, as the main screen's quick log does; up to v0.3.8 it took the received one first. (#175)
- The 3D band scope is on by default (PR #161 by timerealm). An update from v0.3.8 starts in 3D: the waterfall area shows the last 6 seconds of spectrum as traces in perspective, the newest in front, coloured by strength. Switch it off in Settings > Display > WATERFALL > 3D, on the web Display page, on the MFK 4:4 page, or with a key set to "WF 3D on/off". Once off, it stays off over later updates. Under Amber Trey it is off and greyed. 3D glide off takes less CPU.
- The GUI's CPU, of one core: on the X6100 2D 42 %, 3D 69 %; with FT8 open 2D 61.5 %, 3D 68 %, 3D glide off 61 % (6 Nov). On the X6200 2D 23 %, 3D 50 % (6 Nov); 3D with glide off 31 % (5 Nov).
- Tested on both bench radios on 6 Nov: the six screens looked right (3D, spectrum 120 px, 1bit, Amber Trey, buttons hidden, 2D back to 3D). The X6100 was in 3D after the update, and FT8 decoded with 3D on, on its antenna.
- A screensaver, on by default. Matrix rain, as on the VFO-Knob, after 5 minutes without use; the screen never goes dark until you set Screen off. Settings > Display > BACKLIGHT: Screensaver (None, Matrix rain), Screensaver (min) 1-120, Screen off (min), Screen always off; also on the web Display page.
- Updating: Screen off was 5 minutes by default. A radio where it was never changed takes the new defaults: the screensaver after 5 minutes, never dark. A radio where it was changed keeps going dark at its time, with the screensaver before it: at 5 minutes, or a minute before the dark when that comes at 5 or sooner (Screen off 3: the screensaver at 2, dark at 3). Screen off 1 leaves no room: no screensaver there. Screen off 0 (never dark): the screensaver after 5 minutes. To keep the screen in view, set Screensaver to None.
- Screen always off on, with Screen off never changed: Screen off stays 5, so POWER shows the screen 5 minutes, as before.
- The first key, knob turn or hand mic key on the screensaver only brings the screen back. Any TX ends it at once, and none comes while on the air. A Remote > Screen viewer and a screen recording keep it away.
- Tested on 6 Nov. Bench X6200: the rain came, a key woke it with no leftover frame. The GUI's CPU, of one core: lit with 3D 47 %, the rain 26 %, dark 12 %. Bench X6100: a Screen off of 5 set before the update came up as the screensaver at 4 and dark at 5; the rain took 55 % of one core with FT8 open (3D 68 %).
- X6200 S-meter: a band never calibrated still starts at -22 dB, as on v0.3.8: calibrate a band at S9 before Measure. With ATT on, a band never measured adds back 59, 51 or 34 dB by frequency (under 2.5 MHz, to 5 MHz, above) instead of v0.3.8's one 15 dB (#125). On both radios the CW and SSB parts of 40 m and 20 m now share their calibration, on the X6200 their attenuator figure too. The attenuator figure goes up to 70 dB, one per band; v0.3.8's single value stays stored, for an older version. The X6200's preamp default is 16 dB (was 19). On the X6200 Measure wants about S9 to S9+20 (about -55 dBm from a generator). Tested on the bench X6200 into a dummy load (5 Nov): about -118 dBm at -22 dB. The attenuator starts are not checked against a generator here (thanks to ea4iln-debug and DL9RP for theirs).
- X6200 on battery: the power-off at 6.8 V, and a supply under 12.0 V no longer stops it (#160, ea4iln-debug). The base cuts the power by itself at about 6.4 to 6.6 V, without a word and without saving. The radio's own power-off now comes first, at 6.8 V (it was 6.0 V), minutes before. On a supply the radio reads under 12.0 V it goes on: "Low battery, DC IN under 12 V! Turning off in 30s." Xiegu's 12 V 1 A charger reads 11.7 to 11.8 V there, and in #160 it did not hold a flat battery up. Charge a flat battery with the radio off first, or use a supply that reads 12.0 V or more. When the base reports no battery, the power-off never starts. The X6100 keeps 6.0 V, and any supply stops it. Host test (tools/battery) only.
- X6200: the H-22's digit keys go to the band printed on them (#56, joanpao; PR #136 by ea4iln-debug; #145, gilcomnz-eng). 3 to 9 each go one band lower than in v0.3.8: 3 is 60 m, 4 (printed 7) is 40 m, 9 is 12 m. The dot is 10 m. Up to v0.3.8 they took the X6100 mic's table: 7 MHz gave 30 m. The X6100's hand mic on an X6200 keeps its own table, 3 = 40 m. Tested on the bench X6200 (5 Nov): H-22 4 to 40 m, the dot to 10 m; the X6100 mic's 3 and 9.
- KEY Train sends 0 W in CW, in every key mode (#127, dl1sdz). With Train on, in CW and CW-R, the radio holds its TX power at 0 W. In Manual the X6200's base transmitted at PWR from the KEY jack up to v0.3.8. The first key-down says "KEY Train: sidetone only, 0 W". To transmit CW, turn Train off. Tested on the bench X6200 on the air (1 Nov): 0.0 W with Train. A paddle in Manual is not tried (testers are welcome on #127).
- X6200: the digital apps' TX level is learned once more (#126 joanpao, #123 orome). Every X6200 starts once from the default level (0 dB) and learns it again. Its ALC reading no longer turns the level down, no step up comes while the base's power is still rising, and the level never goes past the codec's full scale. Up to v0.3.8 it could climb to 33.4 dB, a clipped square wave. Tested on the bench X6200 on the air (1 Nov): FT8 at 2, 5 and 7 W held its power over eleven transmissions on 10 m; host test tools/txlevel.
- Popups Freq/Info/ALC is now Popups Freq/Info/TX (#166, orome). Its third switch is the TX panel. Off, the panel no longer comes up, and the S-meter box shows PWR, SWR and ALC while transmitting. Up to v0.3.8 the panel came up whatever the switch. TUNE for an external tuner always shows the panel; the digital apps keep their TX strip. PC build only.
- Test beacon takes two presses (#135, sa5qed), in Field Ops and in Bands. The first press takes the radio to the band chosen and listens; a second within 10 s sends TEST DE there, then the radio goes back. Up to v0.3.8 it sent on the wrong band. Tested on the bench X6200 on the air (1 Nov): two presses, the stop and the 10 s timeout. Use band (the Bands app, Field Ops' Bands screen) now keeps CW in CW; any other mode gets the band's phone frequency in its sideband, USB on 60 m.
- LOCK holds the channels too. With LOCK on, or while an app holds the frequency or the mode, a channel recall (the CHANNELS window, the hand mic's UP/DOWN) is refused: "Channel not tuned: dial locked". Up to v0.3.8 the hand mic's UP/DOWN could move the VFO in the middle of FT8. The X6100 mic's XFC no longer swaps the VFOs while an app holds them, and SSTV now holds the frequency, band and mode while a picture goes out. PC build and the audit harness only.
- WSPR always opens with TX % at 0, also after a power-off with WSPR open: no beacon goes out until you press TX % or TX next. Up to v0.3.8 a saved TX % could send a beacon unasked. PC build only.
- Winlink's SWR hold is on one band (#147, edegraaff). After high SWR stops a Winlink transmission, the 5-minute hold covers that band and 100 kHz around it; another band transmits at once. A high SWR in another transmission (a PTT, FT8) holds nothing for Winlink. Host tests only.
- A new card's DATA partition is formatted now (#47). Up to v0.3.8 the first start of a newly written card (new, or wiped by a tool) made the third partition but never formatted it: the radio kept its settings and logs on the system partition, which a new card image overwrites, and a PC showed an unknown partition. The first start after this update formats it as DATA and moves those files onto it. Leave that partition alone on the PC. A DATA that worked and went bad is never formatted now: the start trace and data_fsck.log (Files page) say what to do. Tested on the bench X6200 on 1 Nov, on a test card: the start formatted the partition, moved the files over, kept the callsign and the Panel's pairing. The later hardening: host test tools/data_partition, 215 cases on fake cards with both images' tools.
- Going back to v0.3.8 (the web Update page). The radio transmits anywhere from 0.1 to 55 MHz again. The X6100's two transverters are on again (2 m and 70 cm through a 28 MHz IF) unless you set them. The X6200 sends the fixed PA bias 3360/3820 again and opens its base's band table at the start, its S-meter takes v0.3.8's single attenuator value again, and its digital apps the TX level v0.3.8 had learned. No 3D and no screensaver; the screen goes dark at Screen off as set (5 by default). The F keys work again, CapsLock's chords do not.
- The FT8 library on an X6200: Rollback on the Update page puts back the one it had, its card image's 117022b. v0.3.8's bundle installed over this release keeps the newer one, and v0.3.8 runs fine on it, as the X6100's v0.3.8 always has.
- With a WR12: before a Rollback or installing an older bundle while the WR12 receives, turn WR12 on HOST off or tune below 54 MHz. Older versions never drive the WR12, so they never put it back in its bypass.
New.
- WR12 receiver on the X6200's HOST port (#134, dl1sdz; PR #176 by ea4iln-debug): from 54 MHz up the radio tunes the WR12, takes it out of its bypass and listens on its IF; nothing transmits while it receives.
- Settings > General > WR12 RECEIVER > WR12 on HOST, also on the web Display page (Receiver card); off by default. The radio finds the WR12 by its USB id (0483:5740), sets it up with AT commands and keeps it in its bypass below 54 MHz. It listens on the WR12's 5.06 MHz IF with I and Q swapped; the screen, CAT and TCI keep the dial's frequency. A WR12 tag in the frequency strip says it is in use.
- The X6200's two transverters stay off by default, as in v0.3.8, so 2 m and 70 cm are the WR12's. gpsd no longer takes the WR12 for a GPS at boot. A USB GPS with the same USB id (STM's stock 0483:5740) is left alone at boot too, whatever WR12 on HOST says.
- A GUI that stops puts the WR12 back in its bypass; after a crash the start loop does.
- One ERROR to the WR12's keepalive (AT+AYT) now goes into the log only; two in a row are said on the screen, once. On 6 Nov the bench X6200 showed "WR12: ERROR to AT+AYT" once, right after the WR12 was set up again, and the next answer was OK. Host test tools/wr12; with v0.3.9 on the bench X6200 the WR12 was set up and followed the dial to 2 m. Out of its bypass the WR12 draws more from HOST: with an Ethernet adapter beside it, use a hub with its own power supply (on our bench the adapter dropped out the moment the WR12 left its bypass; docs/WR12.md).
- docs/WR12.md, by ea4iln-debug: the WR12 step by step, in plain words.
- Known: RIT and the CW offset come out the wrong way through the WR12 (RIT -500 acts as +500; CW and CW-R give the same tone). USB and LSB come out right, the image about 34 dB down.
- Tested: ea4iln-debug on his X6200 with a WR12 (4 Nov): 2 m, FM broadcast at -98 dBm on 95.9 MHz, 40 m at 2 W, A/B, a restart, a cold start. Our bench X6200 (5 Nov): the set-up, the switch, an unplug, TX refused on every path with no RF, 40 m free after. 6 Nov with the wattmeter: see X6200: the base itself keys nothing under Good to know; the first AT+BYPASS=0 answered every time after the 10 ms gap fix. Not done yet: reception on an antenna next to an HF reference, a power-off with the WR12 out of its bypass.
- The radio's own PA bias on the X6200 (#173, ea4iln-debug), as Xiegu's firmware reads it from the base; power and current measured equal to before. Versions 0.3.4 to 0.3.8 sent 3360/3820 to every X6200, the values of another radio. Settings > Info > FIRMWARE > PA bias shows it, with where it came from ("3226 / 3958, the radio's own" on our bench). Coming from v0.3.8 the base still holds 3360/3820 ("own or an old version's"): switch the radio off and on once. A normal power-off is enough, battery in and supply connected too: ea4iln-debug read the base's whole block back as at a cold start (#173, base 1.0.8). Tested on the bench X6200 (4 and 5 Nov): the power the same within 0.1 W at 1, 5 and 8 W on 40, 20 and 10 m. ea4iln-debug measured the PA's idle current on his X6200: about 0.82 A with either pair (our bench had no ammeter).
- X6200: register 32 at Xiegu's value (#53, 123tcpip). At each start the X6200 sets its base's register 32 to 0x0a, as Xiegu's firmware does; it was 0. It is set for the short RF pulse at the start of each transmission that 123tcpip's captures showed. Whether it changes the pulse is not known yet: 123tcpip will repeat his captures with this release. On our bench (the radio's meter, 15 starts each, 5 Nov) the starts reading over 1.5 times the set power went from 4 to 0 on 6 m at 2 W, and from 5 to 1 on 40 m at 1 W.
- XARA, a third Winlink modem (experimental) (#120, yo4cil). XARA 0.1.3 by x0rro speaks VARA HF, so the VARA HF gateways answer it. It comes built from source, with three changes of ours: the radio's GUI as its sound card and PTT, a CW ID with any callsign when Pat asks (offered upstream, x0rro/xara#42), and a cancelled call that ends at once. Operating > Winlink > Modem: XARA, its bandwidth (500, 2300 or 2750 Hz) and a VARA HF licence box, off by default: off, XARA runs VARA's free speed levels. The radio only calls: it answers no call. Every key goes through the radio's own TX path: the amateur bands, at most 7 W, the SWR hold. Tested Pat to Pat on a PC, at the free and the full speed. On the bench X6200 it listened at 6-7 % CPU and keyed nothing. Not on the air from a radio, not on an X6100, not against VARA.
- Contestia (PR #154 by Spyyena; #124, johny1time-code). Eight presets in the Olivia window's Mode button (CT 4/250 to 32/1000), made to fldigi's modem, capitals only. Logged as CONTESTI, and in the text log. Tested on the bench X6200: CT 4/250 went out (1 W, 13 s). Not yet against fldigi on the air.
- Amber Trey theme (PR #98 by om0wt): an amber segment LCD over everything above the bottom buttons: the bar graph (S, SWR/CMP, RF/ALC), the filter graphic, the VFO in seven segments, a UTC clock. It draws no 3D. Seen on the bench X6200 (5 and 6 Nov) and X6100 (6 Nov).
- Hide buttons (#128, yo4cil). Settings > Display > INTERFACE: Off (the default), after 3, 5 or 10 s without a key; also on the web Display page. The buttons and knob lines hide on the main screen, and the waterfall goes down to the bottom. Any key brings them back; F1-F5's first press only brings them back. Seen with 3D on both bench radios (6 Nov); otherwise the PC build.
- Waterfall Fixed mode: Jump at and Jump to (#10, firatsoygul). Jump at 40 % (the default), Passband or Line; Jump to Center (the default), Other edge or Scroll; also on the web Display page. The dimmed side moves with the view. Tested on the bench X6200 (1 Nov).
- Power display % + V (#119, Flight3R): the battery's % and, on a supply, its voltage: "85% 13.6v". Tested on the bench X6200 (31 Oct).
- Your own WeFax stations (#116, ericb2B): every channel named WEFAX... shows in the WeFax app's Freq. list after the stations. A channel holds the dial: tune 1.5 kHz below the station. Tested on the bench X6200 (31 Oct).
- X6200: the H-22's side UP and DOWN keys change band, as the BAND keys. They did in v0.3.8 too: its notes and manual said they sent nothing, but base 1.0.8 sends them (ea4iln-debug).
- Pat 1.0.1 (LA5NTA), our six patches on it and a seventh: a text-only template opened as a reply has the message replied to. Forms and templates from the composer open in a window of their own. Pat's gateway list of November 2026. Its options are on its Configure page (Action > Configure). Installed on the bench X6200 (1 Nov), with Winlink off there: no connection through it on the air yet. Host: Pat's UI tests, and the build reproduced with the same SHA-256.
- ardopcf: a call nobody answers ends in "CONNECT TO FAILED". A ConReq or a Ping counts as our ID, so no ID frame keys the radio by itself 9 minutes after a call nobody answered. Host tests (tools/winlink).
- TCI for Hamlib's TCI backend and fldigi's TCI build: commands in any case (
VFO:0,0;),ddsanswered, and what a client only asks takes no control, so a logger polling through Hamlib leaves control with the browser. Host tests; not tried with real Hamlib or fldigi yet (#178). - The browser Panel: the pop-ups (the frequency keypad, lists, sliders, Filter) sit inside the Panel with its colours, in Fullscreen too; up to v0.3.8 they were see-through, and missing in Fullscreen. A status line text that says how things are ("Take control first", "On the air") goes once it is no longer so; the PTT and TUNE buttons drop a refusal once its reason goes (#58). Over Tailscale Funnel the Panel's files carry no time, so a visitor without a key no longer learns when the radio last started; a browser then loads them whole at each visit, about 265 kB. The old ports 50003 and 50004 are closed (see Good to know).
- Other boards and the PC build (cjlawson02): the panel size from BOARD_DISP_W/H with an optional landscape blit (PR #162), the radio's ranges in board.h (PR #171), pointer taps on the F-bar for a mouse or a touch screen (PR #163). Nothing changes on our two radios: both drew as before with #162 on 6 Nov (the X6200 all day, the X6100 in the afternoon), and the front keys worked after #163 on the bench X6200 (5 Nov).
- The manual: the keyboard, the screensaver, Amateur bands only, No TX without data from the base, PA bias, register 32, XARA, a PC on Bluetooth (#117, ericb2B: CAT and a speaker that plays nowhere, no audio input: use the USB cable or TCI), a paddle in the X6200's KEY jack in Manual is a sideswiper (#109, dl1sdz), and the troubleshooting of a black waterfall and of a card with no DATA drive. A normal power-off gives the X6200's base its own calibration block back (ea4iln-debug, #173). The keyboard and Amateur bands only say what the bench X6200 tried. The update section says the FT8 library comes with it, the HTTPS paragraph that ports 50003 and 50004 are closed. The README's base table: X6200 bases older than 1.0.7 are not supported.
- Builders: the update bundle carries
/usr/lib/libft8.sofrom the buildroot target (tools/update/mkbundle.py); local-build.sh says which ft8_lib goes in, and stops when the tree's pin was never installed in target/ or another version went in after it. U-Boot's tarball is made from its GitHub tag when ftp.denx.de doesn't answer, and checked against the release's SHA-256; a version with no SHA-256 stops the build. XARA builds with a pinned Rust 1.99.1. local-build.sh copies in every file that differs from the source.
Fixed.
- ADIF: ft_log.adi and the day's log write the same record and header (TIME_ON with seconds, band in lower case, ADIF 3.1.8, the real program version), so LoTW no longer sees duplicates (#174, joanpao). The Logs page writes an edited band as "20m". Host tests (tools/adif).
- X6200: FT8's CQ with a compound or long call. The update now brings the FT8 library, ft8_lib 73b0db8, on both radios. Up to v0.3.8 an X6200 updated in place kept its card image's older 117022b: CQ OT4ABCD and a CQ of more than 13 characters were refused ("No CQ: callsign ... does not fit FT8"), and CQ PA/ON6URE went out as free text, not as a CQ. With 73b0db8 all three go out as CQs. The X6100 has had 73b0db8 since its first image. Checked with both libraries on the PC; the library in the update checked on the bench radios.
- TX meters start each transmission at 0, not at the last one's power (#53); the TX panel's numbers too; CAT's SWR answer follows the CI-V scale (up to v0.3.8 SWR 5 read as 1.2); TCI's SWR starts at 1.0. Tested on the bench X6200 (5 Nov): the first reading 0.0.
- S-meter attenuator top 70 dB on the X6200, Measure says when it hit the top (#125). On the X6200 Measure also names the band it set ("Done on 40m CW and 40m SSB") and takes a carrier from about S9, and the noise floor's grey moves with the PRE and ATT figures. On both radios a power-off in the middle of Measure puts PRE, ATT and the RF gain back.
- PTT that stayed on (two keys at once): a key let go while PTT was held (a tap say) kept the radio on the air after PTT was let go, until PTT was pressed again; a key held while another was pressed lost its long press. The browser's PTT went the same way (the end of its lease, its three minutes). Each key now counts on its own. Tested on the bench X6200 (1 Nov): back to receive at the release in all four steps.
- FT8 QSOs lost when you moved on (#150, joanpao). A QSO whose report went out is kept aside for 4 slots; its late RR73, RRR, 73 or R+report still logs it, once. Force QSO save leaves the QSO current. A band, frequency or mode change no longer drops a QSO the last decodes complete. Two paths that logged a QSO twice are gone. Host test tools/ft8_qso.
- FT8 and FT4 timing and calls. A transmission starts 0.5 s into its slot, as WSJT-X sends; a late one leaves out what is past, and one more than a fifth late is not sent ("TX not sent: too late into the slot"); FT4 starts at most 1.0 s into its slot. A compound or long call of yours (PA/ON6URE) goes in full as WSJT-X sends it, so a station that has not heard you can answer; your CQ is "CQ PA/ON6URE", without the QTH. A compound call answering your CQ the WSJT-X way (" PA/W1ABC") gets your report. A CQ that cannot be sent is refused at TX CQ. Your grid is the QTH's square in capitals. A short or compound call is no longer refused at random. On an X6200 the CQ with such a call needs the FT8 library this update brings (the item above). Host tests; not checked on the air (DT, a compound-call QSO).
- X6200 CW: the station on the VFO line (#115, dl1sdz; #89). In CW and CW-R a station on the dial is drawn on the line, on the screen and in the Panel; up to v0.3.8 it was a key tone off. ZAP, the CW Tuner's bars and the decoder follow the right way. Tested on the bench X6200 (31 Oct): a carrier on the line in CW and CW-R, ZAP from R+150 to R+2.
- The CW Trainer's "?" (#144, dr4g0nized). The end of the radio's own group is never taken for your answer. The answer waits longer at slow speeds; the decoder goes down to 5 WPM while the trainer is open; long dahs at KEY Ratio over 3 are kept; a character it cannot read shows as
<?>and counts wrong. Host test tools/cw_trainer. - X6200 SWR scan reads the SWR (PR #140 by ea4iln-debug; #107): an FM carrier, each reading after the power is up, about a minute a sweep. The 1 W holds while it runs, whoever sets PWR meanwhile. Tested on the bench X6200 (1 Nov): a 50 kHz sweep, 1.25 s a point.
- Band keys stuck between 20 m and 30 m (#168, W8LID): a band whose saved frequency lies outside it opens at its start. Host harness.
- Independent VFOs: VFO B on a band new to it starts where VFO A has that band (#137, ea4iln-debug). Tested on the bench X6200 (1 Nov).
- The web Memories page (#158, orome, reproduced by ea4iln-debug): one channel above 56 MHz left the list blank, and an Add then wrote over all the channels. The page now lists every channel, changes only the one it shows, keeps the radio's own changes, and takes the frequencies the radio tunes (up to 1000 MHz on the X6200). Host tests (tools/channels, the edit-flow cases).
- Ethernet to a computer without a gateway (#169, orome) leaves WiFi on, with its internet and DNS. With a cable in, the radio no longer restarts the WiFi/Bluetooth chip when WiFi finds no network. WiFi networks restored from a backup are taken at once, no restart. Host test tools/wifi_restore.
- Battery: the % from the voltage no longer flips on a 0.1 V step (#160). An X6200 on a supply close to 12.0 V no longer starts and calls off the power-off over and over. Host test tools/battery.
- POWER in JS8 (#118, erikkt): a half-second press switched the radio off while JS8 was open; the 1 s hold is back.
- Keyboard and lists: closing the WeFax, NAVTEX or SSTV frequency list with Esc or Enter no longer reaches the freed list; a held Esc no longer flips VOL after an app closes; a key held as the keyboard goes away is let go; in SSB after CW the quick log's box no longer has the CW call in it.
- Callsign: an empty field clears it, no longer "Unsupported callsign (too long)".
- Channels: channels.json is on the card before its rename, so a power cut no longer leaves a short or empty list.
- Spectrum off: no see-through gaps under the top bars (PR #177 by cjlawson02, and our follow-up for the DRM path). Seen on the bench X6200 (5 Nov) in Amber Trey and 1bit.
- WeFax: a channel named WEFAX with only '#' after it is listed by its whole name (#116).
- The X6200's POWER key is let go when the base's data stops, so a tap cannot become a power-off.
Tested, and where.
On a radio on 6 Nov:
- Bench X6200 (base 1.0.8), with an external wattmeter on a dummy load: the WR12 and the base's band table (the hand mic on 2 m and 5.060 MHz, 7.100, 5.300, 40.680 MHz and 136 kHz, the TCI key after a refused frequency, the knob over 14.350, Birdie Cal, the CW Trainer's sidetone on 9.500 MHz with no RF, a killed GUI, the block before and after, the first AT+BYPASS=0); the 3D band scope's six screens and its CPU; the screensaver (wake, CPU).
- Bench X6100 (stock base 1.1.8): 3D on after the update and the six screens; the CPU in 2D and 3D, with FT8 open, with glide off, and the rain; FT8 decodes with 3D on its antenna; the screensaver's migration; CW and CW-R on 40 m on its antenna with the tuner: the X6200 as a second receiver heard it on the dial, and the X6100's waterfall drew it on the line.
- Both radios drew their screens as before with #162's display path.
- After the release build, on the bench radios: the closed ports 50003 and 50004 (the https Panel and Screen page still connect), the FT8 library in the update, loaded by the X6200's GUI, and the WR12 set up and on 2 m.
On a radio 31 Oct to 5 Nov (bench X6200, base 1.0.8). Into a dummy load, 4 and 5 Nov: #121's refusals and the 0 W hand mic; #173's own bias; #53's first reading 0.0 and register 32's starts; the H-22's band keys (#56, #145); Contestia TX; XARA listening; the keyboard map and PTT; Amber Trey, 1bit and the spectrum off; the front keys after #163; the WR12 (5 Nov). On the air on 1 Nov (its 10-40 m antenna with the tuner): KEY Train 0 W (#127), FT8 at 2, 5 and 7 W (#126), Test beacon (#135), the PTT release, the SWR scan (PR #140). Receiving on that antenna: on 1 Nov Jump at and Jump to (#10), VFO B (#137); on 31 Oct the CW line (#115, on a broadcast carrier), % + V (#119), WeFax channels (#116). On 1 Nov the DATA first-start fix, on a test card.
Host tests and the PC build only: no TX without data (#152), the X6200's low battery (#160), the X6100's transverters and 0 W hold (#121), the DATA partition's hardening (#47), the popups switch (#166), Hide buttons (but one look on each radio), LOCK on channels, SSTV's hold, XFC, WSPR's TX % at open, the FT8 QSO recovery and timing (#150), the CW Trainer (#144), the Memories page (#158), Ethernet without a gateway (#169), ADIF (#174), TCI's cases and reads, CAT's SWR, ardopcf, Winlink's SWR hold (#147), the keyboard on an X6100.
Not on the air yet: XARA from a radio, Contestia against fldigi, FT8's slot timing and a compound-call QSO, the WR12 on an antenna next to an HF reference. Not tried yet: XARA on an X6100 or against VARA, TCI with real Hamlib and fldigi (#178), the CW Trainer with a real key (#144), a paddle in Manual (#127), a power-off with the WR12 out of its bypass.
Not fixed in this release: on an X6100 with base 1.1.9 the waterfall draws your own CW a key tone below the line (#191, Flight3R). On the stock base 1.1.8 the bench X6100 sent CW and CW-R on the dial and drew both on the line, so the fix will be for bases like 1.1.9.
Thanks to:
- ea4iln-debug (EA4ILN) for PR #176 (the base's band table, the WR12 tag, docs/WR12.md) and PRs #136 and #140; for finding the PA bias (#173) and byte 280, for showing that a normal power-off gives the base its block back, and for measuring the PA's current; for testing the WR12 driver on his radio (#134); for #137, #160 and its battery notes, the measurements of #125, and reproducing #158.
- cjlawson02 (Chris Lawson) for PRs #162, #163, #171 and #177.
- timerealm (Calin Rigo) for PR #161, the 3D band scope.
- om0wt (Pavol Calfa) for PR #98, the Amber Trey theme.
- Spyyena for PR #154, Contestia.
- DL9RP for #125 and its S-meter measurements, and 123tcpip (Bob G6CKK) for the start-pulse captures of #53.
- F4ESO for #175; orome for #123, #158, #166 and #169; joanpao (EA5JTT) for #56, #126, #150 and #174; dl1sdz for #109, #115, #127 and #134; DC6NS for #47; Derek000 for #152; XRF03 for #121; yo4cil for #120 and #128; johny1time-code for #124; sa5qed for #135; dr4g0nized for #144; gilcomnz-eng for #145; edegraaff for #147; W8LID for #168; ericb2B for #116 and #117; erikkt for #118; Flight3R for #119 and #191; firatsoygul for #10 and #58; km4gkd for #107.
- OpenX6100's x6100_gui, whose fixes we took (WSPR's TX %, the frequency list, the held Esc, the channels' sync, LOCK on channel recall, SSTV's hold, the empty callsign, FT8's slot timing and compound calls); gdyuldin's ft8_lib; ardopcf (the ConReq as ID); Pat (LA5NTA); XARA (x0rro).
If you want to support the work: https://buymeacoffee.com/on6ure