listen with mode "chirpmodem" adds the ChirpModem demodulator with SF9 at
62.5 kHz, placed 56 kHz clear of the device's DC spike in a baseband of at
least 256 kS/s. It is not a scan mode (no channel raster). Frames go to
get_packets.
listen sent the mode's bandwidth as rfBandwidth to every demodulator. A
mode whose settings give it under the demodulator's own key ("bandwidth"
for ChirpModem) now keeps it, so no unknown key is reported.
The get_packets, server instructions, readme and receiving guide texts
list chirpmodem.
Verified over the air: listen on 223.60 MHz through an RTL-SDR, the
ChirpModem modulator transmitting from a Pluto through 40 dB: 5/5 frames
in get_packets, +13.2 dB, -12.0 ppm.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New channel Tx plugin modchirpmodem, built on libchirpmodem (format v1),
modelled on modpacket: plugin, settings, baseband, source, GUI, web API
adapter, readme.
- A transmission is the message's frames, separated by a gap, then a CW ID
with the source callsign. It is generated in full on a one-thread
generator pool, never on the DSP thread, and posted back to the baseband;
the source plays it out at the channel rate. A result made for an old
channel rate or offset is generated again; one on air when the rate or
offset changes is cut off and reported.
- Policy is libchirpmodem's: a licensed frame without a source callsign is
refused (EPOLICY) and CW ID above 20 WPM is rejected; no encryption
option. CW ID can be turned off in the licensed profile, with a warning in
the GUI and the report. The unlicensed profile is shown in red.
- Settings: SF, BW, payload mode, profile, src, dst, message, repeat, gap,
amplitude, CW ID and its speed. The GUI shows airtime per frame and the
largest body that fits (chirpmodem_frame_seconds, chirpmodem_max_body).
- Web API / MCP: ChirpModemMod.yaml (settings, report, actions), SWG
classes, ChannelSettings/Report/Actions and model factory entries cloned
from PacketMod (plain setters, safe in the request mapper's reset lists),
webapiutils, webapirequestmapper, webapi.qrc. Action {"tx": 1} with
optional src, dst and message for one transmission.
- Channelizer asked for 4 x BW, as chirpmodem_tx_new needs fs >= 2.2 x
(BW/2 + |IF|) and the channelizer leaves a residual offset.
Verified: FileOutput -> cm_rx 3/3 frames, +17.4 dB, 0.0 Hz; FileOutput ->
FileInput -> demodchirpmodem 3/3 in get_packets; over the air (Pluto at
-30 dB through 40 dB to an RTL-SDR, 223.58 MHz) 15/15 frames received by
demodchirpmodem, matching cm_rx on SDRangel's recording of the same signal
(+17.0 dB, -2713 Hz).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On a DSPSignalNotification the GUI narrows the Δf dial to ±rate/2. With
the first notification at a device's default rate (48 kS/s), an offset set
afterwards was clamped in the display only (signals blocked), and nothing
put the real value back when the range widened: the dial showed +24,000
while the channel was at +250,000. setValueRange was also not signal
blocked, so a clamp there could have sent the clamped offset to the channel.
Block the dial's signals around the range change and show the offset from
the settings again.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New channel Rx plugin demodchirpmodem, built on libchirpmodem (format v1),
modelled on demodselcal: plugin, settings, baseband, sink, GUI, web API
adapter, readme. Rx only.
- The channel is decimated to at least 2 x BW; the library is given the
delivered rate and the signal's residual offset in it (if_offset).
- The chirpmodem_rx handle lives in the sink and is used only on the
baseband thread under the baseband mutex. Settings changes mark it stale;
it is rebuilt once per message, never during chirpmodem_rx_push().
- Frames are copied in the callback and posted to the channel, which
forwards them to the GUI and publishes OK frames (whole MAC frame) on the
"packets" pipe.
- GUI: SF, BW, effort, auto effort, report failures, max bytes; frames table
(UTC, status, profile, src, dst, SNR, offset Hz/ppm, mode, length, text,
hex) with licensed-profile policy flag; channel spectrum.
- libchirpmodem found with find_package(chirpmodem 0.1 REQUIRED), hint
CHIRPMODEM_DIR or ~/.local; linked statically. Builds in SERVER_MODE.
Web API / MCP support:
- ChirpModemDemod.yaml (settings and report), SWG model classes, entries
in SWGChannelSettings/SWGChannelReport/SWGModelFactory, webapiutils,
webapirequestmapper and webapi.qrc.
- MCP data feed: listen to sdrangel.channel.chirpmodemdemod and decode the
MAC frame (flags, base-40 callsigns, body) for get_packets.
Verified on the 2026-10-23 OTA capture (ota.cu8 -> .sdriq, FileInput):
49/49 frames, identical to cm_rx in order, source and body; last-frame SNR
and offset match (-10.1 dB, -2651.5 Hz, -11.86 ppm).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
After adding the SFO drift (4.25 * sfo_hat) to the fractional STO, the
value was wrapped back into [-0.5, 0.5] but the removed whole chip was
discarded, misaligning subsequent symbols by m_osFactor samples. This
mostly bites at high SF with clock drift, where the drift term can
reach ~0.3+ chips.
- Fold the wrapped chip(s) into the returned FIFO advance (-chips * os).
- Wrap via loRaRound() (half away from zero), so |sto_frac| >= 0.5 now
wraps; previously exactly 0.5 did not. Equivalent timing either way.
- Reset non-finite STO before rounding (float->int cast of NaN/inf is UB).
- Log when the FIFO advance is clamped to 1.
Other applications (e.g. SoapyPlutoSDR, used by gqrx) leave the RX LO
powered down when they close, so SDRangel only received the AD9361's
own noise. In FDD mode the Tx LO also stays up while no stream runs,
radiating LO leakage when the Tx device is merely opened.
Set the LO powerdown attribute explicitly: power the LO up when a
stream starts, down when it stops or fails, and power the Tx LO down
when the Tx device is opened. Applies to the input, output and MIMO
plugins.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
RTTY, NAVTEX and ILS demods read logFilename/logEnabled from
getAdsbDemodSettings(), which is null for those channels, so a PATCH of
either key crashes SDRangel. Also fix M17 Mod reverseAPIChannelIndex
(read from NFM Mod settings), Audio CAT SISO settings get and XTRX MIMO
reverse API settings and report (Input instead of MIMO types), and
BladeRF2 Output title (BladeRF1 settings).
libiio 1.x installs its header as include/iio/iio.h instead of
include/iio.h. Look for the 1.x layout first and fall back to the 0.x
one, and set LIBIIO_V1 so plutosdrdevice can select the API at compile
time. The bundled external build (v0.26) keeps the 0.x path.
The input and output plugin headers only need the iio_buffer type, so
forward-declare it instead of including iio.h.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Resolve GCC -Wreorder by moving m_spectrumBinHz before
m_autoUnlockedSamples in the constructor initializer list to
match declaration order in the header. C++ initializes members
in declaration order regardless of initializer list order, so
keeping the list aligned avoids misleading initialization
sequences and helps prevent future initialization bugs.
Signed-off-by: Robin Getz <rgetz503@gmail.com>
Resolve GCC -Wtype-limits warnings by removing upper-bound
comparisons that are always true or false for uint16_t.
The port's type already enforces the maximum value of 65535;
validation of wider input values remains in place before casting.
Signed-off-by: Robin Getz <rgetz503@gmail.com>
Resolve the compiler warning for the unused frequencyHz parameter
in processAcquisitionSample(). Use Q_UNUSED to explicitly document
that the parameter is intentionally unused without changing behavior.
Signed-off-by: Robin Getz <rgetz503@gmail.com>
Resolve GCC -Wmissing-field-initializers by explicitly initializing
the onEdit callback to an empty std::function. This ensures the
callback starts in a defined empty state and avoids relying on
implicit aggregate initialization behavior.
Signed-off-by: Robin Getz <rgetz503@gmail.com>
The GUI is set up on the rollup contents widget, so Qt does not
auto-connect the on_meshPreset_currentIndexChanged and
on_meshApply_clicked slots. They were never called: choosing a preset or
clicking Apply had no effect, USER did not enable the manual LoRa
controls and the preset went back to the saved EU_NARROW on the next
settings refresh. Connect them explicitly as the MeshCore modulator and
Meshtastic demodulator already do.
Fixes#2874
Explicitly delete the copy constructor and copy assignment operator for
UDPSourceSource, which manages resources that should not be shared through
implicit shallow copies.
The implicitly generated copy operations could leave multiple UDPSourceSource
instances referring to the same resources, potentially causing invalid
ownership or double deletion.
Since UDPSourceSource is not intended to be copied or assigned, prohibit
both operations at compile time to prevent unsafe resource ownership.
cppcheck identified the missing copy constructor and assignment operator
with its noCopyConstructor and noOperatorEq warnings.
Signed-off-by: Robin Getz <rgetz503@gmail.com>
Explicitly delete the copy constructor and copy assignment operator for
DVBS, which manages resources that should not be shared through implicit
shallow copies.
The implicitly generated copy operations could leave multiple DVBS
instances referring to the same resources, potentially causing invalid
ownership or double deletion.
Since DVBS is not intended to be copied or assigned, prohibit both
operations at compile time to prevent unsafe resource ownership.
cppcheck identified the missing copy constructor and assignment operator
with its noCopyConstructor and noOperatorEq warnings.
Signed-off-by: Robin Getz <rgetz503@gmail.com>
Explicitly delete the copy constructor and copy assignment operator for
ATVModSource, which manages resources that should not be shared through
implicit shallow copies.
The implicitly generated copy operations could leave multiple ATVModSource
instances referring to the same resources, potentially causing invalid
ownership or double deletion.
Since ATVModSource is not intended to be copied or assigned, prohibit both
operations at compile time, turning potential runtime errors into
compile-time errors.
cppcheck identified the missing copy constructor and assignment operator
with its noCopyConstructor and noOperatorEq warnings.
Signed-off-by: Robin Getz <rgetz503@gmail.com>