ReadTelemetry делал 5-6 синхронных iio_channel_attr_read_* — каждое это USB
round-trip, вместе ~30 мс. Звался из MeterTimerTick, то есть прямо в главном
потоке, дважды в секунду (POLL_MS=500). Замер зонным профайлером: тик метра
занимал 5633 мкс настенного времени при пиках до 51 мс, тогда как CPU-времени
в нём было всего 140 мкс — чистое ожидание USB, ронявшее кадры.
Опрос переехал в TPlutoTeleThread: он раз в 500 мс читает те же атрибуты и
публикует снимок под FTeleLock. ReadTelemetry теперь только отдаёт снимок и
никогда не блокирует, поэтому RadioController.ServicePlutoTelemetry и
MainForm.MeterTimerTick не менялись — вся фильтрация температуры на месте.
Прежнее блокирующее тело стало PollTelemetryBlocking (только из потока).
Конкуренция за контекст libiio не выросла: UI-поток и раньше читал атрибуты
параллельно с iio_buffer_refill в RX-потоке. Сменился лишь тот, кто ждёт.
StopTelemetry идёт первой строкой Disconnect — до iio_context_destroy, иначе
поток продолжил бы читать атрибуты разрушенного контекста.
Результат: ui.meter_tick 5633 -> 110 мкс. Значение телеметрии теперь может
отставать до секунды (троттл контроллера + период потока) — на 10-герцовом
индикаторе температуры это незаметно.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Инструментация (ВРЕМЕННАЯ, не мержить в main):
- PerfLog.pas: зонные таймеры, отчёт раз в 5с в stderr (core% = доля ядра),
EWSDR_PERF=0 выкл, EWSDR_PERF_FILE=путь — дублировать в файл
- Замеры // PERF: сеть (DDC IQ/wideband), Pluto (rx_conv/tx_resamp/tx_fill),
DSP (push_iq/fexchange/slices/tx), аудио, UI (тики/draw/paint/gl_swap,
разбивка draw_spectrum по фазам)
- doc/PERF_PROFILING.md: карта зон, методика, выводы
Оптимизация CPU-рендера спектра (по результатам замеров на эфире):
диагноз — дорога не пиксельная работа, а чередование raw(ScanLine)/
canvas(QPainter) доступа к битмапу на Qt6 (~6 переходов за кадр, каждый
синкает весь битмап; полоса фильтра в тысячи пикселей стоила 4x дороже
полного memcpy кадра). DrawSpectrum перестроен в фазы calc->raw->canvas->raw
(2 перехода): FillBandRaw вместо Canvas.FillRect, DrawSliceFilterMarkers
разделён на raw-полосы + canvas-линии, композиты флагов и альфа-OR в
финальной raw-фазе. Кадр 8.3мс -> 6.1мс; spec_bands 1.87мс -> 0.03мс.
Альфа-OR проход на Qt6 подтверждён нужным (без него canvas-пиксели теряют
альфу и пропадают при блите), сам цикл ~0.08мс — оставлен; тумблер
EWSDR_ALPHA_FIX=0 только для экспериментов. Кисть перед AGC-подписями
теперь ставится явно (была остаточной после перестройки фаз).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Pluto exposes two temperature sensors: the AD9361 transceiver
(ad9361-phy temp0/input, milli-°C) and the Zynq SoC/FPGA (xadc temp0,
raw+offset+scale → C = (raw+offset)*scale/1000). Only the former was
read. Now both are read and shown as "Temp 37/53C" (chip/SoC) in the
desktop status bar and web UI.
- ReadTelemetry gains an out SocTempC param (base + Pluto override);
Pluto reads xadc via the IIO context.
- Controller stores FLastPlutoSocTempC; SoC temp validated with a
plausibility window (no slew-limit — stable local sensor, unlike the
AD9361 sensor that spikes during mid-conversion).
- Status formatting handles one or both temps in MainForm and WebAdapter.
- Widen desktop status-bar field 3 (150→190) so both temps + RSSI fit;
web layout unchanged.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Fix: dragging/wheeling the spectrum in the web UI clamped frequency to a
hardcoded HF ceiling (vfoMax=60 MHz), so on Pluto VHF/UHF it snapped down
to 60 MHz and couldn't pan up (only the VFO wheel worked, via a different
limit). Push the backend's tuning range (BackendCaps.MinFreqHz/MaxFreqHz)
to the web alongside the band plan; JS vfoMin/vfoMax use freq_min/freq_max
from state. The transverter branch (XVTR_CUR>=0) is unchanged — it keeps
the wide-open range since the displayed RF freq is translated to IF and
clamped server-side.
Dead-code cleanup (audit of this session's work):
- SampleRateMode/TSampleRateMode: orphaned when the sample-rate
unification replaced the srmContinuous branch in ApplyBackendCapsToUI;
no readers left. Removed (+ now-unused PLUTO_MAX_SR).
- Pre-existing write-only/unused: TBackendCaps.MaxSampleRate,
SampleRateOverlay.SPAN_NAMES, MainForm.BAND_FREQ.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
TX LO (altvoltage1) is now powered down on RX and enabled on PTT, instead
of staying on continuously. Removes the TX carrier leakage that showed up
at the VFO frequency in the RX spectrum (and persisted after MOX), and
stops idle TX-LO radiation. ApplyTxLO sets powerdown from FTransmitting,
which also recovers the LO if external software (e.g. SDR Console) left it
powered down. RX LO (altvoltage0) is kept on — receive is continuous; QO-100
full-duplex still works (RX LO separate).
Trade-off: re-enabling the TX LO re-runs the AD9361 synth calibration
(a few ms), so the first syllable on SSB may be slightly clipped.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Replaced linear interpolation upsampling (192k→device-rate) with a
polyphase FIR interpolator ×L. Prototype: windowed-sinc lowpass (Blackman,
~-58 dB sidelobes), TAPS*L long, cutoff TX_FIR_CUT=40k, normalized to unity
per-phase DC gain (loudness unchanged). Images at multiples of 192k are now
suppressed ~-58 dB instead of only sinc², cleaning up the uplink spectrum.
BuildTxResampler designs coefficients per FTxRatio in StartTX; FTxDlI/Q
delay line persists across SendDUCIQ calls, reset on StartTX/ClearDUCIQQueue.
Removed FTxPrevI/Q.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Fixed 4096-pair TX buffer was too small for full-duplex above ~1 Msps
("not enough ... samples getting dropped" per SoapyPlutoSDR), causing
clicks/dropouts. Now sized as round(rate/60) rounded up to a power of 2,
clamped [4096..65536] (576k→16384, 1536k→32768). FIFO cushion (~4 buffers)
resizes to match in StartTX.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- Status bar / web now show Pluto/LibreSDR model + serial instead of
"Unknown Board" (BoardDisplayName in controller; UI just renders it).
- Field 3 shows AD936x chip temperature + RX RSSI for Pluto (no PA →
Supply only for openHPSDR). Polled via backend.ReadTelemetry (~2 Hz),
filtered for sensor glitches (plausibility window + slew-limit, with
anti-stick) — fixes rare bogus -15°C readings.
- Add libiio binding iio_channel_attr_read_double.
- Remove stale root pluto_integration.MD (superseded by
doc/PLUTO_INTEGRATION_PLAN.md).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
- RX hw-gain (Pluto/AD936x): backend SetRxGain(ModeIdx, GainDb) — gain_control_mode
(manual/fast_attack/slow_attack/hybrid) + hardwaregain. Controller FRxGainMode/
FRxGainDb, commands SetRxGainMode/SetRxGain/RxGainBy, rfRxGain event, snapshot,
applied in StartRunning. Persisted per-device (rx_gain_mode/rx_gain_db). HPSDR
unaffected (backend no-op).
- UI: separate "RF AGC" panel below the (unchanged) WDSP AGC block, shown only for
Pluto — FAST/SLOW/HYB/MAN mode buttons + manual GAIN slider. LayoutLeftPanel
reflows lower panels via RelayoutBelowBands; no layout shift on HPSDR.
- Cold-init fix: direct sampling_frequency write doesn't load the AD9361 FIR
decimation filter, so a cold-booted Pluto showed a wide centre spike that didn't
track tuning (worked only after SDR Console pre-initialised it). Add optional
libad9361 binding (dynamic) and call ad9361_set_bb_rate() in ApplySampleRate,
which configures BBPLL + loads the FIR; falls back to direct write if absent.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add ADALM-Pluto / AD9361 support as a second hardware backend alongside
openHPSDR, sharing the WDSP DSP pipeline and the existing controller API
(UI stays decoupled from logic).
- RadioBackend.pas: abstract TRadioBackend + TBackendCaps + TRadioDevice
(Kind/URI/Serial). THPSDRNetwork now derives from it (state via virtual
getters); TRadioController.FNetwork is the base type.
- IIOBindings.pas: dynamic libiio loader (runs without libiio present).
- PlutoBackend.pas: scan/probe-by-URI, connect, LO/rate/bandwidth/gain
control, RX streaming thread (int16->24bit BE -> OnDDCIQ), Q conjugated
to match WDSP IQ convention. Verified on LibreSDR (AD9361) over network.
- Unified discovery: TDiscoverThread scans both backends; network Plutos
found via direct ProbeURI (no mDNS needed). ConnectDevice dispatches by
Dev.Kind (EnsureBackend swaps backend, preserving callbacks).
- DeviceStore/DeviceForm: persist Kind/URI/Serial; save Pluto via
AddSavedPluto so saved/autostart devices reconnect across restarts.
- VHF/UHF band plan (BoardUtils, kind-aware): 6m..ADS-B for Pluto; fixes
HF clamps (band detect, mouse-wheel 60 MHz cap, freq-display max).
- SampleRateOverlay: configurable presets (Pluto 576k..5760k, >520 ksps),
auto-width to fit.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
ew8bakandClaude Opus 4.8