Version 3.7.0-rc.1

Compatible versions:

• DuetSoftwareFramework 3.7.0-rc.1
• RepRapFirmware 3.7.0-rc.1

New features:

• Added a Motor Analysis tab to the input shaping plugin that records constant-speed moves per motor and analyzes vibrations at multiples of the full-step frequency, with per-run summaries compared at equal absolute frequencies, findings and a speed sweep view
• Added a Record Motor Profile wizard that derives motor-isolating moves from the core kinematics matrices and records them once per configured speed
• Added a Tune Motor dialog that minimizes the vibration harmonics of a motor, via M970.3 in phase stepping mode or the driver's sine table (M569.2) in step/dir mode, for local and CAN-connected drivers alike
• Added a chart and table summary view with configurable unit and number of harmonics to the motor analysis tab
• Merged the Current Settings tab and the input shapers panel into the Input Shaping tab, which shows the current shaper response while no profile is selected
• Added Duplicate to the context menu of the filaments directory, which copies the files of a filament to a new one
• New Directory creates a filament with empty load.g, config.g and unload.g when it is used in the filaments directory

Plugins and architecture:

• Moved the shared accelerometer helpers of the input shaping plugin into a composable
• Updated @duet3d/motionanalysis to 3.7.0-rc.1 so the estimated shaper effect accounts for damping, and the remaining Duet3D packages to their 3.7.0-rc versions

Changed behaviour:

• The motor waveform tuning tab is only shown when the driver supports it
• Motor analysis results are kept in memory, so switching views does not recompute them
• Motor analysis findings are graded by absolute displacement instead of harmonic ratios
• Motor profile speeds are limited to the frequencies the accelerometer can resolve, and its sampling rate is taken from the object model instead of querying M955
• Added 400 Hz to the default motor recording speeds and limited the selectable number of harmonics to 8, higher orders carry no actionable information
• Recording dialogs size collections from the accelerometer's actual sampling rate and skip the settle dwell in SBC mode
• Motion analysis charts render the frequency and damping a profile was recorded with while the input fields mirror the object model, and estimating the shaper effect is disabled for recordings made with input shaping enabled
• Chart scales are kept while comparing recordings
• M292 no longer waits for G-code replies
• The first HTTP retry after a 503 response waits like the following ones, and the default retry delay is 100 ms instead of 200 ms
• Renaming or deleting a loaded filament asks for confirmation because its unload macro is no longer found afterwards

Bug fixes:

• The G-Code Viewer gauges kept scheduling d3 transitions that never completed, piling up until the browser tab locked up
• The G-Code Viewer framed the bed while a job was heating up and did not zoom in when a job started
• The wrong heater was shown for the bed on RRF 3.6, the G-Code Viewer overlay showed no bed and chamber gauges on firmware without bedHeaterMapping, and it labelled beds and chambers by padded mapping slots
• A cancelled job re-uploaded under the same name and started again kept showing the previous version in the G-Code Stream and the G-Code Viewer
• The firmware install file picker returned an empty file list in Chromium-based browsers, so the upload never started (PR #518)
• Shaper parameter fields were oversized
• Sub-order motor findings were graded at frequencies without a full-step reference, turning low-frequency noise into misleading displacements
• The tune result hint mentioned M970 in step/dir mode
• Motor profile speeds were limited to a fraction of what the accelerometer can analyze
• Apply chips of the input shapers were never shown
• The Home All button in the recording dialogs did not indicate progress and was misaligned
• Recording dialogs could be dismissed while collecting
• Freshly written accelerometer files could be downloaded without their trailer in SBC mode
• WiFi server and PanelDue binaries as well as the DSF and DWC update bundles were only recognised under their exact file names, so a download renamed by the browser like DuetWiFiServer (1).bin did not start an update
• DuetScreen.bin was offered as a PanelDue firmware update although a DuetScreen is updated from its own SD card
• Various layout and truncation issues on the analysis tabs