Fix extrusion of some support layers at wrong Z height
Fixes the issue that PR #12736 (github.com/OrcaSlicer/OrcaSlicer/pull/12736) reverted the changes of PR #13327 (github.com/OrcaSlicer/OrcaSlicer/pull/13327)
This PR supersedes #12225, which originally proposed this feature but
appears inactive.
The feature originated from work I developed earlier in
[BambuStudio-ZAA](https://github.com/adob/BambuStudio-ZAA), a private
fork of Bambu Studio
Compared to #12225, I updated the implementation for current upstream
and fixed the following issues:
- fixed broken tests
- removed references to nonplanar directory
Reviewers may want to compare against #12225 for earlier
discussion/context.
## Summary
Port of **Z Anti-Aliasing (ZAA)** from
[BambuStudio-ZAA](https://github.com/adob/BambuStudio-ZAA) to
OrcaSlicer.
ZAA eliminates visible stair-stepping on curved and sloped top surfaces
by raycasting each extrusion point against the original 3D mesh and
micro-adjusting its Z height to follow the actual surface geometry. The
result is visibly smoother domes, chamfers, and shallow slopes — without
post-processing.
## How It Works
1. The slicer runs normally, then a **posContouring** step processes
each layer
2. `ContourZ.cpp` raycasts every extrusion point vertically against the
source mesh
3. Each point's Z is adjusted to the mesh intersection, converting flat
`Polyline` paths into `Polyline3` paths with per-point Z coordinates
4. The G-code writer emits the adjusted Z values, so the printer follows
the true surface
## Configuration
Five new settings under **Print Settings > Quality**:
| Setting | Type | Default | Description |
|---------|------|---------|-------------|
| `zaa_enabled` | bool | off | Master enable/disable switch |
| `zaa_min_z` | float | 0.06 mm | Minimum Z layer height; controls
slicing plane offset |
| `zaa_minimize_perimeter_height` | float | 35° | Reduce perimeter
heights on slopes below this angle (0 = disabled) |
| `zaa_dont_alternate_fill_direction` | bool | off | Keep fill direction
consistent instead of alternating |
| `zaa_region_disable` | bool | off | Disable ZAA for a specific print
region/material |
## Key Changes
- **Core algorithm**: New `src/libslic3r/ContourZ.cpp` (~330 lines) —
raycasting engine
- **3D geometry**: `Point3`, `Line3`, `Polyline3`, `MultiPoint3` extend
existing 2D types
- **Arc fitting**: Templated to work with both 2D and 3D geometry
- **Pipeline**: `ExtrusionPath::polyline` changed from `Polyline` to
`Polyline3`; new `posContouring` step in `PrintObject.cpp`
- **G-code**: `GCode.cpp` writes per-point Z when `path.z_contoured` is
set
- **UI**: ZAA settings exposed in Print Settings > Quality panel
- **Documentation**: `docs/ZAA.md` with usage and implementation details
57 files changed, ~1800 insertions, ~200 deletions.
## Test Plan
- [ ] Load a model with curved top surfaces (sphere, dome, chamfered
box)
- [ ] Enable **Z contouring** in Print Settings > Quality
- [ ] Slice and verify G-code has varying Z values within contoured
layers
- [ ] Build on macOS (verified), test on Linux and Windows
* Add extrusion role change G-code options
Introduces new G-code options for handling extrusion role changes at the process and filament levels. Updates configuration, GUI, and GCode logic to support 'process_change_extrusion_role_gcode' and 'filament_change_extrusion_role_gcode', allowing custom G-code insertion when the extrusion role changes.
Co-Authored-By: Rodrigo Faselli <162915171+RF47@users.noreply.github.com>
* Optimize extrusion role-change gcode handling
Cache gcode template strings and current filament id, and guard creation of DynamicConfig/placeholder processing behind a check that at least one role-change gcode is non-empty. This avoids redundant config lookups and DynamicConfig/placeholder parsing when no custom role-change gcode is defined, improving clarity and performance without changing behavior.
---------
Co-authored-by: Rodrigo Faselli <162915171+RF47@users.noreply.github.com>
In some rare support edge cases, Orca could start extruding a new layer before moving to the correct Z height.
This happened when two support layers were generated back-to-back and the next layer started exactly where the previous one ended. In that situation, there was no movement that naturally updated the Z position first.
The code was clearing the “pending layer change” flag too early, so it lost track of the fact that a Z move was still required.
This change ensures that if a layer change is still pending, Orca will always move to the correct Z height before the first extrusion of that layer.
This guarantees that every layer starts at the correct height and fixes the missing / incorrect support layers seen in those edge cases.
Adjust initial- and slowdown-layer speed logic to correctly account for raft layers. Replace the previous layer checks with object_layer_over_raft(), simplify initial-layer speed selection, and split the slow_down_layers interpolation into separate branches for configurations with and without rafts so the lerp uses the correct layer offset. Also avoid applying overhang speed adjustments to object layers that are over rafts.
# Description
This commit adds the enhancement of https://github.com/OrcaSlicer/OrcaSlicer/issues/13116
- Adds checkboxes to enable or disable the fan speed override during print or after printing is completed. This allows users to, for example, only override the fan speed during printing whilst leaving the fan speed as is set in the machine gcode at the end.
- Two new flags are added for this: `activate_air_filtration_during_print[extruder]` and `activate_air_filtration_on_completion[extruder]`. These can be used to more finely control machine gcode
- Current filament settings remain as they are: `activate_air_filtration[extruder]` settings are unaffected and by default both new flags are set to true, ensuring the same set behavior as before.
# Screenshots/Recordings/Graphs
<img width="856" height="208" alt="Screenshot From 2026-04-09 18-57-14" src="https://github.com/user-attachments/assets/e71e7de3-2def-4046-b5dc-55bf3b516ce5" />
As you can see there are now checkboxes left of the fan speeds. They have their own tooltip too, which also helps identify the correct flags for users who want to adjust their machine gcode.
## Tests
I have thoroughly tested this on my own computer (CachyOS) and on my printer. I have also carefully checked the gcode in every possible state this is in and ensured that the unsaved changes dialog properly displays the labels for these settings. Flags are set properly, sliced files properly use the flags if you check for them in machine gcode and default behavior is unaffected for those who have already changed settings for air filtration. From what I can see, this does exactly what it should be doing without any issues.
* make initial layer travel move acceleration and jerk configurable
* Update spelling to match UI pattern
* Update min integer and re-order to match patterns
---------
Co-authored-by: Thomas Scheiblauer <tom@sharkbay.at>
Co-authored-by: Ioannis Giannakas <59056762+igiannakas@users.noreply.github.com>
* Fix crash-inconsistent draft shield
* Skip empty skirt G-code and fix skirt_done handling
Track whether a skirt was actually generated for the current print Z and avoid appending empty skirt G-code. Only pop m_skirt_done when a skirt was emitted to prevent corrupting skirt state across multiple instances per layer. Also use a temporary string (moved into the main gcode) to detect emptiness before updating state. Preserves existing first-layer clearing behavior.
* Union ex brims
Revert "Union ex brims"
This reverts commit bbc9a39faf318dc2df093eb2bdcebf19a4162fe9.
Update Brim.cpp
* dont repeat paths
* Update Brim.cpp
* multimaterial brim independiente
* Normal brim if is by object
* fix print order
* cleaning 1
* cleaning 2
* Normal brim if multimaterial on first layer
* fix artifact
* combine_brim optional
refactoring
* refactoring gcode.cpp
* refactoring brim.cpp
Update Brim.cpp
* Remove multimaterial first-layer check for brims
Stop detecting extruders used on the first layer and remove the is_multimaterial_first_layer guard. Simplify can_combine_brims to only consider combine_brims and whether printing is ByObject, allowing brims to be combined across extruders unless printing by object or combine_brims is disabled. Cleans up unused code and simplifies brim-generation conditions.
* Remove material specification from unified brim comment
* Update PrintConfig.cpp
- Handle z contouring in variable speed flow when emitting GCode
- Add logic to restore nominnal z height for regular extrusions
- preserve z_contoured flag when splitting extrusion paths
Previously, wipe tower behavior was determined by checking if the printer
was a QIDI vendor. This introduces a configurable enum (Type 1 / Type 2)
so any printer can select its wipe tower implementation. BBL printers
remain hardcoded to Type 1. Qidi profiles default to Type 1.
When flush_multiplier is applied to flush_volumes_matrix values, the
result can be a decimal (e.g., 272 * 1.3 = 353.6). Some printer firmware
metadata parsers (notably Creality K2) crash when parsing decimal values
in flush_volumes_matrix, causing prints to get stuck at "File selected".
This fix rounds the multiplied values to integers, ensuring compatibility
with firmware that expects integer values in this field.
Fixes compatibility with Creality K2 and potentially other Klipper-based
printers that parse the CONFIG_BLOCK metadata.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-authored-by: Claude <noreply@anthropic.com>
The WipeTower2 code path never called construct_mesh(), leaving
wipe_tower_mesh_data as std::nullopt. GCode export then dereferenced
it unconditionally, triggering an assertion failure / crash.
Added checks in ByLayer print sequence to handle ordering for Intra-layer ordering option. Prevents new ordering setup when "As object list" is enabled.
print_machine_envelope() used get_extruder_id(extruder_id)*2 to index
machine limit arrays that only hold [Normal, Stealth] (2 entries).
For multi-extruder setups this went out-of-bounds, causing wrong M201/M203
values in the G-code which then override the estimator's correct limits.
Same class of bug as c6d1c11ebb but on the G-code writer side.
Changes:
- Remove unused extruder_id param from print_machine_envelope()
- Use .values.front() for M201/M203, matching M204/M205 in same function
- Change get_option_value() fallback from .back() to .front() so any
future out-of-bounds index returns Normal mode instead of Stealth
Wipe tower interface features and preheat fixes
Fresh PR branch rebuilt on upstream/main (squash of origin/BBL-studio-wipe-tower-merge) to avoid merge-history issues.
* fixes: keys_map is initialized with itself [-Winit-self]
* fixes: operation on repeats may be undefined [-Wsequence-point]
* fixes: warning: suggest parentheses around && within || [-Wparentheses]
* fixes: moving brim_points to itself [-Wself-move]
* fixes: operation on unprintability may be undefined [-Wsequence-point]
* fixes: extra tokens at end of #endif directive [-Wendif-labels]
* fixes: converting to non-pointer type int from NULL [-Wconversion-null]
* review result: simplifies fix 'operation on repeats may be undefined'
Port Z Anti-Aliasing from BambuStudio-ZAA (https://github.com/adob/BambuStudio-ZAA)
to OrcaSlicer. ZAA eliminates stair-stepping on curved and sloped top surfaces
by raycasting each extrusion point against the original 3D mesh and micro-adjusting
Z height to follow the actual surface geometry.
Key changes:
- Add ContourZ.cpp raycasting algorithm (~330 lines)
- Extend geometry with 3D support (Point3, Line3, Polyline3, MultiPoint3)
- Template arc fitting for 2D/3D compatibility
- Change ExtrusionPath::polyline from Polyline to Polyline3
- Add 5 ZAA config options (zaa_enabled, zaa_min_z, etc.)
- Add posContouring pipeline step in PrintObject
- Update GCode writer for 3D coordinate output
- Add ZAA settings UI in Print Settings > Quality
- Add docs/ZAA.md with usage and implementation details
ZAA is opt-in and disabled by default. When disabled, the slicing pipeline
is unchanged.
## Problem
Some Creality printers (e.g. Ender-3 V3 SE) rely on metadata comments in the first lines of G-code files to display print info on the screen (estimated time, filament usage, layer height). This follows a format originally used by Cura/Creality Print:
```
;FLAVOR:Marlin
;TIME:3708.97
;Filament used:6.21m
;Layer height:0.2
```
OrcaSlicer currently has no way to write content before the `HEADER_BLOCK` in the G-code output, and does not expose `print_time_sec` or `filament_length_m` as usable placeholders. As a result, these printers show zeroed-out print info on screen.
## Changes
### 1. New placeholders (post-processing)
Added two new G-code placeholders resolved during post-processing:
- `{print_time_sec}` — total estimated print time in seconds (double)
- `{filament_length_m}` — total filament length in meters (double)
These values are only available after G-code generation, so they use inline marker replacement (`@PRINT_TIME_SEC@`, `@FILAMENT_LENGTH_M@`) that `GCodeProcessor::run_post_process()` substitutes with actual computed values — the same pattern used by existing M73 and layer count placeholders.
### 2. New "File header G-code" option (`machine_top_gcode`)
Added a new G-code field in **Printer Settings > Machine G-code** that writes content at the very top of the output file, before `HEADER_BLOCK_START`. This allows users to add firmware-specific metadata that must appear in the first lines.
### 3. Creality Ender-3 V3 SE profile update
Pre-configured the Ender-3 V3 SE profiles (all 4 nozzle variants) with a default `machine_top_gcode` value matching the Cura-compatible header format, so print info displays correctly out of the box.
# Screenshots
| Before (current OrcaSlicer) | After (this PR) |
|---|---|
|  |  |
## Tests
- Sliced test model with Ender-3 V3 SE profile
- Verified G-code output contains correct values in first lines:
- `;TIME:` with actual print time in seconds
- `;Filament used:` with filament length in meters
- `;Layer height:` with correct layer height
- Verified `{print_time_sec}` and `{filament_length_m}` work in both Machine Start G-code and File header G-code fields
- Verified empty `machine_top_gcode` produces no extra output
- Built and tested on macOS (arm64)
* Emit Disable Power Loss Recovery
Now only works if it's enabled but the goal it's to force disable it.
With this change it will always emit the command for BBL or Marlin 2.
Co-Authored-By: Michael Rook <54159303+michaelr0@users.noreply.github.com>
* Refactor power loss recovery G-code comments
* Return empty power loss recovery when no compatible printer
* Update power loss recovery comments
Update label and tooltip for power loss recovery
* Add enum for power loss recovery mode
Refactored power loss recovery configuration to use a new PowerLossRecoveryMode enum instead of a boolean. Updated GCodeWriter and related logic to handle the new enum, allowing for 'printer_configuration', 'enable', and 'disable' options. Updated config handling, legacy value conversion, and default values accordingly.
* Update PrintConfig.cpp
---------
Co-authored-by: Michael Rook <54159303+michaelr0@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
* Adjust resonance avoidance speed calculation
* Only adjust if it's strictly below max
* If the speed is over half of the set speed range,
use the max speed to prevent slowdowns
Signed-off-by: Kuran Kaname <celtare21@gmail.com>
* update comment
---------
Signed-off-by: Kuran Kaname <celtare21@gmail.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
* Add ability to disable Power Loss Recovery
* Fix typo in PrintConfig.hpp for power loss recovery
* Attempt to resolve Unknown option exception: disable_power_less_recovery
Add disable_power_loss_recovery property to any json which had scan_first_layer
* Revert "Attempt to resolve Unknown option exception: disable_power_less_recovery"
This reverts commit ddaf34b317f8797540235b8f7d179b1c9b6d10f8.
* Fix typo
* Change attribution from BBS to Orca in PrintConfig.cpp
* Mini refactor power loss recovery handling in GCode export
- Moved power loss recovery G-code generation to a new method in GCodeWriter.
- Support Marlin 2
* Update comments and power loss recovery handling
* Implement power loss recovery G-code commands
Added functions to start and end power loss recovery with appropriate G-code commands and comments.
* Add power loss recovery methods to GCodeWriter
* refactor and fix build errors
---------
Co-authored-by: Michael Rook <michael@rook.id.au>
Co-authored-by: Ioannis Giannakas <59056762+igiannakas@users.noreply.github.com>
Co-authored-by: SoftFever <softfeverever@gmail.com>
