Mihail Dumitrescu
87faa73bee
Smoother motion by fixing calculating trapezoids and ISR stepping. Fix rounding directions in calculate_trapezoid_for_block(). Fix off-by-ones errors in ac/deceleration steps in block_phase_isr. Half-initialize ac/deceleration_time to smooth the speed change shock that happens between segments, which is critical as jerk/deviation adds to this. The result is a smoother motion profile that follows the imposed acceleration limits with a well defined 0.5-1.5x error factor (or 2x if axis is starting from ~0). Errors are due to converting a real-valued motion profile into discrete numbers of steps. Fixes are general and improve S_CURVE_ACCELERATION too (no endorsement implied). Tested by looking at the generated step/dir impulses with a logic analyzer. Enjoy the smoother motion or use more aggresive acceleration/jerk/deviation values for faster prints. Also improves: #12491 Remove nominal_length, remove MINIMAL_STEP_RATE, add min_entry_speed_sqr, initial clean up of reverse_pass_kernel and forward_pass_kernel. Removing MINIMAL_STEP_RATE allows for correct handling of moves with low acceleration, including fixing judder that's caused when the planner computes an entry speed based on minimum_planner_speed_sqr that's then promptly overriden by MINIMAL_STEP_RATE. Added min_entry_speed_sqr to avoid a specific potential source of judder due to working with discrete steps rather continuous real-valued physics. The first step of any segment runs at initial_rate. If it is too low compared to acceleration_steps_per_s2 it will result in too much accumulated acceleration_time (see stepper.cpp) which will mean the following step will be at a much higher speed, and the speed change will significantly surpass the set acceleration_steps_per_s2 limit. Making sure we can match this limit is why we have minimum_planner_speed_sqr in the first place. Optimal number of sqrts and trapezoid calculations. Remove forward_pass(). Call forward_pass_kernel() from recalculate() instead. Fix potential for large speed changes if planner falls behind. group for clarity and review as described combined float sq match modified names Optimal number of sqrts and trapezoid calculations. Remove forward_pass(). Call forward_pass_kernel() from recalculate() instead. Fix potential for large speed changes if planner falls behind. Save and use block->steps_per_mm to avoid many divisions.
Marlin 3D Printer Firmware
Additional documentation can be found at the Marlin Home Page. Please test this firmware and let us know if it misbehaves in any way. Volunteers are standing by!
Marlin 2.1 Bugfix Branch
Not for production use. Use with caution!
Marlin 2.1 takes this popular RepRap firmware to the next level by adding support for much faster 32-bit and ARM-based boards while improving support for 8-bit AVR boards. Read about Marlin's decision to use a "Hardware Abstraction Layer" below.
This branch is for patches to the latest 2.1.x release version. Periodically this branch will form the basis for the next minor 2.1.x release.
Download earlier versions of Marlin on the Releases page.
Example Configurations
Before you can build Marlin for your machine you'll need a configuration for your specific hardware. Upon request, your vendor will be happy to provide you with the complete source code and configurations for your machine, but you'll need to get updated configuration files if you want to install a newer version of Marlin. Fortunately, Marlin users have contributed dozens of tested configurations to get you started. Visit the MarlinFirmware/Configurations repository to find the right configuration for your hardware.
Building Marlin 2.1
To build and upload Marlin you will use one of these tools:
- The free Visual Studio Code using the Auto Build Marlin extension.
- The free Arduino IDE : See Building Marlin with Arduino
- You can also use VSCode with devcontainer : See Installing Marlin (VSCode devcontainer).
Marlin is optimized to build with the PlatformIO IDE extension for Visual Studio Code. You can still build Marlin with Arduino IDE, and we hope to improve the Arduino build experience, but at this time PlatformIO is the better choice.
8-Bit AVR Boards
We intend to continue supporting 8-bit AVR boards in perpetuity, maintaining a single codebase that can apply to all machines. We want casual hobbyists and tinkerers and owners of older machines to benefit from the community's innovations just as much as those with fancier machines. Plus, those old AVR-based machines are often the best for your testing and feedback!
Hardware Abstraction Layer (HAL)
Marlin includes an abstraction layer to provide a common API for all the platforms it targets. This allows Marlin code to address the details of motion and user interface tasks at the lowest and highest levels with no system overhead, tying all events directly to the hardware clock.
Every new HAL opens up a world of hardware. At this time we need HALs for RP2040 and the Duet3D family of boards. A HAL that wraps an RTOS is an interesting concept that could be explored. Did you know that Marlin includes a Simulator that can run on Windows, macOS, and Linux? Join the Discord to help move these sub-projects forward!
Supported Platforms
| Platform | MCU | Example Boards |
|---|---|---|
| Arduino AVR | ATmega | RAMPS, Melzi, RAMBo |
| Teensy++ 2.0 | AT90USB1286 | Printrboard |
| Arduino Due | SAM3X8E | RAMPS-FD, RADDS, RAMPS4DUE |
| ESP32 | ESP32 | FYSETC E4, E4d@BOX, MRR |
| LPC1768 | ARM® Cortex-M3 | MKS SBASE, Re-ARM, Selena Compact |
| LPC1769 | ARM® Cortex-M3 | Smoothieboard, Azteeg X5 mini, TH3D EZBoard |
| STM32F103 | ARM® Cortex-M3 | Malyan M200, GTM32 Pro, MKS Robin, BTT SKR Mini |
| STM32F401 | ARM® Cortex-M4 | ARMED, Rumba32, SKR Pro, Lerdge, FYSETC S6, Artillery Ruby |
| STM32F7x6 | ARM® Cortex-M7 | The Borg, RemRam V1 |
| STM32G0B1RET6 | ARM® Cortex-M0+ | BigTreeTech SKR mini E3 V3.0 |
| STM32H743xIT6 | ARM® Cortex-M7 | BigTreeTech SKR V3.0, SKR EZ V3.0, SKR SE BX V2.0/V3.0 |
| SAMD51P20A | ARM® Cortex-M4 | Adafruit Grand Central M4 |
| Teensy 3.5 | ARM® Cortex-M4 | |
| Teensy 3.6 | ARM® Cortex-M4 | |
| Teensy 4.0 | ARM® Cortex-M7 | |
| Teensy 4.1 | ARM® Cortex-M7 | |
| Linux Native | x86/ARM/etc. | Raspberry Pi |
Marlin Support
The Issue Queue is reserved for Bug Reports and Feature Requests. Please use the following resources for help with configuration and troubleshooting:
- Marlin Documentation - Official Marlin documentation
- Marlin Discord - Discuss issues with Marlin users and developers
- Facebook Group "Marlin Firmware"
- RepRap.org Marlin Forum
- Facebook Group "Marlin Firmware for 3D Printers"
- Marlin Configuration on YouTube
Contributing Patches
You can contribute patches by submitting a Pull Request to the (bugfix-2.1.x) branch.
- We use branches named with a "bugfix" or "dev" prefix to fix bugs and integrate new features.
- Follow the Coding Standards to gain points with the maintainers.
- Please submit Feature Requests and Bug Reports to the Issue Queue. See above for user support.
- Whenever you add new features, be sure to add one or more build tests to
buildroot/tests. Any tests added to a PR will be run within that PR on GitHub servers as soon as they are pushed. To minimize iteration be sure to run your new tests locally, if possible.- Local build tests:
- All:
make tests-config-all-local - Single:
make tests-config-single-local TEST_TARGET=...
- All:
- Local build tests in Docker:
- All:
make tests-config-all-local-docker - Single:
make tests-config-all-local-docker TEST_TARGET=...
- All:
- To run all unit test suites:
- Using PIO:
platformio run -t test-marlin - Using Make:
make unit-test-all-local - Using Docker + make:
maker unit-test-all-local-docker
- Using PIO:
- To run a single unit test suite:
- Using PIO:
platformio run -t marlin_<test-suite-name> - Using make:
make unit-test-single-local TEST_TARGET=<test-suite-name> - Using Docker + make:
maker unit-test-single-local-docker TEST_TARGET=<test-suite-name>
- Using PIO:
- Local build tests:
- If your feature can be unit tested, add one or more unit tests. For more information see our documentation on Unit Tests.
Contributors
Marlin is constantly improving thanks to a huge number of contributors from all over the world bringing their specialties and talents. Huge thanks are due to all the contributors who regularly patch up bugs, help direct traffic, and basically keep Marlin from falling apart. Marlin's continued existence would not be possible without them.
Project Leadership
| Name | Role | Link | Donate |
|---|---|---|---|
| 🇺🇸 Scott Lahteine | Project Lead | [@thinkyhead] | 💸 Donate |
| 🇺🇸 Roxanne Neufeld | Admin | [@Roxy-3D] | |
| 🇺🇸 Keith Bennett | Admin | [@thisiskeithb] | 💸 Donate |
| 🇺🇸 Jason Smith | Admin | [@sjasonsmith] | |
| 🇧🇷 Victor Oliveira | Admin | [@rhapsodyv] | |
| 🇬🇧 Chris Pepper | Admin | [@p3p] | |
| 🇳🇿 Peter Ellens | Admin | [@ellensp] | 💸 Donate |
| 🇺🇸 Bob Kuhn | Admin | [@Bob-the-Kuhn] | |
| 🇳🇱 Erik van der Zalm | Founder | [@ErikZalm] |
License
Marlin is published under the GPL license because we believe in open development. The GPL comes with both rights and obligations. Whether you use Marlin firmware as the driver for your open or closed-source product, you must keep Marlin open, and you must provide your compatible Marlin source code to end users upon request. The most straightforward way to comply with the Marlin license is to make a fork of Marlin on Github, perform your modifications, and direct users to your modified fork.