LaserBoard is an open source controller for fiber laser marking systems, based around a Teensy 4.1 MCU board. LaserBoard provides a USB interface compatible with most low-cost fiber laser sources and XY-100-speaking galvanometer scanners, and supports 4 axes of step/direction motion control. LaserBoard is released under the terms of CERN-OHL-S; see LICENSE.txt, or cern_ohl_s_v2.txt for the full text.
At the component level, laser marking hardware is remarkably cheap and high quality, and easy to interface with. Sadly, laser marker control boards aren't! Generally, they're bundled with a license for low-quality, proprietary software, greatly inflating the price - a control board with 4 auxiliary axes (eg. JCZ DLC2-M4) cost about as much as a quality laser source and scanner ($2000!).
Advanced features such as high-precision lens calibration, field stitching, active z-axes control etc. are also very difficult to implement, as the commercial controllers behave like a black-box and speak an undocumented host/controller protocol. Some reverse-engineered 3rd party software exists, such as LightBurn (commercial, user-friendly high-quality GUI) or Balor (open-source python interface library), but controller support is limited.
Lasers can be very dangerous! Unenclosed 10+ watt invisible lasers with peak powers in the low 10s of kW are absolutely not safe, and a quick scan of YouTokstagram "laser influencer" content will show many poorly educated people doing very dangerous things with them. Laser safety regulation and enforcement hasn't caught up to the ease with which anyone can import a turnkey class-4 laser capable of instablindess.
Lasers are also very useful and all of the risks associated with them may be mitigated with proper engineering and procedural controls. It is essential to fully understand these risks BEFORE purchasing a laser, and if you already have one, before modifying it or further using it.
If you're in a context with a laser safety officer, talk with them. Laser safety officers are your friends, even if they seem scary at first.
LaserBoard is designed for use with pewpew-laser. If that link appears to be a 404, it's because I'm ambivalent about running an open-source project and haven't figured out the liability issues related to releasing a complete system. Additionally, there's no GUI or even a standard toolpath generation script - although there might be some Rhinoceros plugins in the works.
The following table provides an estimate for a minimal laser marking system with a small Q-switched laser source, a 50x50 working envelope, a mid-market galvo scanner, and a manually-focused z axis.
| Component | Example Manufacturer and PN | Budgetary Price |
|---|---|---|
| Control Board PCB | LaserBoard-Rev5 | $5-50 |
| Control Board Components | Mouser | $90 |
| 30W Laser Source | Raycus - RFL-P30QS | $1000 |
| Galvo Scanner | Sino-Galvo SG7110 | $400 |
| F-theta Lens | HY-SL-1064-50-63 | $70 |
| Beam Path | "Fiber Laser Beam Path" | $70 |
| Stand | "Laser Fiber Lift Table" | $120 |
| 350W 24V Power Supply | Meanwell LRS-350-24 | $35 |
| 2x 15V 2A Power Supplies | CUI VGS-35C-15 | $28 |
Coincidentally, this comes to around $1850 - around $900 cheaper than a plug-and-play commercial system with roughly the same grade of component. There does seem to be a large amount of price flexibility on laser sources from certain AliBaba vendors, to the extent that some JPT MOPA sources may be available for around 50% of list price. However, importing laser sources can pose significant customs issues. Note that this does not include key safety equipment, such as an enclosure, properly rated laser safety glasses, or an adequate ventilation system; not does it include wires, terminal blocks, etc.
Revision 5a is a minor revision, primarily fixing screwups in Revision 5. The major changes are as follows:
-
Pulldown resistors were added to laser output pins, as turn-on behaviour was undefined.
-
Board footprint expanded to keep both DB connector faces in-plane and space was added between the connectors to accommodate cable's overmolded plastic.
-
C15 was switched from an SMD part to a PTH part to simplify soldering.
-
The pull-up/pull-down selection jumpers for inputs actually work with the debouncing network and don't require bodge wires.
-
The ethernet status LED is routed to the correct pins on the magjack.
Revision 5 is, in some ways, a simplification, and breaks pin mapping backwards compatibility. The major changes are as follows:
-
The MCU board is changed from a Teensy 4.0 to a Teensy 4.1, in order to access more GPIO and remove a messy SMD header on the bottom of the board.
-
Proper power supply input - small fiber lasers generally run on 24VDC supplies, and galvo scanners required -15/15V supplies, so there's always a readily available DC rail to generate a 5V rail for the teensy. The input protection circuit and switcher is essentially the same as published here, although it omits reverse polarity protection and uses a small polyfuse due to much lower current demands. The board may also be powered off a USB connection by shifting the location of a jumper.
-
The Q-switch protection circuitry is removed, as the current firmware seems to be robust enough to trust with not generating Q-switch clock glitches, and the broader frequency range required for MOPA sources renders it useless.
-
Previously, the standard 8-bit parallel laser power interface used a shift register to reduce the GPIO requirement - now, 8 GPIO pins directly control the power. Supporting MOPA serial data stream with the shift register would have required bit-banging a serial stream to bit bang a serial stream - although this was technically feasible, the code complexity would have been annoying, and the switch to a parallel interface only required four extra pins.
-
Four limit pins are properly debounced with a low-pass filter and Schmitt-trigger and have (jumper) configurable pull-up/down resistors and indicator LEDs.
-
The teensy 4.1 ethernet phy is broken out to an on-board magjack.
Rev 4 fixed a bodge wire needed for rev 3 and slightly modified the laser pinout to support JPT long-pulse sources, although that is as of yet untested.
Rev 2 incorporated the following changes:
- Support for a two-axis galvanometer scanner communicating over XY-100, requiring adding a DB-15 interface and four differential line drivers.
- Motor interfaces were moved to 4 polarized 4 pin molex 0,1" headers.
Rev 3 is still in active use.
Revision one was ordered 2021/5/4, and had the following features:
- One DB-25 interface for a q-switched fiber laser, with a one-shot timer used to protect the laser from abnormal q-switch pulse trains, a shift register for the power control interface, and a hardware interlock to require a good q-switch clock signal, a good e-stop signal, and software enabling of the laser.
- Four step/direction interfaces, level-shifted to 5V and broken out on a single 0.1" header.
- A schmitt-triggered and debounced e-stop input.
- 9 GPIO broken out on a 0.1" header, with no level-shifting or signal conditioning.
Revision one was used with a stepper-driven micrometer stage to get a proof-of-concept working for some surface-patterning applications, and was quickly replaced with rev 2 - seemingly a mostly cosmetic modification .