AxCut: An Open Source Laser Cutter

If you’ve always wanted a laser cutter and you have £1500 lying around (approx. $2500 as of today) — and you have access to a 3D printer — then you’ll want to take a look at [Damian’s] open source laser cutter: axCut. The project has evolved over the last few months from some mockups in OpenSCAD to a working prototype.

You’ll want to dig through his blog posts as well as his YouTube channel for all the juicy details, but from what we can gather, [Damian] is on the home stretch. The current implementation includes a 40W CO2 laser with functioning laser control and an impressively quiet watercooling system. Although the build’s wiring remains a bit of a tangle, the prototype cuts (almost) as expected. His next hurdle is ironing out the air assist, which should prevent some fire hazard issues and keep the lens free of debris.

Check out a couple of videos after the break, and if you’re interested in getting into laser cutting but want to start smaller, have a look at the MicroSlice from a few months ago.

47 thoughts on “AxCut: An Open Source Laser Cutter

      1. I have two 100w lasersaurs – I can absolutely assure you that it’s 100% opensourced. Every single line of code is in the GitHub and the schematic for the single circuit board is available. There is an active ‘community branch’ of the software that I’ve been modifying and improving – which is feeding into the ‘core branch’ that’s what most people download. Obviously some of the components it uses (stepper drivers, laser PSU, laser tube, etc) are off-the-shelf commercial – but that’s going to be true for pretty much every project of this complexity. It’s computers are a beaglebone black and an arduino-compatible Atmel chip…so you’re running OpenSourced Linux on one machine and bare-metal C++ code on the atmel CPU.

        The file formats it works best with are open standards (G-code at the low level and SVG at the high level) – and you can make stuff for it using (for example) Inkscape – which is OpenSourced.

        I don’t know why the site says “97%” – that’s clearly not the case.

        The Lasersaur is immensely powerful and capable – 100 watts over a 2’x4′ bed is impressive…but that comes at a price – the bill of materials runs to over $7,500. Using a smaller bed and a lower powered laser drops the cost dramatically – so it doesn’t surprise me that the Axcut is as cheap as it is.

        We have over 4000 hours on our first lasersaur (“The Death Ray of Ming the Merciless”) and close to 1000 hours on the second (“Illudium Q36”). There are few (if any) hobby laser cutters that have been thrashed to the degree that ours have. We use our Lasersaurs to run our business ( – and they have paid for themselves at least ten times over. Our business is Kickstarter funded and uses *only* OpenSource software and OpenHardware for anything beyond a basic PC/laptop.

      1. If (you had not been confusing things) { the lasersuar should not claim eventual 100% open-source-ness; } /* they would have known at day one that it was a lie */

        I think open-source (and that alone) has become some kind of quality-mark, quite undeservingly. there is plenty of useless open-source code out there. And I do personally prefer open-source, so I am not a hater putting down the many other great open-source products. Excatly like there is a ton of horrible closed-source options among the many good ones. My point is that the type of source is only one datapoint to consider. And we all have different priorities (and we sometimes have different priorities on different projects).

        1. I agree with you.

          Years ago when I first started getting into web development, I would download and install “Open Source” web apps that were free and supposed to be good. There was also the standard licensing dictate that required one to submit any changes to the program that was made. I never submitted my changes due to the fact that I found myself spending hours debugging them and in most cases, actually rewriting the procedures in order to get the program to run.

          I got the distinct impression that the authors were just handing out junk functions in order to get someone else to do the hard work of debugging to get it to work and then lay claim to the working code.

          That in my mind is steeling.

          There were and likely still are, way to many people claiming an ability to do without the ability to do.

        2. The lasersaur is absolutely 100% opensourced software. I own a couple of these machines – and every single line of code is out there in the public GitHub. I modify it and share the results on a ‘community branch’ GitHub where there are half a dozen other contributors.

          The hardware designs are also 100% open – there are schematics and board layouts for the electronics, CAD diagrams for all of the assemblies. The only parts of the machine you can’t buy from off-the-shelf manufacturers like Misumi and Mouser are some of the aluminium brackets. However, there are CAD diagrams for those too – so you could get them made on a water jet cutter or a mini-CNC mill if you really wanted to. My first machine uses those exact same designs, cut from 1/4″ acrylic. They work just fine – but tend to crack over time and need replacement – so aluminium parts work better. Of course I have a laser cutter – so I can make spares myself.

          So, I don’t know why people would think that Lasersaur isn’t OpenSourced – it most certainly is…and I know that for 100% sure because I own two of them.

      1. Yes – but it’s a 100 watt laser with a 2’x4′ bed…and that pushes the price up considerably. It also assumes you’re going to buy all of the parts from places like Mitsumi and Mouser. If you shop around, you can do much better.

        The lasersaur is for very serious users – it’s overkill for most hobbyist users.

        1. The Lasersaur isn´t cheap, but it is big.
          It is powered by a Beagle Bone Black, which is very capable hardware for the task.

          Unfortunately, the two Programmable Realtime Units that fit the board are sitting idle, because motor control is implemented by a custom GRBL firmware running on a Arduino board (connected to the BBB over a serial link)… that is enough for a 3D printer moving at 1-2 inches per second and a build volume of few cubic inches, but not for a build like the Lasersaur !

          Controlling the motors/fan/laser directly on the BBB, with a proper RT kernel and nice PRU assembler routines would yield a dramatic impact on performances:

          The hardware build in itself is good, and gets better with every version. The google group contains much valuable information about the problems encountered, and how issues were solved.

          1. Separating out the ‘hard realtime’ tasks onto an AtMel CPU makes a ton of sense. This design will to be around for a long time – and SOC PC’s like the BeagleBone come and go and change over time at an alarming rate. The lasersaur design puts code that is highly portable and not hard-realtime into the BeagleBone – and the design makes it trivial to replace that with (for example) a Raspberry Pi – or even a laptop. All of the hard-realtime and safety-critical stuff (which you really don’t want to be running on a ‘soft-realtime’ operating system like Linux) are running on that bare-to-the-metal AtMel chip – which will be around for a very long time. It’s not like the cost of the AtMel chip (~$1) makes even the slightest dent in a $10,000 machine…so it’s a luxury that’s easy to afford.

            So, yeah – sure that code COULD be running in some weird dark corner of the BeagleBone – but good luck with that when the next generation of SOC comes out and you suddenly can’t buy BBB’s anymore! Lasersaur has transitioned gracefully from running on some external computer, to the BeagleBone ‘white’ and the BeagleBone Black and the Raspberry Pi over the last few years. Good luck with future-proofing if you use those compute units on the BBB.

            …and all of that porting effort to save a dollar? Nah – spend your time where it’s really needed. Get the price of the frame and the drive system down. Make versions that are smaller and larger than the Lasersaur. But for chrissakes – think safety all the time. These are not toys – they are very dangerous machines with the capability to self-destruct, blind you, set fire to your workshop – and if not used wisely, poison you and shred your lungs.

    1. The cutter seems to require a parallel port to run, so while it’s open source, the suggestion of making a new machine that uses an abandoned computer connector puts it in the reject pile for me. I’ve sold off my PlasmaCAM partly because of that. It was a used one from several years back, but even new models use parallel ports and it’s not reliable, and I’ve tried seemingly everything to get the machine and computer communicate in the mode PlasmaCAM considers ideal.

      My USB-based Versalaser didn’t give me that kind of problem, it just connected.

      This one is new to me, I’ll have to look closer.

      1. With either buildlog or axCut, the physical build of the machine can easily be separated from the control system. Simplifying things a bit the physical build is really just two stepper motors and an on/off laser control. (In the sense that a 3D printer is just 3 stepper motors and an on/off extruder control.)

        The gcode / Mach3 / parallel port / stepper driver route is simple. I chose it because I’ve already got the same setup for a CNC mill and I’m used to the toolchain. It works really well for vector stuff but is awkward for raster engraving. The other end is an off-the-shelf laser DSP – much slicker but not cheap.

        Any board that can control a 3D printer should also be up to the job of controlling a laser from a hardware point of view. I’m not sure that there’s any readily available open source DSP software (the equivalent of a slicer) available at the moment. Maybe someone else can enlighten me if there is.

        1. LAOS doesn’t seem to be a board, but a firmware for boards that seem to be hard to find.

          Smoothieboard has been made, but they acknowledge their laser firmware isn’t mature yet.

          I’m also kind of making it a point that it’s disappointing that people are still building new machines using parallel ports in the 2010s. That’s their business, but I have a very hard time getting on board with it, and it makes me wonder what other dubious choices are in the design.

          1. Parallel ports are proven reliable and there is no issue in using them. I run my CNC and Laser from it with no issues and know many many people who do the same. There is no issues with drivers, its plug and play in its simplest form.

          2. >>>
            I’m also kind of making it a point that it’s disappointing that people are still building new machines using parallel ports in the 2010s. That’s their business, but I have a very hard time getting on board with it, and it makes me wonder what other dubious choices are in the design.

            I have a hard time getting on board the notion that everything must be on the USB port. I have always had problems with USB port mice, but the PS2 mice have always worked just fine.

            Then there’s the people who always demand that to correct such issues, one must purchase a faster system. This notion is not sustainable. As history has shown, there is a limit to just how fast and powerful a computer can be, so demanding faster and faster processors, or more and more RAM is illogical.

            In some cases, retaining legacy ports and protocols makes more sense as it reduces the amount of data being transferred over the USB port. Dividing tasks amongst differing ports has always been more efficient than force feeding all the data down one single port. And the alleged speed of the port is quite irrelevant. At some point, it will always become a bottle neck that prevents the desired outcome.

            Besides, it’s better to re-purpose an older system when possible rather than sending it to the landfill.

          3. I’ve not seen a problem with any USB mice, I really don’t understand how you could possibly be getting so much trouble with them.

            Not only do most computers have more than one USB port, the USB protocol is very good at sharing different streams of data. Maybe not as good as say, Ethernet (an even better port for printing, etc.), it’s worked incredibly well.

            The laser engraver I bought in 2006 has a USB port. The connection has worked fine all along.

            I’m all for repurposing old computers, but there’s a limit to that. I think the laptop connected to my laser engraver is 12 years old now. It does have a parallel port, but it has only been tried on a PlasmaCAM, and it only worked passably.

            The laptop I just bought that might replace it, built in 2009, doesn’t have a parallel port.

            Even Linux distributions don’t always maintain support for old computers, for example, a new version of Mint won’t run on a 32 bit CPU that doesn’t have PAE support.

          4. I’ve read that Laptops have issues with parallel ports for the purpose of CNC. I can only speculate the cause.
            I would look at power saving settings and make sure nothing goes to sleep.

            As far as USB mice, yeah, I don’t understand the problem either, I get delays, I have to actually click the mouse button sometimes to get the thing working, even though I have the settings set to “not sleep”.

            I used to have a trackball which I liked very much, then it died. In it’s place I decided to get the Logitech wireless track ball that is claimed to be ergonomically designed. Seriously, I would fire an employee for lying to me like that. In the future, I will not purchase a Logitech anything since they have shown that they prefer eye candy over true functionality. I posted a scathing review on because it is so poorly designed and not ergonomic at all.

            It’s pretty much like Java. Someone always claims that: “the newest version is faster and more efficient”; then I install it on my computer and find that it is just as slow pokee as the previous version. I think they make millisecond savings and then declare it to be faster which is correct on a technical level, but a lie on the actual user level as milliseconds of speed are not perceivable to us mere mortals.

            I have always faired better with mice on the PS2 port. They tend to respond faster. I guess if you have the latest greatest fastest computer, it won’t matter, but as I said before, there is a point when that won’t work either.

          5. I’ve tried several different computers on that machine, and that laptop beat all but one tower. That tower was just not able to work well with our work flow.

            Oh well.

          6. I’ll grant the port generally isn’t crap, but I think it’s annoying to use, and it seems improper for new builds. I expect my laser engraver will have a 10 year life or more (mine is about 7 years old now), if you build a laser engraver with a parallel port now, how easy do you think it will be to keep a computer for it to year 2024?

            And difficult to find on notebooks, notebooks have begun dropping the port eight years ago now, and this counting the workstation notebooks too. It’s not on my Dell Precision M6400 notebook, and that was introduced 2008.

          7. If you spend the money on a laser engraver, then buying a couple of expansion boards to future-proof yourself is not that much more expensive. another option for your particular notebook could be something like this:

            Theres also usb-parallel converters but those kind of things are a bit hit-and-miss with regards to build-quality and may not be supported by/work with the machine. so that is only a “hack” solution for the worst-case.

            I really doubt you will have that much trouble in 10 years, as plenty of industrial equimpent use it and will not be written off over such a short time. But i concede that it will likely not become easier to find in the future, especially not on “minimal computers” like notebooks, where space is a premium and they have to cover all the “new” required ports (a handfull of usb, a couple of different video-outs etc) parallel ports take up a lot of space (but is still often traced out to a header on the motherboard, even on laptops – incase one dont have issues with opening up and messing inside)

            I fully respect whatever you decide, i am not going to try to force you to use something against your will – and it is your money to spend as you please. I just didnt agree with your absolecense(sp?) argument and just wanted to point out that there is still plenty of options (which may or may not fit your particular situation/economy of course) and that you probably worry a bit too much about it. If its as simple as “rather wanting usb” then that is fine too, just dont take it out on the poor old parallel port :)

            USB might even carry its own “too new” issue in some shops – where the old trusty PC from last millenium not even having USB available is still running things – with no update in sight. industry tends to move slow as heck – if it still runs, its not getting replaced – but this is merely a thought, not something i can back up.

          8. The problem with laser cutters (which makes them entirely different from 3D printers and CNC machines and such) is that the speed at which the head moves has to be controlled carefully alongside the laser power. Using a PC with a parallel port to drive the stepper motors results in motor stalls if for some reason the PC takes a long interrupt for some OS-related reason and doesn’t get the stepper pulses out fast enough. When cutting some materials, the resulting glitch is enough to start a fire – or at least to wreck your job.

            You really NEED all of those low-level and safety-related things to be done either in dedicated hardware – or in bare-to-the-metal microcontrollers where you know where every microsecond of time is going. The price of microcontrollers has fallen to levels where the cost it totally negligible compared to your $1000 laser tube…and the g-code to stepper/laserPWM software already exists. So there is really no point in not going that route.

            Given that you have that low level controller doing the low-level grunt work – you might as well sent it’s high level commands via a more modern interface like USB.

            So the idea of using a horribly outmoded parallel port to drive a laser cutter is simply ridiculous.

            Lasersaur takes it one step further by putting an entire Linux computer inside the box – so now you access the laser cutter via the web – either WiFi or cabled Internet. The issue of software portability goes away – all you need to access it is a web browser. I can (and do) control my two lasersaurs with Apple laptops, Windows PC’s and Linux laptops – I’ve even used my phone to control one of the machines while I was debugging a motor controller problem!

            Parallel ports are truly obsolete. They need to go away now.

      2. Insulting the parallel port (on a HW-hacking related site no less – where many USB related hacks have been build by simply offering USB to an already welldesigned parallel/serial port device) seems pretty narrowminded.

        As far as reliablility of Parallel vs. USB – that is very likely a per-product issue. ive had a lot of horribly behaving USB devices too.

        Feel free not to buy or build parallel port products. but calling them abandoned is close to deliberate lying – it might not be on your average cheap DELL. but it is not like it is hard to find parallel port options (buildin or as expansion boards).

        1. As @JRDM say, the parallel port has been dead for years, along with the serial port, Firewire, IDE, floppy drives and DVDs.

          Good riddance to them all.

          The parallel port hung on for longer than expected, but if you really need one then get one off eBay. Very few motherboards have the parallel port, and even the header is started to go missing.

          The advantage of the parallel port is simplicity; all you need is a break-out board whereas with USB or Ethernet you need a driver board – far more expensive but getting cheaper.

          My machines are a mix, the ones I build or convert are parallel, purchased are USB.

          1. It’s pretty trivial to drive the steppers with an Arduino (the lasersaur does it) – then your high level operations can be done on a PC, transmitted as G-code to the Arduino, which handles the ‘hard realtime’ tasks.

            With a laser cutter, you can’t have the machine just stalling when the CPU is too busy to update the stepper motors. The head has to be smoothly accelerated and decelerated – and the laser power dynamically adjusted to the speed that the head is actually travelling at or else fine, delicate etching will dig big ugly holes into the material whenever the CPU has to go off and service a hard drive interrupt or something.

            Putting the ‘hard-realtime’ tasks off onto the Arduino means that the PC can do whatever it needs to stream G-code into the Arduino – and not have to worry about managing the system at that low level. Since you can buy and Arduino clone for under $10 – that’s a deal.

            Furthermore, because it’s just a USB interface – I can use any computer to drive it.

            In my case, I have a BeagleBone Black (running Linux) driving the USB port to the Arduino, and (using the lasersaur software) presenting a web-based interface for downloading drawings, scheduling cuts, setting power levels, testing, aligning, etc.

            With that setup, the embedded Linux can be reached from anywhere in my home/office/workshop – even from a phone or tablet.

            Parallel port interfaces suck because you can’t easily buy them anymore – and they presume that your PC is capable of doing things like accelerating the head smoothely without every losing a millisecond to go off and do something else – and that’s just not reasonable.

  1. Nice work. Damian mentions as an inspiration. That’s always worth checking out if you’re thinking about building your own laser cutter.

    The thing that put me off a buildlog-inspired design is that there are a few pieces that are difficult to make at home (e.g. gantry ends). I haven’t looked at the design in detail, but if Damian’s avoided pitfalls such this, then that’s great.

    I ended up going down the cheaper and easier route of buying one of the DC-K40 eBay specials and converting it to something workable. (They don’t cost much more than just a laser tube and PSU anyway.)

  2. thanks all… couple of replies:

    Hack Man – the Lasersaur looks like an excellent machine, and I’ve certainly used much of the published info as reference/inspiration. For me, cost was the major factor in not just replicating the design. I also wanted to modify the working area (to suit A1 stock) and alter the frame design to allow pass-through materials.

    Hans Peter / JRDM / Oxfred – the buildlog 2.x laser is also a great design, but with limited build area and dependence on machined parts. I wanted to make use of my 3D printer for as many connecting elements as possible, this has also helped keep the cost down. The key point here is you need access to a 3D printer to replicate my design!

    I’ll be publishing the BOM fairly soon (will add to the git repo) if anyone is interested in where I sourced my parts. If you’re not in the UK, then the Lasersaur BOM/supplier list would be a good place to start.

      1. halfdead14 – very good, but depends what you’re aiming for – metal will always be stiffer. The overall machine is already quite rigid, and that’s without the cladding panels (which will be 3mm dibond). the cladding is intended to further square/stiffen the frame.

    1. Hey Damian!
      It’s a cool build No doubt. The reason I brought the system to the table is that I’m building that one.
      You’re absolutely right about the limited working area. it might become too small. I like your design since it really utilizes the envelope of the machine. I chose to redo all parts in 3D as well, and redesigning them to metric units. Also, changing the lift plates to printed ones, while maintaining the floating nut principle.

      The worst thing is currently to get MakerSlide pieces :)

  3. In the comments there is discussion about parallel ports..I’ve looked but can’t see anybody stating the fact that since windows 7 , parallel port bit bashing isn’t supported by windows 64 bit osperaing systems, making the choice a little more obvious on which way to go I think. So for example Mach3 is not supporteed on windows 64bit OS. Aplogies if this has already been covered

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.