Building A CNC Milling Machine For Less Than $1300

CNC milling machine

[Mynasru] tipped us about a homemade CNC milling machine that his friend [trochilidesign] recently made. We have to admit it may be one of the best ones we’ve featured so far on Hackaday, mainly due to its elegant design (see picture above) and its all metal structure with linear guide rails. In the very well detailed write-up, we can gather that the CNC machine was designed using SolidWorks.

The main frame is built around 2 Maytec 40x80mm profiles and 2 endplates made from 10mm thick aluminum. 3 Nema 23 stepper motors and their drivers power the build, all of them bought on ebay. Finally, the Mach3 CNC software was chosen to interpret the G code and send the appropriate control signals.

Due to licensing restrictions the original author can only provide us with PDF files detailing each part of the machine, but we’re sure this should already be enough for interested persons out there.

34 thoughts on “Building A CNC Milling Machine For Less Than $1300

    1. The rails handle dust pretty well if you use the little plastic plugs that go in the mounting holes over the rail screws. Even without them they handle crud pretty well.

        1. You could put bellows on. Two problems with bellows. First they usually end up taking away from your stroke since they do not compress all the way. Second, assuming standard pleated bellows like on most linear actuators, they are expensive. He will spend almost 1/3rd of the cost of the machine to put bellows on it.

          And for what? The chances of the machine getting enough use are incredibly slim. These rail will last a very,very long time if they are kept lubricated.

  1. Does anybody have a link or pointer on the ‘license issues’ that kept him from releasing the SW files? For all I know you will get a ‘educational use only’ stamp over all your drawings, but you can still share them. Or did SW really sober up and started proscribing you should not use SW for open source projects?
    Because that would really be a good advise. ‘do not use Solid Works at all’ would be better.

    1. Yes, I do ;-) I’m the author of the Instructables. Since I’m working on a educational version of Solidworks at my university, they do not allow me to share the CAD models (the drawings were all made by myself).
      Secondly, since Solidworks is completely parametric, the design could be modified easily and used for commercial purposes. Thats not the intention! But I’m sure this is already enough for interested persons out there to build one or modify the design for the opensource community.

      As is the case with the popular Ultimaker files. They share their drawing, but CAD models are not available.

      1. So, don’t do that.

        What’s the point of making something if you can’t share it? Can’t you export the model in another format? Can you export the drawings and allow someone else to construct a model from them?

        I don’t blame you, but what other alternatives are available?

      2. Sounds like the University is simply making sure that nothing escapes into the wild that could aggravate Solidworks or would let something of value to the University out, although a Solidworks viewer would allow a non-editable version out, but then, so do PDF files, which can be valuable as a starting point for a similar design.

        1. Also, because this was made using “university resources” generally, it is all owned by the University.

          Remember students, EVERYTHING you do while at a University is OWNED by the university you attend, even if it had nothing to do with what you worked on! The burden is generally on you to prove that the intellectual property rights are yours.

  2. I’ve been interested in building something like this. I have been wondering about the shear resistance of the long axis with a single center leadscrew, i.e. if you push the left side back and the right side forward, how much does it flex?. This would be the same kind of load the machine would see when milling off center. Some people use two leadscrews on this axis but have wondered if this was necessary.

    1. One way to resist those deflective forces is to provide a wide bearing platform on the Y-axis rails. On this design the bearing width appears rather short, though I haven’t dug into the drawings.

      The gotcha with a longer bearing platform is that is quickly reduces the available amount of Y travel.

      I don’t see a spec or source for the bearings or rail used to support the gantry. That is an important detail, as those bearings can be rather expensive. There are a few ways of solving that problem, inexpensive acetal bushings among them.

      Unfortunately, we rarely get to read the results of engineering analysis and measurement of various implementation methods.

  3. So how would this rank against the Shapeoko v2? Obviously, the two haven’t been compared side-by-side, but I’d truly appreciate a professional’s opinion on the differences, as I’m in the market.

    1. Shapeoko 2 is looks a lot smaller. This project looks considerably beefier. It’s possible SO2 is faster for very light cutting, but I think you’d want this machine over SO2 for most work. There are a lot of hobby CNC router designs, so I’d do a lot of research before settling on any one design.

  4. On the topic of guiderail material choice, I must disagree with this comment from the build log:

    “Since you are constantly fighting the forces from the endmills against the material of the workpiece, a lot of support is recommended.”

    The most significant forces involved are from the stepper motors, as they attempt to accelerate the substantial mass of the machine. Especially in a moving gantry design. The moving gantry designs tend to deflect and bounce under acceleration. That is not to say that machining does not generate force, but those forces are often less significant.

    One need only attach a dial indicator to the head of the machine, and move the X and Y across a flat surface, to observe the deflection that results. On this style of machine, the X moves are usually fairly stable, but stability in the Y is a major challenge.

    A dial indicator can be attached in other locations, and force applied via lever arms and hanging weights, to isolate and quantify the sources of deflection. Of course you could also go the theory route, and model the system.

    1. It’s probably brilliant, but there’s so much old info and linkrot. You can only access Joe’s forums if you buy his plans. The link to the 2006 model linked to a thread at (and it goes to an r1, I think). I bounced between several discussions on cnczone (one is hundreds of pages long) and then I surfed long enough to see references to an R2 and found a link for R2, but it’s a rapidshare link that 404s. (So I’m still looking for the r2 link).

      It’s frustrating for such an old project (that’s obviously successful) to be so uninviting to someone new to it. The signal to noise ratio is pretty low.

      Sometimes engineers forget how to make a project easily digested.

      Putting some info behind a paywall means I can’t decide if I want to buy the info (but r2 isn’t a pay version anyway…. so r2 just is out of luck with his forums I guess) and putting other info on, which is an open-ended conversation, is just frustrating. Forums suck as a primary source of info unless you have a specific question you need an answer to and are part of the community.

      I’m not a CNC guy, but I’d like to be.

    2. Movement without any load should be smooth and require very little power unless you’re accelerating the gantry at high speeds and even then the forces will not even begin to compare to those experienced when machining. Any machine with half-decent linear guides or rails should be easy to push with a fingertip and should not bounce or deflect even during rapids.

  5. I guess nobody at Hack-A-Day has seen for the past, dunno, 6-8 years? Oh ya, this “new” instructable pretty much what the Joe2006 is. The new 4×4 is buildable for about $1100, and the one in my garage is pretty much on par with the Shopbot that my local hackerspace has.

    Just because it isn’t on Instrucables doesn’t mean it doesn’t exist in the DIY community guys. Sorry to sound grumpy, but it just annoys me when I see stuff that is waaaay behind the curve get posted over there, and hailed as an amazing new step forward.

      1. See that’s the thing, it’s like “get the word out”. Ok, I’ll start submitting links to stuff that’s 5-6 years old if it helps. I found all of them just by googling “DIY CNC homebuilt” years ago, and hit upon several massive open forums in the first page of results. I guess I just figured anyone interested in it would do the same. Instead, everyone just waits for an instructable or a kickstarter to make the news.

        1. OK, first, if this article adds superlatives, then maybe it shouldn’t, but this article didn’t say it was revolutionary that I can see.

          However, it’s a big world out there. So many things out there, CNC routers just one of them. For one thing to get noticed, it’s more likely to stumble across something than just think it up and do a Google search.

          I don’t think it’s fair to expect a casual observer to know things about specific designs compared to other designs. There’s a lot to it, and tiny differences matter.

    1. Bah, I replied to the wrong person earlier. I meant for my reply to be here. Search ‘neutrino’ on this page for my discussion about Joe’s CNC (which I’m still interested in, still clicking on links all these hours later, but no less bothered by its lack of nice docs in an easy to get to place).

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.