DIYLILCNC: Do-it-yourself CNC Mill


The DIY LIL CNC project is the newest member of the homebrew fabrication scene. This is a three-axis CNC mill that can be built by anyone with basic shop skills and about $700 in their pocket. Many of the materials can be acquired from the likes of Home Depot: the basic framework is assembled from Masonite, while other cost-cutting measures include the use of skate bearings and a common Dremel tool for powering the cutting bit. About half of the cost is for the HobbyCNC driver and stepper motor package that runs the show.

The instructions for the DIY LIL CNC are distributed under a Creative Commons license, allowing for modification and distribution with few restrictions. They’re well-written and quite thorough, including all patterns and a complete bill of materials with suppliers, part numbers and costs. As documented, the ’bot can produce parts up to 12 x 14 x 2 inches, but the project’s creators offer some suggestions on adapting the design for larger work. It’s not self-replicating like the RepRap aims for; you’ll need access to a laser cutter for some of the parts. If you can clear that hurdle, this looks like a great introduction to CNC production.

39 thoughts on “DIYLILCNC: Do-it-yourself CNC Mill

  1. I’ve been looking for a decent way to make PCBs. I’ve been following the postings on this site, but have often wondered if a homebrew CNC machine could perform this task fairly well (among numerous other tasks).

    Also, does anyone know of services that provide laser cutting for fairly cheap? If so, what types of drawings does one need to provide to these types of services? I’ve done some basic technical drawings in AutoCAD back in the day, but certainly don’t have these tools available to me now. (Granted, there’s pretty much an open source tool for everything :) )

  2. Its a very well designed and built machine. But it uses a dremel… Which is okay if you only want to draw with pens, cut foam, or really soft material. I tried using a dremel on my first CNC build and it failed miserably. Upgraded to a router and made all of the difference.

    So if you are going to drop $700 on a CNC build, might as well go with a router and not a dremel unless you only want to cut foam.

    Aside from this, the build is very nice.

  3. @Jon
    Its possible to mill pcb’s on a homemade cnc machine, I do that. I cut some shortcuts and got an old Agfa flat bed scanner with dual rails (unlike the new ones with only one rail) and used it as a base for my cnc machine. If you want more precision I would advice getting a Proxxon IB/E rotary tool, which is 20k max rpm and very low runout thanks to the aluminum body and 3 lip steel collets. Check out for all of your cnc needs.

  4. How would this (homemade cnc) compare to the toner transfer method for doing PCBs? I’m interested in using it primarily for through hole as i haven’t gotten to smd yet, and figure i can kill two birds with one stone by using this for pcbs instead of investing into toner transfer as well.

  5. I have always figured it must be easier in the long run to do PCBs with a CNC than the toner method; plus having the CNC is going to be an advantage to the hobbyist more than a laminator…

    But $700 is still pretty high, especially in this economy. Although Konstantin’s comment about a simple CNC made out of a flatbed scanner gives me some hope for a low-cost solution.

  6. i just don’t understand why people want to make pcbs with something like this. cnc routers do not work well for pcbs. toner transfer is spit easy, and costs next to nothing if you already have acess to a laser printer and an iron. it yeilds fine traces suitable for qfps and 805 or smaller smd bits. or you can spend $700+ and several months building a complex machine like this, figure out the right software to convert your board design to a vector files, and if you’re lucky, create a loosely spaced, sloppy pth board.

  7. Everything I have seen about doing (decent) toner transfer has looked ridiculously convoluted and time-consuming. In fact, the last one posted on HaD appears to be an all-day event:

    Even if it only takes half the effort and time that the video(s) seems to indicate, the vast majority of people would rather stick with protoboard than go through all that. If the longest part of your project is masking and etching your PCB, you might have a problem.

    Direct printing onto the PCB makes a lot more sense and would be the best option for the individual, but only a few printers can be modified for it from what I have seen.

  8. The big advantage of milling a PCB is that you don’t have to spend time drilling all the pin holes, milling the board if it’s non-rectangular, and routing holes for any large through-components.
    That, and no need to have any chemicals on hand, no need to be continually involved in the process (put the PCB in the CNC machine, have a cup of tea and a sandwich), and no need for three or four pieces of equipment to do one job.

  9. Here is a video of my flatbed scanner mod to cnc.
    It is a nice thing to have to prototype fast, like if you are doing your design in Eagle and you run the gcode-pcb plugin you get the gcode to mill the pcb.
    For production it is faster to use toner transfer method imho, like in I must say that the toner transfer paper in this vid is a beauty, because it just peels off totally when wet and all you need basically is the paper, a printer and iron.
    It will take a while to get a working CNC machine with close tolerances to do decent job, plus time to generate gcode, plus more money on bits and bit sharpening.
    CNC is intersting as a whole and very nice learning but if you need to produce pcb’s fast there is no better way than toner transfer for the moment.
    Printing with a laser printer directly to the pcb would be the holy grail, anyone has a link where it has been done successfully?

  10. The site http://www.buildyourcnc.Com shows step by step videos to create a complete cnc machine with basic tools (no laser cutting needed.) This is a great site to see what is needed to get into the cnc hobby without a lot of money. I built a machine simillar to his video tutorial and it works great for what is costs.

  11. DIY CNC machine designs are always cool, but they always seem to be designed for soft materials like balsa wood.

    I have been searching for a while, but can’t find a good inexpensive solution for milling metals like aluminum. The best design I have found is this amazing one:
    .. but this person went all-out so it’s costs are in the thousands.

    Other than the harbor freight mill cnc conversions, are there other solutions out there that I have missed?

  12. I just had to post again to say this project is great in the fact they have a parts list that specifies not only exactly what you need, but also where to buy it and costs. That is thorough! Somebody deserves a beer.

  13. The place I work has a cnc mill that was made to make PCB’s it works great thing is it cost $12k and it take tweaking to get it to work right. I am building my own mill to do PCB’s and what not but they are not the easiest to do and get fine traces.

  14. Ive made an awful lot of custom PCBs in my day (toner transfer, photoresist, sharpie resist, CNC, ghetto dremel human cnc). Ive even made a couple of homebrew CNC machines specifically for making PCBs. Ive discovered a few things.
    1) Etching is generally easier/ cheaper/ more portable than milling.
    2)etched boards are much easier to troubleshoot if done correctly
    3)Making a CNC specifically for PCBs is just not worth your time. If you are going to make a CNC machine, make sure you make it tough enough to handle other stuff.

    4) If I were to build another mill: I would design a rock solid 2 axis machine that could be mounted to the table of a drill press. I would continue to etch my boards, but I would use the 2 axis table to drill the holes.

  15. I’m on the side of those saying that “milling” and “PCB” shouldn’t normally be used in the same sentence. I have access to a PCB mill and it’s clear to me that it’s not the best solution. And besides, you still don’t get PTH which is a PITA.

    The cost of the milling parts are expensive, comparable to the cost of sending it out (not domestically though) and just about as fast for jobs that are 0.5 sq.ft double-sided. You waste a significant amount of base material and tool wear getting it to work right, adding to the time and cost.

    To me, direct print or laser resist-removal (ideally *clad* removal!) seems to be the most practical, affordable for a local group of like-minded makers. One place for the equipment, chemistry set and storage/disposal.

    One thing that I never hear anyone talk about is the possibility doing PTH at “home”, but I’m not familiar with the process. Are there no feasible, accessible chemical/electroplating processes for this?

    Really guys, just say “no” to PCB milling, unless you’ve well over $10-20K for equipment, a taste for substantial recurring tool costs and a lot of time/patience.

    My $0.014 worth.

  16. my school has a pcb milling machine, which cost $10k, i’ve used it several times and it can cut some very fine traces. i’ve even seen smd’s that it’s cut and it works great even for sot-323 transistors, which are extremely small. i’ve been building my own, machine but i don’t expect that kind of quality. i do hope i’ll be able to make some decent through hole boards though.

  17. Another very nice project is the MTM mini developed by Jonathan Ward, a low cost (less than 400dol) milling machine to make printed circuit boards, that uses small motors instead of a Dremel. The plywood structure can be milled in less than one hour, and complete assembly takes about 2-3 hours.

    We used it for PCB fabbing and also wax mold milling and the results are great, for example we did this Isp programmer with it && The precision of the machine is far better than our Modela (using 1/64th endmills).

    Personnaly, I think that for milling small single sided FR1 copper-clad laminate (phenolic paper) board + SMD components a small machine like the MTM mini is perfect. It takes usually 30 min for a small board (something like the maximum size of eagle free edition) and it is very secure and clean (you can do something else during this time). More infos on the PCB milling process here and also more infos on components used in the class (carbide endmill, where to buy FR1) at


  18. The operative word is “can”. The one I’ve used “can” cut some very fine traces, but too often it doesn’t. Part of the board (or other boards on the same blank) will be ok but other parts may not be.

    For small boards and/or one-offs, it can be quite useful.

    Hopefully one day we’ll get it tweaked so that we can count on it. But in the meantime, it’s very demanding.

    You’d think it were made by Jaguar… :-)

  19. Wouldn’t it be simpler to control the thing with the RepRap electronics? It’s cheaper (only $128 for the boards, not including steppers), and you can use Skeinforge/ ReplicatorG to convert your model files to toolpaths. Seems like it’d be easier.


  20. @svofski
    I havent made a post about the machine, since its not a big deal I guess. I have some photos of it here

    If your mill has the same Z level on the whole pcb then you can mill just the surface of the copper off the blank, I can do that on my mill, it takes a while to set up but it is doable. While milling only the copper away you mill as less as possible and the overall quality is very good.

    Higher cost brand mills face same problems too. The spindle selection is paramount if you want as low run-out as possible, again I go for Proxxon IB/E.
    In my mill I had to make a custom 1/8 bit holder using a spindle from an OOOOOld hard drive (it was of 129MB with bad sector count printed on the label and it hosted 8 plates) and turning a shaft from round stock to accept Dremel attachments.

    Dremel collets are made from aluminum and they are far from good quality (for pcb milling that is).
    One thing very gratifying is to have the pcb drilled perfectly, I just swap the bits when the cnc program (mach3) asks for it.
    Prototyping this way is fast this way, for production of a well designed board I’d say go for TT or outsource.

  21. I recently isolation routed a small run of 20 boards for a robotics class. They were double sided and contain a mix of SMD and through hole.

    It really does work great and each board was taking ~20mins on a pretty slow machine for both sides including 3 drill sizes. Done a lot of etched boards too…etching sucks! CNC rocks!

    Biggest drag with home made boards is no plate-through but that is problem with CNC or etching.

    A bit more (gettting rather old) info here…

  22. @10bulls: how big were your boards? I’m guessing about 2.5 How many cutters and endmill bits did you go through? 50 double-sided on our mill is probably $80 worth of bits if you end-mill out the guts, and probably $55-60 if you isolate only ~16mils.

    Can’t agree more about missing PTHs. Tin coating is nice too, but at least that’s doable. Plus, if it’s only for prototypes, you’re throwing them away anyways.

  23. @ericwertz
    Yes, boards are smallish :
    Cutters last me ages…I use 60 or 45 degree solid carbides which cost me about 7EUR ($10). I tend to snap them before I wear them out (doh!).
    What rpm are you using? $55-$60 of cutters sounds a lot! I try to just break through the copper so I don’t cut much fiber glass which will wear the cutters quick.
    Yes, I only do isolation (up to ~19mil) – flood fills take too long and aren’t really needed for the prototypes I do.

  24. @10bulls
    I’d love a name/source/whatever for your bits. After this first project, I’ve been wanting to find some acceptable non-manufacturer bits. The manufacturer’s cutting bit has a suggested tool life that appears to yield 30 meters of use at $18/bit. I probably went through about 3.5 bits for 80 x 2 sides (150 sqin or so total).
    Drilling, fill removal and routing (180in) probably doubles that cost.

    RPM’s 30K if I remember correctly.

  25. @ericwertz: I’ve used these…

    There is also a guy in the states:
    $6.50ea !!! – I’ve not used his, but I know he has a good reputation on the pcb-gcode forums.

    Yeah, if someone was selling me a cutter for $18, I’d be a bit wary if they say it will only do 30 meters – bit like buying a porsche and them saying ‘You’ll need to buy a new one after 100k miles … or it’ll explode!!!’ ;-)

    If your cuts start to get a little fuzzy, a few wipes with a diamond hone may help to give your cutters an edge.

    Also, having a good spindle with no runout is a must.

    Good luck!

  26. Hey guys, I am looking to possibly build a CNC machine to cut some 14gauge steel I have. Do you guys know if a set up like this is capable of cutting material like this, and how accurate it would be compared to something like plasma or lazer cutting?
    Thanks for any help in advance

  27. Well, I have a Sable CNC (not the small 2015, a larger version 3000-something) and it cuts very fine traces. Also using a small probing g-code I can adjust the height of the g-code so it doesn’t matter if the board is level or not. It’s true, it takes longer than the toner transfer but at least it’s working every single time and I don’t have to work with chemicals that I have to dispose of later. So instead of working one hour on a board, trying to get the sides aligned, drilling the holes by hand, etc, I just go to the machine, put the board, run the drilling code, 5 minutes later I change the tool, run the isolation code for one side, watch tv for 40-60 minutes, go back, flip the board during the commercial break, watch tv for another hour, and done, i have my board ready.
    Now I just have to find a way to do through-hole plating…

  28. There’s probably more than one product out there that does this, but LPKF does sell a product for this. I think that it’s a (one-part, I’m assuming) conductive epoxy that you squeegee through the holes, bake it and then drill through it.

    You might want to check it (and similar products) out. That’s the general process though.

Leave a Reply

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