Having a few machine tools at one’s disposal is a luxury that not many of us are afforded, and often an expensive one at that. It is something that a large percentage of us may dream about, though, and with some commonly available tools and inexpensive electronics a few people have put together some very inexpensive CNC machines. The latest is the Minamil, which uses a rotary tool and straps it to an economical frame in order to get a functional CNC mill setup working.
This project boasts impressively low costs at around $15 per axis. Each axis uses readily available parts such as bearings and threaded rods that are readily installed in the mill, and for a cutting head the build is based on a Dremel-like rotary tool that has a similarly low price tag. Let’s not ignore the essentially free counterweight that is used.
For control, an Arduino with a CNC shield powers the three-axis device which is likely the bulk of the cost of this project. [Paul McClay] also points out that a lot of the material he needed for this build can be salvaged from things like old printers, so the $45 price tag is a ceiling, not a floor.
The Minamil has been demonstrated milling a wide variety of materials with excellent precision. Both acrylic and aluminum are able to be worked with this machine, but [Paul] also demonstrates it in its capacity to mill PCBs. It does have some limitations but for the price it seems that this mill can’t be beat, even compared to his previous CNC build which repurposed old CD drives.
I like the design :)
Thanks Niklas!
For a capstone project, I built something only slightly more polished out of spite for the professor who didn’t think it could be accomplished for under several hundred dollars. Pretty sure I used the same Harbor Freight grade “spindle” in my design
Any so called professor who discourages ideas is no professor. Q.E.D.
Sounds like rb’s professor encouraged them to me!
And if rb hadn’t been the type to rise to that challenge then maybe rb wouldn’t have been the type to complete the project anyway.
The professor sounds old and bitter.
A bit ageist?
My worst university instructors were typically young and cynical.
My favorite instructor (did his 2 calc and linear algebra and discrete math courses) was late 50s/early 60s, patient, and passionate about math. My worst college instructor was just a year out of her history PhD, and would not brook introduction of any ideas outside of her published lesson plan.
I am now in my 60s and teach an embedded systems course at a private high school. We had to dismiss the youngest instructor on our team (age 28, MSEE) because of his poor and narrow-minded presentations and lack of passion for teaching kids and the subject matter.
Whatever. YMMV.
I hope Prof appreciated & rewarded your expectation-defying accomplishment!
nice butter and maybe foam mill. A bit more tape and some hot glue would be perfect
Acrylic, MDF/HDF, copper-clad FRn, etc.
Not entirely impotent with aluminum.
Haven’t tried inconel yet ;)
So I’m all for reasonably priced DIY tools, but to me, this doesn’t cut the mustard. It’s this weird combination of ultra cheap materials and very advanced construction techniques requiring relatively expensive tools (laser cutting). This strikes me as somewhat odd if you don’t already have a laser cutter, because the expense of laser-cut parts is mostly in the machine time and not the material. So I’m really not sure what I should make of that, it’s a commendable project, but I just think the effort would be better spent in not trying to make it as cheap as possible given access to expensive tools and free scrap material, but in developing something with both reasonable price and performance. Of course where that point lies is entirely debatable, but I for one don’t see the point in making CNC mills that work mostly on plastic–we have reasonably cheap and easy-to-use 3d printers for making small plastic objects nowadays, so there is just very little reason to have a low-performance CNC for that (maybe transparent acrylic, but to build this CNC mill for the proclaimed low cost, you need a laser cutter, so that’s kind of taken care of already).
You really don’t need your own laser cutter to get parts laser cut, maker spaces and postal services exist, and would be a great tool to leverage to get you a useful tool for your needs at home.
And to me you miss a big point with mini mills, even ones only capable of easier to work materials – they can do things a 3d printer can’t – like be sure that flat face really is flat (assuming its stiff enough to actually cut ‘flat’ – but a 3d print even in a heated chamber etc is bound to develop some warping, and really quite awful if the geometry of the 3d print doesn’t suit. It can also cut PCB traces, probably play Vinyl cutter with a change of tool bit. Its a very versatile tool that compliments a 3dprinter, and for folks maybe even replaces it – your FDM printer has layer lines of meaningful size that cause weaker and uglier parts, and the UV-resins tend to be expensive and brittle, so if you need fine detail parts with some durability better to mill them from cheap solid plastics etc…
I do agree I wouldn’t have tried to build anything this cheap, say double or triple the cost per axis with either the goal of similar easy machining materials but over a larger working envelope or much more capable over similar size. Still be dirt cheap really but much more useful, and if you design it for yourself with a size in each axis that meets your expected needs (at least within your budget/workshop space) its better than the off the shelf budget ones.
Entirely fair to put this in the questionable margin.
To clarify the idea here, it’s actually *not* about expensive tools and free scrap — that’s much better done many times over elsewhere.
The *single* fancy tool is a laser cutter — and I suggest that for increasingly many people laser cutting is actually *more* accessible than traditional fabrication. Foldi-One already made a similar observation.
See the project for cost details for buying all new parts and materials for the working parts. The frame that holds the Z axis in position over the XY axes can be random available material, with very crude scrap shown in the photo to make the point.
__f wrote a very substantial response to a (near) duplicate of my reply to his comment — which vanished when that dup instead of this one got cleaned up. Bummer, because he wrote a lot and raised several points that I wanted to respond to. And I didn’t get email notifications for replies to that half of the dups, so no fallback. But I did write some notes, which could lead to a huge comment here, but maybe a project log entry or three instead.
Comments prompted by just a sliver of __f’s lost reply:
Availability of/access to different kinds of tools varies with location for sure.
What value could this project offer someone who already has access to a laser cutter — and presumably 3d printers and other fab tools because laser cutters are rarely the first tool in a place? My own access to a laser cutter also gets me FDM & SLA 3d printing, an incomparably more capable CNC router, and a bona fide CNC mill, so how could this have any value for me? First, I have to go somewhere else for any of that, for however long it takes to sort out how to do and then do whatever I’m trying to do, assuming tool availability. Also, my access to the CNC mill is only theoretical until I pay the time and effort to qualify for actual access. Then subtractive CAM has a taller learning curve than laser cutting or 3d printing and often, at least for me, requires more iterating to sort out how to make a thing. I can do a lot more learning up that curve at home with a machine that’s just there, never occupied by anyone else, and much less severe about punishing errors.
FDM printers are very accessible and can make lots of useful plastic stuff. But as Foldi-One already commented, subtractive machines can do things that FDM printers can’t, like precise dimensions and surfaces, and use materials that resin printers can’t. Click thru to the lead video on the project for parts that I wouldn’t expect from a common FDM printer.
Thanks for the reply. This might differ from region to region, but at least in my case, having access to expensive tools is a problem in the first place. Sure, I can order cheap laser-cut stuff online, but then I can’t specify the materials. Or I can try and get it produced reasonably locally, which is nigh impossible if you’re not a business with massive order volume. With access to a community-owned laser cutter or other more elaborate tools, there’s little point in building your own ultra low cost tool, because you already have reasonably easy access to much more capable tools. As far as plastics are concerned, I still think run-of-the-mill FDM printers make much more sense for hobbyists than subtractive manufacturing nowadays, since FDM material is readily available and FDM printers are quiet and reasonably clean, so you can just put them in any room that has space, instead of needing something like a dedicated workshop for a mill. I’d be very interested in the kind of metal milling you’ve been able to do. Just stuff like front panels and signs that you could do by hand with a Dremel and a template, or milling a 3d structure from a block of aluminium? Your machine seems to be allright for PCB milling and there’s really not much to complain about there; I don’t quite see the advantage over etching with photoresist, but the machine seems to be quite capable of milling PCBs with sufficient detail.
As for the accessibility of laser cutting: I guess it depends. Table saws and drill presses can be had for quite little money (with quite little precision), they just take up workshop space. A laser cutter is a specialized tool that hasn’t been picked up by many hobbyist-grade manufacturers. It also requires careful maintenance and is quite dangerous when handled incorrectly (as are table saws), but the laser tube and mechanics are just not super cheap. This makes even pretty basic laser cutters far more expensive than pretty basic table saws or drill presses or even 3d printers. In my opinion, if you want to construct something that’s geared towards people without access to fancy tools, constructing it from industry-standard aluminium extrusions and 3d printed parts is a reasonable approach: 3d printers have just become so inexpensive that it’s pretty much a no-brainer to get one if you are into constructing things, so it is far more likely that someone you know has a 3d printer than it is that you can get access to a laser cutter.
3018 desktop CNCs are now in the $100.00 price range. Not sure if it is worth rolling your own anymore.
I was surprised, but found several in the $70 – 150 range. For example,
https://tinyurl.com/45ee3puj
90$ with 70$ shipping, so 160$.
I’m seeing a few down to 130 + free shipping
Still quite cheap. What you get for that price is anyone’s guess
Several years ago I set out to design my own CNC mill. But when I went to price out things like stepper motors, Acme threaded rods, and aluminum extrusions, I quickly surpassed the price of the mills available at the time. So I bought one. It’s probably better than what I would have built, and almost certainly better than the machine described here, in terms of accuracy, precision, and repeatability.
Yes, this is the thing for building a “cheap” machine. However frugal the build, conventional mechanics have a minimum cost.
For example a neat cardboard CNC cleverly constructed from a $1 sheet of cardboard… and ~$200 of steppers, threaded rods, etc. + $150 electronics: http://web.archive.org/web/20200804170432/https://danchen.me/lab/cardboard-cnc/
Maybe buying a 3018-type machine is the current best way to buy a kit of parts for something like that.
Part of the idea here was looking for *different* mechanics to break below the cost floor for NEMA+Acme mechanics.
For accuracy/precision/repeatability, I think this machine does remarkably well within limits of light loads, like PCB milling, and short time. I wouldn’t expect to hit a 0.1mm peck on center again after a week of casual handling, but PCB results on that scale compare well with what I could find of other people posting results from 3018-type machines.
I don’t know what you bought but I’m sure it can bury this in chips.
I’ll second what Paul and others are saying here. A 3018 is a kit of parts that you can get running now, figure out why it’s crappy, and then build into something that works. You’ll probably want to swap out the motor as well, and ….
Check out the YT channel of Minh’s Man Cave:
https://www.youtube.com/channel/UC2WnVQvs841c7HwUIXtaA0Q/videos
About a year or two ago, he did one of the best 3018 conversions I’ve seen — although in the end there’s almost nothing left of the original machine. :)
Well, yeah, that’s why I bought it – if nothing else, I got a pile of 8020 extrusions, some stepper motors and lead screws, and a GRBL board, for much less than they would have cost separately. I figured the motor would be crap, and it is, but it’s sufficient for wood and plastics, which is most of what I needed it for.
There are a number of YouTubers who have gone crazy upgrading their 3018 mills, including one who replaced the side plates (originally laser-cut acrylic on that particular model) with aluminum castings.
Minh’s Man Cave on YT – he’s put up lots of neat stuff.
Minh used his original 3018 for one of the PCB examples that I found to compare against — favorably. (linked in reply to HalfNormal)
Then he spent $250 tweaking up the 3018 before designing a new machine to make with a 3018 + $375. So there’s ~$500 if you skip tweaking the 3018 as use it as-is to make parts for the new build. Then he kept going with ballscrews, more metal, etc. He’s got some $ in that thing.
It would be great if some eager CNC vendor out there sent him a new machine priced similar to what he’s spent on his homebrew so he could compare the two.
I don’t think that one includes anything but the frame, the spindle motor, and the stepper motors. The pics show a power cord but no power supply. Always assume you’ll get nothing more that what the photos show, and sometimes it won’t be quite exactly what’s in the photos.
I bought one for 160 USD shipped, came with everything needed to start (even a box of V bits)
YMMV though.
That’s an existential question for this project: whether or not there is any space for a *reproducible* kit between an improvised one-off build from available scrap and a commercially (i.e. efficiently) produced entry level CNC mill.
You’re right that there are click-to-buy options for converting a little more money into a working machine. Does this design offer any compensation for the hassle of building and smaller result?
Here’s an attempt to compare PCB milling results with what other people say they’re getting from 3018-type machines: https://hackaday.io/project/174370/log/197657-compared-to-what
And I’ll claim smallness is a feature :)
Cleaning up near-duplicates of my first few comments here unfortunately killed the instance of this reply to which BrightBlueJim responded. From email notice of his comment:
“If you are only doing watchmaking or jewelry, maybe smaller is always better, but it’s hard to imagine that you live in a space where a footprint of less than a 400 mm square is too big. A 3018 CNC “mill” takes up less area than the typical 3D filament printer, and not much more than a laptop computer.”
(the deleted near-duplicate to which he’s responding was more explicit about space where I live)
There’s the size of a place, and what an occupant wants to do with it, and imposed constraints. I don’t have a 3d printer here either, and laptops can vanish when not in use. A closed-up box with footprint of 240 x 160mm that I can grab with one hand is more like a laptop that can disappear — and half the area. The telescoping axis design makes the parked machine smaller than in operation.
400mm square is a pretty big footprint, especially if you can’t actually put things on top so it eats all the space above it too – lots of folks don’t live in the large areas of America with those huge cheap timber houses, even Americans can live in apartments etc, and anybody with a small space can really find the tools they want/need have to be bespoke just to fit….
You also have other things than just the floor/desk footprint from a practicality standpoint – if you want to say mill your own die for that RPG you love a little larger than standard you end up needing a machine with much more Z than any of the cheap off the shelf CNC mills I’ve seen – by the time you actually put the tool and some work holding in those you have so little working height available.
So if you have tight space constraints and expect to do more that sheet work you probably end up needing to roll your own, or convert one of the ‘mini’ freestanding mills to CNC (assuming you can fit one – as they tend to have a pretty big working box in comparison, but also are actually pretty large for a ‘desktop’ machine.).
As always with any complex tool the third one you build bespoke for yourself will be perfect (or nearly) for your needs, but anything you buy and the first one you build are bound to have some compromises you have to work around or rebuild to fix…
Okay, so I guess that really shouldn’t be that hard to imagine! I hadn’t thought about the perils of shared spaces.
If you are only doing watchmaking or jewelry, maybe smaller is always better, but it’s hard to imagine that you live in a space where a footprint of less than a 400 mm square is too big. A 3018 CNC “mill” takes up less area than the typical 3D filament printer, and not much more than a laptop computer.
I have a counter weight like that on my microphone stand. If you move the spindle too fast, it will start swinging around.
Here the spindle moves only vertically and with low acceleration (large mass/tiny motor).
The motion at the end of the tilting beam is not *perfectly* vertical, and I suppose it could be *possible* to find a resonance. But that would be a strange tool path.
Reminds me how i got started wit 3d-printing. Furniture sliders, ton of hotglue and random pieces of wood. worked well enought to make some better parts for next revision.
And that got you started! Although I didn’t use it to make any Minamil parts because this was more about *not* making parts apart from the laser-cut pieces, the link at the end of this article points to the “random pieces of…” precursor to this project.
I just thought you’d taken the curtains down to wash them.