CNC mill built from a 3D Printer!

CNC mill built from a 3D Printer!


So i’ve always wanted a CNC mill, arguably
my interest in automated manufacturing in general was what got me into 3D printing in
the first place, since 3D printing is sorta the same thing from a technical standpoint,
just way more accessible. But i still wanted something that could machine
materials and not just extrude them, even if it was only for the educational value and
less so for actually opening up a job shop and running the machine 24/7 to produce parts. Essentially, it should just augment the 3D
printers i already have and be able to occasionally machine a thing or two from aluminum, acrylic
or wood. So i set out to do what i had planned for
a long time with the MendelMax 3: I converted it into a CNC mill. Now, even though the MendelMax 3 is a fairly
sturdy 3d printer, it’s not quite up to the task of machining materials out of the
box, at least when it comes to my basic understanding of the cutting forces and resonance issues
involved with milling. So my padawan and I started replacing a few
essential parts that were originally only made from relatively flimsy 1mm steel – which,
surely, sounds like a lot, and actually is the same ballpark thickness that a large part
of your car’s chassis is made from, but the geometry plays a huge role, too, in how
easily a part will bend. For a lot of applications, a chunky, 3D printed
plastic part can actually be way stiffer than anything that’s just a flat sheet of metal. So after removing the parts that made this
beast a 3D printer, we grabbed some aluminum flat stock and replaced the Z and Y-axis motor
holders and the X-axis carriage, since that’s where the new toolhead will go. The great thing with these OpenBuilds wheels
is that you really don’t have to work precisely at all when it comes to the positions of the
mounting points for the wheels, since you’ll adjust the pretension of the wheels anyways. We also stiffened up the print bed, since,
again, that was simply sheet metal from the factory and didn’t come with any significant
stiffness. To make it less prone to twisting, we spaced
out the openbuilds wheels a bit more, which does make the Y-axis a bit shorter, but that’s
ok. What also helped was adding a two-story MDF
top-plate. The thicker bottom part was screwed to the
original sub-bed and is going to stay there, forever, and the 10mm thick top plate is essentially
a waste plate that will get screwed, milled and drilled into and is just something that
is going to be replaced on a regular basis. Electronics-wise, we didn’t change anything. The motors, drivers, control board etc are
still the original parts, but the firmware is now running slightly slower maximum accelerations
and speeds as well as slightly higher current for the extra weight on the X and Y axis as
well the cutting forces themselves. So the original idea was to use this Proxxon
rotary tool, because it came with a nice mounting flange, has a collet chuck and a wide range
of adjustable RPMs. And it worked great for a while, using these
cheapo 3mm two-flute bits, it milled happily through MDF and even aluminum with, like,
super reduced cutting depths and feeds. Engraving into acrylic also worked beautifully,
but it didn’t last long. The bearing in these tools aren’t great,
well, at the very least they’re not made for machining like this, and after replacing
the original Proxxon tool with a newer that supposedly had an extra bearing for that exact
purpose and still having it fail in almost no time, the MendelMax sat unused for quite
a while. I was frustrated, after spending so much time
on modifying the machine and buying two of the Proxxon tools, it was almost all the way
back to the start. But thanks to the internet, you now have access
to an almost infinite catalogue of parts, and eventually i found this 400W, 48V spindle
motor with an ER11 collet. Sure, it’s technically only a large-ish
brushed motor with a collet chuck on its shaft, but if it worked, i’d have a more powerful
tool, a better chuck and still pay less than buying yet another Dremel. So after getting that and a matching 48V power
supply, the machine was back in business really quickly, and it turned out that the spindle
motor was actually quite solid. I can adjust the RPM by just tweaking the
output voltage of the power supply, but i still want to add something like this DC motor
controller to make it a bit more elegant. Right now, the surface finish can be tweaked
to be, i think, fairly amazing, it definitely needs a finishing pass on any surfaces you’ve
machined and tricks like trochoidal milling won’t hurt, either. Now, the converted MendelMax isn’t perfect
by any means. It’s still a 3D printer frame after all,
never intended to be used for these heavier applications. For example the Z-axis nut holder are still
way too weak for this and will be replaced by printed versions, also the electronics
and the Y-axis rails for the openbuilds wheels could use some extra protections from the
chips, with electronics, obviously, for not shorting stuff out with aluminum flakes, but
the OpenBuilds wheels are also extremely prone to just picking up chips that sit on the rail
and then end up rolling everything but smoothly. I also sorta want to replace the regular GT2-2M
6mm belts, which are often just called GT2, with something maybe at least a bit wider
to make it stiffer, maybe even a leadscrew or a ballscrew if the little MendelMax doesn’t
present me with any more severe issues. Though that would probably also warrant stiffening
up the frame significantly. And obviously, both 3D printing and CNC machining
are very heavily dependant on what software you use to prepare your jobs, and while CNC
obviously has a ton of options when it comes to high-end CAM packages, the equivalent of
a slicer, there’s not that much great software out there that is somewhat powerful and still
easy to use. Sure, Autodesk Fusion is an incredibly powerful
too, but the “free as long as you’re a tiny startup” and the learning curve involved
kept me looking for simpler options. I personally can program something like a
Haidenhain CNC controller directly and use a five-axis CAM like TEBIS, but neither of
them apply here, obviously. What i found to be a nice middleground was
the shareware-like ESTLcam, which is a fairly simple, 2D or 2.5D-centric cam that does just
about enough for me without having to relearn the entire tool whenever i want to machine
something. It also does picture engraving and 3-axis
3D milling, and while that’s good to have, it’s not a major selling point for me. What i want is just a simple, visual tool,
ESTLcam is just about ok enough for this. It even creates Marlin-compatible gcode out
of the box! So yeah, that’s the first part of my adventure
with turning a 3D printer into a mill. While it’s not perfect and was a lot of
work, it has most definitely turned out better than i expected. I mean, it mills aluminum, that’s a success
whichever way you look at it! I’ll definitely be making a few more mods
in future, but at this point, it’s already an awesome little conversion, i think. If you want to try something similar, i’ve
linked the spindle, power supply and milling bits i’m using in the description below. I hope you learned something in this video,
if you liked it, give it a thumbs up, consider subscribing to the channel, and because Youtube
is still being weird about it, remember to also click that bell next to the subscribe
button or you might end up missing some videos altogether. Also check out the affiliate links from the
video description to shop on Amazon, eBay, Matterhackers and iGo3D, those don’t cost
you a single penny extra, or if you want to support this channel with a spare dollar or
two, head over to Patreon and get access to monthly Q&A hangouts and more.

100 thoughts on “CNC mill built from a 3D Printer!

  1. I was toying with the idea of putting together a self build 3D printer and converting it so it could to light CNC milling as well. I even identified the newer Proxxon hobby tool as a possible spindle.Comparisons all seem to rate it as far more durable than the Dremel.
    In the end though I guessed that the build would always be too compromised by starting with a printer. It seems like you're going to spend quite a bit in combined time and outlay to beef the printer up in order to get a very lightweight mill that may still shorten the axis motors' lives significantly. You're already starting to look like Theseus.

    I think the only satisfactory way to go, at least if you only want 1 machine in the room, is to save up for one of those Chinese $1,500-$2,000 5 axis CNCs and build a print head and heated bed to fit it.
    That outlay typically gives you:
    – 1 to 2kw water cooled spindle;
    – 5 axis CNC milling (with the software by the looks. Pirate?) and wide range of compatible end mills available;
    – a tank with some of them, if you need to cut while using coolant;
    – a CNC router that takes a useful range of bits;
    – an engraver;
    – possibly even a 3D part scanner by attaching a measuring tool;
    – and finally a 3D printer as well.

    Repeatable placement of tool placement is typically quoted as <20 microns.
    It seems so obvious that you should be able to have all in one machine, but I think the CNC spec' defines the build and then it won't even notice when it's carrying a printer on its shoulders or routing soft materials.

    The only thing I profoundly lack is any kind of justification that has a hope in hell of passing the wife test :'-(

  2. I realize that I'm a month behind, and you may have already settled on a solution to the belt question, but maybe going from 2GT to 3GT belts might take care of your issues in that area? By all accounts there is a marked strength improvement from 2nd gen to 3rd and since they share tooth profiles it'd be maybe $35 for the whole thing.

  3. Hi Thomas, really enjoyed your video after finding you recently. Have you pondered converting medium-duty CNC mill to 3D print? I picked up one for about $700 USD that had a fairly ample working envelope of roughly 300x400mm, though currently the Z axis only has 80mm of travel. I paid a little extra for a fourth axis which shipped with a full-size 3-jaw chuck you might see mounted on a mini-lathe, very impressive. It's very sturdy with 12mm aluminum plate used throughout the entire machine, and ballscrews for low backlash. It's not for working steel but very competent and quite accurate for everything else (Uses 16X microstepping for .000625mm/step). Uses a 24V supply with what they call "57" stepper motors, very quiet and works great with an Arduino Uno running GRBL (Though GRBL doesn't support the fourth axis); it shipped with a 25-pin printer-port-style connector but since I didn't want to run the steppers from my PC I used GRBL to do the G-code work onboard. I use a free package called bCNC to drive it from my Linux/Ubuntu PC, which includes quite a few features for simple things like leveling a surface, engraving text and shapes etc., and provides profiles for materials and end mills/tooling, along with buffering and loading G-code files, interpreting DXF, and even ASCII STL. It even has support for using a USB camera to do measurements and homing, though I haven't gotten mine properly calibrated and mounted yet; I thought it was a very ambitious feature for such a simple, powerful package.
    Anyway, the mill was an attractive buy since it shipped completely assembled in two large, heavy wooden crates from a company in California (I'm in the US), and arrived in perfect condition due to some serious, steel tie-downs and wooden mounting braces. I got it on eBay, here's the item number for a similar mill (They're fairly generic and are re-sold by a number of importers): 262850371022. –What do you think? Would it work for 3D printing or is it too slow/limited? Given the prices for some of the nicer 3D printers it's not terribly expensive, so it seemed like a good possible match for mounting an extruder. I've tried mounting a laser for light-duty paper cutting and burning text/images into wood, which worked quite well (No visible stepping artifacts, even under magnification), though a bit slow (1000mm/min) since I wanted it to be accurate without a lot of slop.

  4. take a look at mach 4 machining software, it's reasonable, powerful and compatible with most motion boards. tons of plugging and a support forum.

  5. Please stop the music… Its so annoying and i wander off to browse the other videos on
    offer instead of listening to your commentry and watching the full video.
    Shame because you have some great video's. No thumbs down but no thumbs up either.

  6. This quickly turned into a, "my grandpa's axe" story. Only replaced the head twice, and the handle three times!

  7. Hey nice video. Just wondering what need to be changed in marlin firmware when using ESTL cam to do this conversion?

  8. Are you sure it was the bearings? I have something similar and I went through several sets of motor brushes but the bearings (2x 608) were fine.

  9. Den Messschieber zum Anreißen nehmen ts Das hätte mecker vom Meister gegeben.
    (Using the caliper to mark your measurement tsk This would've caused some remarks from the master craftsman.)

    Whatever floats your boat 😀

  10. I have this project on my bench for to long now..
    I realy love the ideia of making my pcbs, dont rly need to cut trough aluminum just a bit of copper will do the job
    https://www.instructables.com/id/3D-Printed-Desktop-CNC-mill/
    https://www.instructables.com/id/Make-a-CNC-mill-with-a-laser-cutter/

  11. My experience with the low cost Chinese spindles was terrible. Most have simple 608zz bearings in them. The runout on the one I picked up was so terrible that it was unusable. Switched to the Dewalt DW660 and it was much better (although it is larger).
    As for conversion, you might be better off building a dedicated cnc machine such as the Mostly Printed CNC or Root 2.1

  12. hi there, i'm trying to reuse a old cnc router to cut windows it's verry big, can you advice me some software and hardware? the board is veeery old, 1985, and i can not find enithing related with the fabricant of the machine, [email protected] .

  13. I have literally spent the last 2 weeks watching your videos….downloading most of them…Today…you literally made my day…Estlcam….a lifesaver for my DIY CNC project. The more you upload, the more we get the learn….especially those small hints you give out.
    From your videos i learned how to setup an arduino…which arduino to buy..how to configure Marlin…the list goes on…..Thank you
    One request…if you can literally show how to DIY build a filament extruder that would be awesome..i mean literally from scratch……here in Africa cost of shipping alone would exceed the price of the extruder or parts….Thanks again

  14. I wonder if you could reuse the plastic trash you produce with the 3D printer to make recycled plastic sheets that you would then mill on the converted 3D printer (with a hot press or something similar). Also, using the aluminium flakes in the recycled plastics would provide some fabulous glitter.

  15. hi, I am really having hard time to gcode that will operate my Marlin based 3D printer like a cnc… plz do an instruction video about the software, how to convert an STL file to Gcode that operated 3D printer like CNC. I tried Estlcam to create Gcode directly from that software, but that gcode wont work in my Marlin based 3D printer. You have done a wonderful job already, PLZ HELP ME, i am desperate 🙁

  16. maybe not as ambitious as a CNC (I have a 2.5HP 1950's Mill with 53" table, weights 2880lbs… that's what you need to mill metals. but perhaps you could use an attachment on your printer to convert it to a CNC Drill for making PCBs? Update: Have successfully made attachments for my WanBlo i3 to use FlatCAM and Geany to CNC Drill and Engrave (sharp springy dragging style) PCB's with through hole VIA's accurate to +/-0.25mm over a 14x14cm double sided PCB (after Y-axis tram adjust mod using long threaded rod). 3D printing pulleys and cycloidal gearboxes to convert my 1950's 2.5HP 3PH Cincinnati Knee-mill to CNC.

  17. I'm looking to do the exact opposite. I'm going to try to add a 3D print head to my Bridgeport CNC mill. I've always wanted to try 3D printing, but not sure if I have any real uses for it.

  18. you bought 2 proxxon engravers, tried routing on a 3d printer, and then call it "milling", but didn't notice the proper mini mill in their product line?

  19. I thought that I could convert a standard Prusa i3 with 8mm rods to a wood milling machine. I was wrong. But after building another reprap with 12mm rods I was successful on hanging there a 600W spindle that works surprisingly well. Cutting speeds not super fast but it works. And the best thing – the final total cost of such reprap cnc is about 200EUR. Of course it can be converted back to 3D printer 🙂

  20. is it possible to put the whole moving table in a bag of robust transparent plastic to prevent the particles flying around? Assuming you provide enough cooling.

  21. I would replace the squirt gun and vacuum hose with a pump/nozzle/coolant with lube/recovery tank with filter setup. At least for acrylic and metals.

  22. You need to match the material removal rate to the power of the spindle.

    The feed rates looked way, way to fast for the Proxxon's 90W.

  23. I have a Tevo Black Widow that I am very happy with. I'd like to retrofit it so I can add a CNC head to it. One thing I am curious about is that it is belt driven for the X and Y.. Not thread. When CNCing how much pressure would be put on the bit during normal operation and would the belt be sufficient to hold the bit in position.
    Many thanks

  24. What do you do after you put the motor and drimmel on to make it cut? Do you have to use a program? I want to do this on my 3d printer but I'm not sure how. Any help would be greatly appreciated.

  25. i wouldnt consider a startup with 100k revenue as "tiny" so fusion is still the best option imho… i mean if your business runs, you can easily afford buying a license for it anyways. so autodesk made the move that adobe should have done years ago with the creative suite 🙂

  26. I just wonder what frimware you use for cnc mill , Im kind of intersting to do the samething to one of my 3D printer, that would be great if you do a video on firmeware and loading file machine and tunnig

  27. Yes, you tube is being lame, especially if you are a free thinker, anti globalist, anti-U.N.AGENDA, a non liberal-socialist, anti-china, or basically not a democrat/lefty/new worlder-Cultist…..

  28. Hallo Thomas ein super Video wie immer. Mit der Proxxon hab ich das gleiche Problem gehabt müsste ich auch 2 kaufen. Das kleine micromotor mit separate Netzteil hab ich danach verwendet und hat super funktioniert zum gravieren und Holz zu Scheiden.
    Mach weiter so mit deine Video ?

  29. Very well done!!!
    The frame of a 3D printer was never designed to take the loading that would be present on a mill but you got it to work ~ congratulations!!!
    Another big issue is that the mass of the spindle motor is very high but I have seen someone (on his own design PCB mill) use the motor out of a RC car, then he used the motor out of a quad-copter, both of which he was able to do speed control using g-code.
    The only issue was that the bearings of these motors were not designed to handle the side loading that they were now being given.
    Last time I saw him he was building his own spindle which would absorb the loading and coupling it to the motor which would only have to provide the rotation.
    But he was driving using toothed belts and these had a very short life span.

  30. The people at MTW are probably crying after seeing this. Who thinks this chassis has the rigidity to support ANY form of milling/routing? What a waste and complete loss of credibility! Unsubscribed!

  31. I have a Anet A8 3D printer and would like to turn it into a CNC mill I have done all the upgrades/mods belt tensioners, braces, and marlin firmware. I have a plan and would like some advice. I have a Black & Decker RTX and a 24" flexible shaft. I plan to mount/hang the router and print a mounting bracket for the shaft this will reduce the weight on the carriage, then I can replace all the 3D printed parts with metal. what I would like to know is do I have to reflash the firmware or can Marlin do both. I also plan on getting a laser and would like to be able to swap between the 3 print head ( is that the correct term for extruder, mill and laser?) Second question I use cura for slicing the videos and what I've read other people use other programs to generate gcode and save it as .nc file not .gcode. I like raspberry pi and octoprint could I use that? 🙂 I love the show and you have helped so much already 🙂

  32. $1100 for the printer $100 conversion price. $1200 ish. You can get some really nice desktop CNC for that price.

  33. We were impressed by the video, looking forward to the next. We are a small 3D modeling software startup completely bootstrapped and tooks us 5 years to develop our real-time collaborative CAD.? Would be great if our channels could collaborate.

  34. Hi Tom can I convert my Tevo Tarantula in the same way, I have already converted it to laser engraving, with your help, many thanks

  35. an extractor next to the nozzle reduce hoover pipe with smaller pipe 10mm you can 3d print an adaptor from 30mm to 10 🙂

  36. Thanks a ton for this information! Would love to hear about any future developments you do with respect to building a machine like this.

  37. Does the power supply just drive the spindle? What about the electronics? They usually take 12/24v, so I wonder if you are running two power supplies or stepping down the current for the controller board?

  38. Your feeds look crazy for milling, no wonder you wrecked the bearings. Any cutter will have a maximum amount of material it can cut away in one revolution, if you exceed that it causes tearing instead of cutting. That's where all the extra stress was coming from.

  39. Any idea how to get around the fact that the stock top becomes Z0 when you place your bit on it and the printer won't move into -Z space?

  40. Even getting it to cut aluminium is an achievement I have a bit of milling and turning experience and even multi ton machines aren't truly solid take to big or to fast a cut and get a lot of chatter and vibration

  41. Rather than wd40 give paraffin a try for lubing aluminium one of the best cheap readily available lubes I've ever used on ali

Leave a Reply

Your email address will not be published. Required fields are marked *