I decided I want a CNC Router - I would really like something like the X-carve, but my budget doesn’t stretch to it. My budget does stretch to smaller devices, but all of them look like they may have pitfalls built in (for example some of the Chinese 3 or 4 axis CNC routers on ebay look like they are fairly good mechanically, but need to be driven by a Windows XP PC with a parallel port - and the USB/Win10 compatible counterparts are above my price ceiling).
There are devices that have all of the technical features, and are within my budget, but they are typically small - perhaps too small to be useful.
So… I reckon I should build my own - starting with a packaged set containing stepper motors, controller board and software (consuming about half my budget) - and I build the table, linear drives, etc. I’m not interested in getting dirty with Arduino or building stepper motor control boards myself, but there are kits that appear to have the full set of technical bits that I need.
How hard can it be? (That’s an earnest question) I don’t think the mechanical build of a 3-axis tracked system (with leadscrews to drive it) is beyond my skillset.
Has anyone here built anything like that? Care to share advice or experiences?
I’m on my phone right now and can’t find the link, but I know Sparkfun sells a DIY CNC kit that has gotten good reviews. Not sure about the price, but it may be worth checking out.
Jumping in to say my son has built two CNC machines from plans and parts he found on the internet. He built the first one mainly from plywood and MDF. The second, larger machine is more metal. He now makes a small income selling custom items made on the machine. He is mechanically inclined.
Yeah- that’s what I’m finding - a lot of these devices seem to have started out life as ‘budget’ versions, then morphed into the mid-range cost bracket.
I guess maybe I should check if there’s a nearly-mature kickstarter project that I can jump onto.
We have 2 CNC routers my spouse built. I should add he had a prior professional as a professional engineer designing assembling robots and assembly lines.
My advice to you is:
Do a LOT of research on any system you are interested in purchasing and building. Then do more research.
You might be able to get a better system from someone who purchased one then either lost interest in the project, or found it beyond his/her skill set to build. Of course, there are also pitfalls with doing that, but it might be a way to get more for your money.
It will almost certainly take longer and cost more to assemble than you initially think, which is typical of most major build projects. You also have to include costs of peripherals - for our largest router we also purchased the parts for and built a heavy-duty dust collection system. You WILL want a dust collection system. You also might want a good dusk mask/respirator with dust filters, too, and I also recommend hearing protection, they are noisy as hell.
If your system is not designed with a sudden stop Big Red Emergency Button attach one to the system. When stuff goes wrong it can go VERY wrong. You want the option for a sudden shut down from a safety perspective. A fire extinguisher might also be a good investment. Consider it industrial equipment, and take precautions so you don’t have an industrial accident. From our on-line fellow CNC routers folks I know of at least one case where a body part was permanently removed in an accident.
3 and 4 are examples of why these projects easily wind up costing more than you initially thought.
Vibration and registration can be issues. Speed of the bits can be an issue - we were doing plastic signs for awhile and we had some issues with the plastic melting due to friction which does the machine no good. Wood doesn’t melt, but it can burn. Be prepared for experimenting and testing. Again, this is another reason costs go up.
It’s important you’re comfortable with the software you are using. It’s important the software can talk to the machine. From what I’ve seen a lot of folks put a lot of time and effort into researching the hardware but not so much the software. Both are equally important
Have fun. Try not to get too frustrated (it will happen). When you get it to work it can be magical.
I can’t speak to the mechanical part of the home-build, but here’s some thoughts on the computer side.
Do you have a beat up old PC laying around somewhere? The reason is hinted at by your post (the cheaper interfaces needing XP and parallel port): CNC machines really run better when they have a dedicated PC in the “dirty” side of the shop, running the G-code, and not stopping to pop up a dialog asking if you want to install the latest Windows updates.
If you can scrounge up such a machine, either set it up with Windows XP or your favorite Linux distro. Then you can use Mach3 ($) on Windows or LinuxCNC (free) on Linux.
Then download Fusion 360 and have fun! They give a free 1-year license to their excellent product, and once the year is up you can continue on for free with a hobbyist license as long as you have exceeded a certain generous threshold in profit (e.g. 100K?).
Fusion 360 is particularly cool because it is a pro-level product with both CAD (designing your part) as well as CAM (calculating tool paths and generating G-code). You then put the generated G-code on a thumb drive and load it on your junker PC via sneaker net.
I’ll be watching here to see how things go for you. I’m also interested in what kinds of projects folks are doing with their CNC setups.
I’m currently pondering my own home shop. In a former life I was a machinist, working on Bridgeport mills and the like. I might be able to purchase a manual milling machine this year, and I am reading everything I can about CNC conversions.
Thanks for that - the startup/hobbyist licence for Fusion 360 is really cool.
I went to a hardware store today and I saw some 40mm box-section alloy tube priced at £10 per 2.5 metre length - I’m fairly sure I can make the frame and mechanical parts at a vastly cheaper price than any kit or prebuilt CNC router set I have seen
I’ve collected a bunch of parts to make a big X-Y-Z table for my plasma cutter.
I’ve never had the time to finish it, and some parts will need to be fabricated on a large milling machine. You might want to check out these guys: http://www.techno-isel.com
The biggest issues for a DIY machine is getting decent slides, making sure everything is square, and finding a low-backlash driving solution.
I was so worried about these issues that I way overdesigned my table, with precision zero-backlash ball screws (actually, not that expensive), and very, very expensive low-runout bearings.
I’m thinking that perhaps I should blow about half my budget on a smaller kit and use this to run some of my smaller project ideas, build my CAD skills and meanwhile, continue saving with a view to building a larger model (including manufacture of some of the components on the smaller CNC machine.
I’m wondering about in the cybersecurity aspects. From a security perspective, I like the idea of an isolated computer running the CNC tool, but you have to keep the computer updated, and load data files on it - both sources of security weaknesses. How do people who do CNC work typically secure their software and data files? What do people typically do to prevent an evil hacker from modifying the design or tool path files? Bit of a hijack, but I am curious.
If you have a standalone machine to control a piece of equipment, and you’re the only user, security is more or less a wash - it doesn’t get online updates for the OS or antivirus, but it’s insulated from most threats because it’s not online.
It’s exposed to threats such as infected USB media, but if those are coming from protected machines, then the standalone machine gains security by proxy.
The biggest risk would probably be if customers send in media and it is inserted directly into the standalone machine, but again, because the machine is off the network, the risk is contained.
That’s not as much protection as you hope. If bad guys start targeting CNCs, or their data files, or the cloud-based CAD/CAM/CAE software, builders could be in a bad way. the I’ll have to think on suggestions for improvements.
We have a computer for the CNC equipment that is NEVER permitted to connect on line. NEVER. If it’s never on line it can never catch anything on line.
We do not use software residing in the cloud, nor do we use such datafiles. EVERYTHING is on the stand-alone machine. Which NEVER goes on line.
Anything going from anywhere else to the stand-alone goes through a thoroughly scrubbed and protected intermediary first, is downloaded to a USB, then inserted into the stand alone.
Customers NEVER have contact with the stand alone.
When the machine is not in use the computer is turned off.
I’m having trouble imagining how the “bad guys” could target our CNC unless they physically broke into the place where it resides. They’d still need passwords to get into the computer.
I’m not sure why you think the CNC computer needs updates - well, I suppose if you change the software you’re using but you shouldn’t need to do that. The code running our computer-controlled lathe hasn’t changed since the 1990’s, for example.
Heck, most of our design work is done on the stand alone - we almost never have to transfer anything to it. It really is physically isolated. Computers don’t actually need to be networked, wi-fi’d, or on the internet to function.
Here’s a cool video from How It’s Made showing the manufacture of player piano rolls. Perhaps the neatest thing of the video is not the buggy-whip-era product, but their own CNC tool.
At about 1:00 you can see a musician recording a ragtime piece using a MIDI keyboard and an older iMac. Then at 1:14 you watch as a worker inserts a 5.25" floppy disk into the stone-age Apple II computer driving their machinery.
I’m fairly sure it’s as much protection as is necessary.
Why would they? Where’s the profit for them?
The biggest real-world risk associated with standalone machines is that they tend to get used way past the point where they are obsolete - then they break down, and the breakdown, although fairly minor, is actually catastrophic in effect, because the world has moved a long way on and repair or partial replacement isn’t an option.
I think I’m going to chew on the idea for a while. I know I have the tools and skill to make my own device, and my eagerness to have one to use is clouding my judgement. Time to cool off.
Of course, I can spend the interim thinking about how I would build it if and when I do build it - so here’s my outline plan:
Static bed, moving XY gantry, with mounted Z axis.
Build the Z axis first, then build an X axis gantry on a Y axis carriage that’s capable of accommodating it - so many projects I’ve seen seem to finish with the Z axis, but it seems sensible to start with that - as any unforeseen problems with size etc can be engineered around in the rest of the build.
I could use metric stainless threaded rod for the leadscrews - M10 probably. I can easily make anti-backlash nuts using two M10 nuts, some small pieces of ply, a spring and some small containing bolts. The pitch of metric threads is quite shallow, so this means the leadscrews could be driven directly by the motors without compromising resolution.
The runners/rails are probably the biggest unknown. I can manufacture HDPE wheels that would run in channels or on rails with quite low friction, or I could make or buy bearings that run directly on cylindrical rods. Not sure about this one.
I keep seeing the DIY Smart Saw advertised on fb and other sites. The site claims it can be built for $160 if you have access to scavenged and used parts or $500 if you go with new parts.
I know a lot of stuff that’s advertised on fb is of the X10 camera caliber, but I can’t see any obvious scaminess to this one.
I’ve also been thinking about buying a hobbyist CNC, but I don’t know where to begin. I just looked at the X-Carve referenced in OP, which would suit my needs, but are there competing products? I’m a hobbyist woodworker, I was going to upgrade my cabinet saw to a SawStop (~$3.5k) this year but I think I may opt for additional capability rather than an upgrade of something I can already acomplish. What are some options for those that would like to buy a complete package? I’ve seen systems at Rockler and Woodcraft but they are $$$ and not as affordable as the X-Carve seems to be.
Shapeoko is probably the most obvious alternative to xcarve, but to be honest, xcarve has made cnc much more accessible because of their Easel app, which simplifies the design and control processes