3D printers used to be finicky, rickety hobby kits that required a second hobby just to keep them running. My first machine, a RepRap Darwin derived RepMan, arrived as a box of laser-cut acrylic pieces and threaded rods. The only instructions were a PDF of 3D renders. I had to build the hot-end by hand, measuring the resistance of nichrome wire, wrapping it around a brass tube with Kapton tape, and encasing it in barbecue cement. We sliced files using an early, parameter-dense abomination called Slic3r, and printed exclusively in ABS simply because HDPE was worse.
Today, the landscape is unrecognizable. Modern machines boast heated chambers, textured plates with automated mesh leveling, dynamic motion control, and filament run-out sensors. Software like Cura and PrusaSlicer handles the heavy lifting with smart profiles and auto-generated supports. As a colleague recently noted: she wanted to make 3D objects, not take on 3D printing as a lifestyle.
This shift from "enthusiast project" to "reliable appliance" has rippled across the maker market. Laser cutters like Glowforge and xTool, driven by cheaper, more powerful diode lasers, have become highly accessible. UV printers are moving the same way. But one category has stubbornly resisted this plug-and-play revolution: the desktop CNC.
Making a user-friendly CNC machine is an order of magnitude harder than building a 3D printer. In 3D printing, the print head experiences almost zero load. You can rely on simple belt-driven extruders gliding on linear rods.
In CNC machining, the cutting bit is in constant, forceful contact with the material. If you use a standard belt-driven system to carve into anything harder than foam, the cutting tool deflects, the belts slip, and the part is ruined. This physics problem compounds rapidly with harder materials. Screw-driven systems might handle wood, but cutting non-ferrous metals like brass and aluminum—let alone steel—requires immense structural rigidity.
Until recently, consumer-grade CNCs (like the Genmitsu kits) were essentially engraving machines. They were fine for wood and capable of light surface passes on aluminum, but lacked the rigidity for full-depth metal cutting or true dimensional accuracy. Anything that could manage this was expensive and heavy with the userfriendlyness of a rabid badger. Although that has begun to change
A few years ago, Makera released the Carvera a fully enclosed, highly accurate desktop CNC capable of cutting metal. It featured an automatic tool changer, depth probing, and a 4th-axis module. But at $6,000 and 50kg, it was firmly in the professional workshop tier. They followed up with the Carvera Air, stripping the automatic tool changer to bring the price down to $2,500 and the weight to 30kg, but it was still a hefty investment.
Then a year ago they announced the Makera Z1.
Designed as an entry-level machine, the Z1 promised a footprint closer to a standard 3D printer (weighing 17kg) while maintaining the rigidity to cut brass and aluminum with 0.02mm precision and with a usable work area of 200*200*100mm in 3 axis mode. I jumped on the Kickstarter for the Z1 which had a base price of $899 although I picked the upgraded Z1 pro for a bit more money featuring closed-loop stepper motors to prevent lost steps, and ball screws instead of standard Acme lead screws for better rigidity and longevity. I also picked up the cyclone dust collector and the 4th-axis rotational module.
Fast forward to August 2026. A massive, awkward box arrived via UPS just as I was heading out to PAX West. It sat in my living room, taunting me, until I returned to find out if Makera had finally built the "Bambu Lab of CNCs."
The machine relies on Makera Studio, a new app that consolidates their older CAM and Control software. It includes automated modes for common operations and custom modes for advanced users as well as predictably for the current hell times some AI bullshit to make models for you. Makera is also attempting to build a Thingiverse-style repository where users can download and run CNC projects from them or other users with a single click.
Getting the machine online, however, was a headache. The Wi-Fi connectivity was finicky, struggling with special characters in my SSID. Even after a firmware update, it took significant fiddling to establish a stable wireless connection to my PC. For a machine aiming for the seamless "Bambu Lab" experience, network setup needs to be flawless.
For my first project, given I had just return from pax and had gotten some free dice, I designed a simple wood dice tray in FreeCAD. It featured a main pocket, rounded corners, and recessed holes for magnets to attach a matching lid. I decided to machine the thing out of some poplar I had lying about which is technically a hard wood although on the softer end of that spectrum.
Moving the model into Makera Studio immediately highlighted the inherent differences between 3D printing and CNC machining. 3D printing is additive and forgiving. CNC machining involves spinning a razor-sharp piece of metal at thousands of RPM and slamming it into solid material. Get the feeds and speeds wrong, and you will snap the bit, ruin the part, or damage the machine maybe even damage yourself, possibly all of the above. Furthermore, unlike a 3D printer bed, a CNC workpiece has to be securely clamped down or else it could break free and cause damage death dismemberment etc so is usually secured with clamps. The software doesn't know where those clamps are, so you have to manually ensure the tool path avoids them. For the feeds and speeds Makera provide a library of suggested values for different material types for all the many bits they provide for the machine. For anything you use that isn't in their database or isn't quite the same as they are expecting you would need to manually select these variables.
I skipped Makera's automatic 3-axis mode as the simulated paths looked risky and opted for manual setup. Here I hit a frustrating undocumented quirk: Makera Studio accepts standard STL files much like any 3d printer slicing software, but you cannot select specific faces to generate toolpaths on an STL. You need a .STEP file to do that. A quick trip back to FreeCAD to reexport the file solved the issue, but a simple software tooltip could have saved me a lot of googling.
I set up a ramping spiral cut for the magnet holes, a standard pocket clearing for the tray, and a contour cut with auto-generated tabs to free the final part. I did a quick simulation to see if things looked good then loaded the paths into the machine over wifi. After running the built-in laser probe to sweep the bounds of the cut and verify it would clear my clamps, I kicked off the job.
The cutting process was impressively smooth. Swapping to the 1/8-inch single-flute end mill was effortless thanks to the quick-release lever, and the machine automatically probed the tool length on a built-in bed sensor. The machine was surprisingly quiet; over the sound of the vacuum, you could barely hear the bit cutting the poplar.
Near the end of the 19mm deep contour cut, I realized my 12mm-long bit was cutting it uncomfortably close. The plastic collar on the spindle looked like it was about to bottom out against the wood. I hit the e-stop just to be safe and finished the last sliver with a flush-cut saw. The final part looked fantastic—accurate dimensions, perfect magnet fit, and only minor tool marks in the pocket that sanded out.
While the machine impressed me, the official Makera dust extractor did not. It hooks up via two hoses: one blows air to clear chips from the bit, and the other is supposed to suck them down into a collection bin beneath the dust collector machine. The blower worked fine, but the extraction was abysmal. The CNC's interior was left completely caked in wood chips and dust. It works fine when you disconnect it and use it manually as a shop vac, but as an automated, integrated system, it fails completely.
So Is it the Bambu Lab of CNC?
Not quite and it may never be, largely due to the nature of the technology.
Makera has built an incredibly capable, rigid, and accessible machine at a groundbreaking price point (now $1,099 for the base Z1 and $1,499 for the Pro). For the hardware alone, it is a triumph.
But the "click-and-print" dream falls apart on the software and community side. You can't just download a CNC file and run it the way you do an STL. Even if you have the exact same machine as the creator, you need the exact same cutting bits, identical material hardness, and the exact same clamp placement. There are simply too many destructive variables for a true "click and forget" experience. CNC machining fundamentally requires the user to understand what the toolpaths are doing.
Makera Studio still needs bug fixes and UX improvements, and the dust collector isn't worth the money. But if you are willing to learn the basics of CAM and respect the physics of subtractive manufacturing, the Z1 Pro is a massive leap forward for desktop fabrication. Is it for everyone, no it's a much more niche tool than a 3d printer for someone who likes to make things and has always wanted to be able to make small metal parts its perfect and there is almost nothing like it at this price point. That may change as I've seen several other "bambu labs of CNC" projects popping up so competition may heat up and promote more innovation in the field.
As for me so far I'm pretty pleased with the machine, a little let down by the dust collector, and hopeful the software improves.
Comments
I have a big CNC router at work, 12ft x 6ft bed. A number of issues you have highlighted do not change as you go industrial.
There is no good answer to feeds and speeds, there are calculations but you always have to adapt to the machine and material. One thing you may not already be aware of is that you need to produce chips not dust (talking wood, I haven't done metal.) Chips remove heat from the router bit and prolong the life of the bit, going slow and making dust actually overheats the bit and reduces lifespan. So you are always trying to find the sweet spot, going as fast as the machine is capable, whilst maintaining a good cut edge and not breaking the bit. There are no sensors in the big machines either if you mess up speeds etc they will tear through your material, smash bits, tear themselves apart and tear you apart. They are unforgiving! We have destroyed two spindles on the old machine, and gouged a big hole in the bed.
Dust extraction is never as good as you would like. We have a separate hoover that lives with the CNC for post cut clear up. Also, two bits of ply or MDF to just scoop up the mess.
We use VcarvePro for creating the files, I like it. It has a relatively simple CAD interface.
We started with a relatively cheap used Chinese machine with UK brand/seller. It was somewhat unreliable, as we gained experience that improved to an extent. We asked other compaines with much more expensive machine if they got more reliable if you paid more. The advise was if you want a working CNC, buy two. After a day's problem free cutting you can turn it off and the next morning it has failed in some catastrophic way with only 12h of entropy to explain it!
In summary, I don't think CNC routing is ever going to achieve "click-and-print" except in perhaps some highly ring fenced environment...
(BTW I always read your tech dives with interest, cheers)
Enjoyed that read too, EMW! Amazing to think how far they have come. We used the CNC for our 3rd year project robot but I handed the design and measurements and the engineering dude did all the work. They kept it completely away from anyone.
Have you got a project in mind that you would like to do?
I've always been interested in Cuburo marble runs but they have always been fantastically expensive. Felix reckons you could 3D print the bricks with a bit of pliers action. I wonder if your CNC would do it?
I'm very much in the "reliable appliance" demographic, with zero engineering or 3D modelling background, but a variety of expensive hobbies with physical components.
I remember looking at the more engraver-style laser desktop CNC machines a while ago, which very much over-promised (I think I was looking at wargaming scenery manufacture at the time). 3D printers were, at that point, flaky and complex, and I thought "I could simply laser-cut some wooden parts". I quickly understood that it wasn't that simple. Having read the above from you and Fish, I now realise it's a bloody nightmare!
I think it's time one of us said "how hard can it be?". So there we go.
Yeah it's a lot to learn in terms of new ways to think about how arrange the operations and what parameters to use. Feel like even having done one easy project there are a bunch of things I would do differently. The chips part is something I was vaguely aware of but good know I think I also used metal bits on the first go with wood which I knew was not really ideal. I picked up a few other ones since then which had actual wood cutting bits so I can use bits actually designed wood when cutting it
The dust collector part annoys me because I feel like I could have just bought a cheap shop vac and it would have worked as good if not better and it wasn't a cheap add on. Maybe the community or the people at makera will come up with some improvements but so far it seems like it barely gets anything when its running.
I've seen vcarve on a bunch of peoples cnc guides although it's probably a bit expensive for my hobby use case.
I've not had any really specific project in mind for it more I've always wanted to be able to machine metal to make more sturdy sorts of parts and make the sorts of things that can't be done on a laser or a 3d printer and this machine came along at the right price to performance point that I could justify adding it to my collection of toys.
I think you probably could make those marble run blocks to some degree the more complex ones might need multiple cuts where you reposition or flip them over to allow the tool to access the part to make the holes/paths which can get tricky if a previous cut means it can't sit flat or it's hard to align it such that your cut is in the right place.
That sort of complex part like that is actually a good task for the 4th axis you can have it index the part rotating it to do each side and then cut in the other three axis as one operation with it locked in a stable rotating axis. Although for the Z1 the 4th axis working area is pretty small 80mm(Diameter) * 150mm(Length).
I feel like CNC is a hobby in and of itself with a lot of pitfalls and cost beyond the machine itself (decent milling bits are pretty pricey you can get the cheap chinese ones but they'll snap or wear down in no time) where 3d printing is now almost at a commodity level (I did see 3d printer filament in Target the other day and I'm pretty sure Walmart also stock it now) given the extra level of challenge CNC will likely never cross over that boundary.
Plus given the costs involved if you only want a part every now and then made in metal or some such a lot of the chinese pcb companies also offer either casting or 3d metal printing or cnc from a model for pretty reasonable prices and quickish timeframes
3D printing is definitely commodity level. Felix has printed and design a shit load of stuff without any input from me. While Felix is definitely up the brighter end of 16 year olds, he's still not a hard-bitten-reprap kind of hobbyist. It's great to see him turning his Nerf blasters into WH40k weapons. The combination of three hobbies right there.
I enjoyed this video on the Z1.
I gave the 4th axis module a go. Adds a good deal more complexity to the tooling paths which caused me some troubles. You need to think about clearances and tool angle in all 4 dimensions which is tough. If part of the model sticks out it can shadow another part so the rotary cutting can't get to it also a hazard for the tool to crash into on some other pass.
Out of two attempts got one reasonable model out if it at the end at the cost of two blocks of epoxy tooling material and one 30 degree v carve bit that shattered when it went sideways into the model at the base when doing a 3d relief cut to try and get to some of the bits the rotary cut couldn't get to. If it had been a bit with more of a cutting surface on it's edges like an end mill or something it might have been fine but the vcarve engraving bits are mostly designed for plunging straight into the material and really only cut on the V of the tip so it just shattered when too much of it caught on the material.
Still I learned a good deal about how to cut things using the module and what sort of things work well so the lesson was useful.
I also remembered this time to see what the timelapse looked like., Got a great view of top of the 4th axis stepper motor getting covered in epoxy chips and not much else.
Got a video of it in action, EMW?