Hey, maybe I guessed right! 
Olive is some beautiful wood, so far I haven’t found a local source and it’s always nice to be able to see for yourself what you’re buying.
I’d advise building a 60x60cm, just to figure it out and get a hang of it. You’ll soon figure out if its enough or if you want to make a bigger and improved one, from what you’ve learned.
1 Like
First, just be clear since I did not understand it when I built my first MPCNC, Z working height must contain the material, the bit, and clearance for the clamps, and perhaps some extra for double-sided milling. So 80mm of working height will allow you to mill things in all situations to around 28mm in height, and 100mm will allow you to mill things up to around 38mm of height. There are games you can play to mill things somewhat taller than the values I outline. Note that the MPCNC is very sensitive to Z height. The taller you build it, the more leverage on the Z axis, and the more deflection you could have. This will impact your accuracy and/or speed. So keep it short if you can, and if you cannot, research ways to mitigate the height. And there are some ways such as replaceable legs, multiple height spoil boards, and drop tables.
As for the XY size of 120x60cm, I’ve seen builds mentioned on this forum of this size. I don’t have any figures for support, but the general thinking by experienced people on this forum is that Ryan’s other design, the LowRider, is better for machines over 1000cm in size. Note if plywood is only an occasional thing to cut, it is possible to index the material using pins and cut larger pieces in multiple cuts. It is a pain, and there is a risk of not getting things indexed perfectly, but it is doable.
As for relief cutting like the bowl, this machine will do it, but cut times might be very long. Times will depend on the bowl size, type of wood, and how fussy you are about the surface finish as it comes off the machine. Comparing the MPCNC to other (and usually far more expensive) machines in its “class,” the MPCNC can accomplish the same things, but the depth of the cut is often significantly shallower. This means for a relief carving like this bowl, many more passes are necessary to get down to the full depth of the carve. And Z height of the machine mentioned above will impact the depth of the cut and/or the speed of the cut.
All the materials you list are doable. The recipe to getting good cuts from each will change depending on the material, and research plus trial will be necessary to figure out the recipe. Since the MPCNC can be built in so many different sizes, recipes can be very different between the machines.
1 Like
I also thought about that, but then again I don’t want 2 MPCNC of different size (would be cool if I had enough space). I could also just upgrade the size of the rods and that would be probably totally fine if I wouldn’t use an enclosure. But to avoid unecessary dust and noise I plan to built an enclosure before doing my first cuts on the MPCNC. I don’t know if someone ever thought of some kind of “modular” enclosure or that is adaptable in size 
Since I know all too well from experience that usually a short time after I do something I come up with ‘I couldas’ and ‘I shouldas’, I’m thinking I might make a temporary enclosure by hot gluing insulation foam panels together to make sure I like it before going on to wood.
Lol, I thought 100mm mean I can mill wood that is 100mm thick. Then I probably would need a z axis of 200mm which is to instable.
I never thought about indexing using pins, but that could do the trick. But I am always like “the bigger the better” and if it works once I don’t want to touch it again to adjust size 
Can you tell me whats “very long”? I saw a guy cutting a 30x15x8cm bowl in about 3h I think, but that wasn’t a MPCNC. Do you think a MPCNC takes even longer?
I think they tell you the z-axis shouldn’t be much larger than 80mm or 100mm, so basically then I can only cut 28-38mm thick wood? Is this in general so, that the z working space is so weird defined?
I am also thinking about a Stepcraft D840, which has a working space of:
- Working Space (X,Y,Z): 600 x 840 x 140 mm
Are those real 140mm, i.e. I can mill wood 140mm thick or do I have to subtract the bit etc.?
How would you compare the milling speed between this Stepcraft CNC and the MPCNC, similar size and milling motor?
As mentioned a bigger size is always nice to have, because I can probably do things with it I haven’t thought of yet. But at the moment I would probably need 120x60 cm for plywood like wood with a thickness up to 3cm and probably up to 30x30 oder 40x40cm for solid wood up to 8-12cm.
Perhaps it makes sense to buy 2 cnc one after another, because the purposes are different? But it would still be cool if one could do both 
Probably I can cut much faster and precise e.g. on a Stepcraft D420 or D600 than on a 120x60cm CNC anyway?
The MPCNC will take far longer to carve the bowl (assuming you’d built it to handle something 8CM high to handle the job). It is hard to quantify how long since there are so many variables: bit choice, stepover, types of tool paths, quality of the CAM/CAD tools, your skill with CAM/CAD tools, type of wood, amount of extra stock beyond 8cm, how finished you want the piece coming out of the CNC, etc. The 8cm height of the bowl makes me think that the bowl you mention was carved with a higher class of machine. Right off hand, I cannot think of any hobbyist level machines (under $2500) that (unmodified) can carve something so tall. Sometimes hobbyist machines can have an opening in the spoil board to handle larger things, but that does not modify the maximum cutting height.
z-axis shouldn’t be much larger than 80mm or 100mm
As mentioned, there are ways to design around this limit. A drop table (opening in the middle), swapping out leg sizes for doing larger work, multiple thickness of spoil boards to bring shorter material closer to the core, adjustable legs…
Are those real 140mm, i.e. I can mill wood 140mm thick or do I have to subtract the bit etc.?
Taking a quick look, I do believe you need to subtract the bit and stuff. This would give you around 60mm of material height. Note that there are games that can be played in some situations. As just one example, say you re only cutting 1/2 of the way into the work piece, and you clamp from the sides, and are not creating pegs for two sided milling, and you had a bit of just the right length, you could mill something 90mm high in the 140mm of working space. But that is a lot of conditions.
I am also thinking about a Stepcraft D840
I’ve only used my MPCNC so information other machines comes second hand, but I’m not impressed with the Stepcraft D-series machines compared to the several other machines in this hobbyist category…especially for the cost.
solid wood up to 8-12cm.
This thickness will be a killer with most hobbyist class machines. Again you might be able to design around the issue for certain kinds of projects.
Probably I can cut much faster and precise e.g. on a Stepcraft D420 or D600 than on a 120x60cm CNC anyway?
I don’t have hard numbers, but I would guess that the MPCNC has a similar precision to most other hobbyist machines. Speed is another matter, and not necessarily a simple one to figure out. Machines that are stiffer and have stronger steppers to go along with that stiffness can push the end mill through the material faster and/or deeper without suffering deflection. On the other hand, machines that have lead screws for X and Y like I believe the Stepcraft machines do, are notably slower on rapids (fast non-cutting moves). Fast is also defined by the size of the end mill. A 1/4" end mill will remove more material in a pass than a 1/8 end mill. The entry level spindle for the Stepcraft machines is a Dremel 4000, which only takes 1/8" mills (or probably 3mm if you are buying in Europe).
That sounds bad, but probably you are right
Maybe I start then with 2-4cm thick wood and buy a better one later on.
One day I also want to be able to do such stuff:
www.cnc-aus-holz.at/index.php?thread/341-schale-in-freiform/
But I think that is not really possible with the MPCNC or it would need 24h?
I think I could be happy with 6cm height, but not entirely sure if it is enough.
By the way, a larger z-axis makes it more instable, but is that true anyway?
E.g. is it also instable if I cut 8mm plywood or just if I need about the full height of 6cm? (Most of those 6cm carvings would need be worked on from both sides, if that matters)
I planned to go either for the recommended Dewalt or “China” spindle here. The latter should also be noticeable less loud?
To do the kinds of things that are shown at your link with what I consider a hobbyist class machine, you need to get creative. Here is someone who carves larger things using a Shapeoko. I consider the Shapeoko in the same class as the MPCNC, but it is stiffer so can be run faster/deeper.
And here he uses the same opening in the table but using a different jig to carve a large globe.
By the way, a larger z-axis makes it more instable, but is that true anyway?
E.g. is it also instable if I cut 8mm plywood or just if I need about the full height of 6cm? (Most of those 6cm carvings would need be worked on from both sides, if that matters)
For the MPCNC, the issue is the length of the Z axis below the core. The longer the Z axis below the core, the more leverage it puts on the core, the less force is needed to deflect the bit. So cutting the top of a 40mm piece puts less leverage on Z axis than cutting 8mm plywood. That is why I have two spoil boards…on short and one tall. When cutting thin material, I bring it up to shorten the Z axis when cutting. I have a taller machine since sometimes I carve 53mm builders foam.
The “China” spindle is quieter, but someone on this forum who has had both a low-end spindle and later a router (I think they had a Maikita) claim the router is significantly more capable.
Edit: Here is a spindle discussion.
1 Like
That’s some neat stuff! Thanks for posting this. 
- The reason a 1200mm MPCNC with supports is floppier than two 600mm ones is that the gantry has no supports. It’s not just sag, the gantry tubing needs to hold the X and Y position tight and if you have a 1200mm work area, the tube is going to be more like 1500mm and that is going to make it hard to keep it’s position. You say you are ok with 1mm error, ok. But if the bit wanders that much, it will be a very uncontrolled cut and the bit will either bite nothing, or take a huge chip. Not good.
- There have been lots of big builds and they work ok. But you need to be patient and set your expectations. Some people are very happy with them. Doesn’t bother me that they are happy.
- The best buts are about 20mm long. Carving a deep bowl is an advanced expert level job.
- You can make the machine longer with new tubes and maybe new wires.
I would suggest building it near the recommended size first. And you will learn more in the first week than reading 1000 forum posts.
1 Like
Maybe I should really start with a 60x60x10cm MPCNC. Maybe design the enclosure that longer plywood sheets would also fit and do that indexing pin stuff if I really need it. I think 90% of time I would be fine with 60x60, but sometimes it would be handy to have 120x60cm (now I also need to cut the 120cm sheets in half but that is not a big deal) e.g. for carving a sword.
I thought such a simple thing as a bowl would be an entry level project to 3D carving. Maybe I should start with less deep holes or something like (wall) plates. That should probably also be fine with 2-4cm.
Probably it is really the best to start with the recommended size and maybe build another one or increase the size later on, or maybe even buy some 2000$+ machine once I have more experience.
1 Like
There’s a shift of expectation that has to happen when considering projects on the CNC. It is no longer the intricacy of the moves that defines “difficult,” it is the ability for the machine to maintain rigidity and repeatability. Often this comes down to work holding/alignment and tool choices.
The elements of the leaf bowl pictured above that challenges the MPCNC is the tool “stick out” for the depth of the carving, which compromises rigidity and the need to flip the piece over and re-secure it in a precise location to mill the other surface of the piece.
Several folks on these forums have had early successes milling boxes and lids, where at first glance the fit of the top would appear to be the challenge but these are usually milled from a single setup where neither the top nor the bottom of the box are milled as deeply as the leaf bowl appears to be.
I would roughly guage these projects in increasing difficulty:
- Pen plotting
- Through hole cuts. Cutting an outline, and making some internal holes
- 2.5D, which will combine some through hole cuts, but also some dados or rabbets. Cutting some pockets to a specific depth, not just blasting through
- Carving, with a V bit. This requires good control of the Z because there are places where just the tip of the tool would be carving. Any error in height results in an error in width.
- Jobs with tool changes. Including using a flat bit to rough out a carve.
- 3D carving, with a bullnose bit. This requires good height control across the whole area, and usually multiple tools. Even small projects can take a really long time.
- Multiple sided projects. Carving the top and bottom can be a real challenge. It adds several more places to make mistakes and at least doubles the run time.
That’s based on the assumption that more steps == harder. More time == more opportunities to fail. More steps == more things that could go wrong. Also, you need to know everything to do a 2 sided 3D carve.
Honestly, none of this is rocket surgery. But there are real, measurable differences in difficulty between these steps. The only thing that is harder would be to make it a lot larger, or including pauses between steps, maybe a translation in XY to cover more area.
Another thing that isn’t obvious from the start is trying to estimate time. The time is basically the depth times the length. If you go twice as deep, it will take about twice as long. If you cut out a square hole, the time is the sum of the length of the 4 sides. If you want to cut out the pocket to half the depth, then you need to travel back and forth across the square, so it would be the length of one side times the number of passes that will take to clear. That can easily be 50x longer than just cutting it out. A 3D cut, especially a pretty one, will need to go over the whole area with a fat bit to get rid of most of the material, then go over it twice with a smaller bit (and smaller bits take more passes). You’d go over it twice to go left/right and then do it again going up/down. That will result in a smoother surface. If you want to flip it and do the other side, you’ve again at least doubled it. If you’re getting into really long cuts, you also sometimes want to go slower to just give yourself that extra guarantee you won’t be starting from scratch after so much work is done.
Cutting out a 4" square from 1/2" material can be done in a few minutes. Cutting out the letter ‘A’ in 4" tall font is just a few minutes too (it doesn’t matter much how detailed, just the length of the cut). Making a 2 sided bowl that is 4"x4" could easily be 3 hours. If the bowl was 8", that’s 4x longer, which is more than 12 hours. And, unlike a 3D printer. You can’t ever leave the CNC alone. It starts fires pretty quick if you bury the bit. The 600W router is not very forgiving and the whole thing is flammable.
1 Like
I am wondering how they mass produce bowls like the leaf shaped? Probably they won’t spend 12h for such a bowl to sell it? But how you produce such a thing in a reasonable time?
Likely big, heavy, expensive commercial CNC machines.
So probably I should just start my hobby with a 5D+, 200000$+ cnc machine to be happy with 
1 Like
Odds are you’ll be pretty pleased with what you’ve built with your own hands.
1 Like
There are a lot of shortcuts you can make when you are building thousands, or hundreds of thousands. I imagine the machines at ikea are pretty interesting. I wouldn’t be surprised if they have a machine built for each piece of a desk.
Making a simpler shape is easier too. I bet anything that can be made on a lathe can be made on an industrial lathe in seconds.
You posted this link as an example of what you wanted to create with a CNC machine, so I was curious and poked around a little bit. I only have access to parts of the site you reference and I’m reading translated German, but I believe he created this bowl using this machine. Based on the components in the picture, this looks like a CNC in a class above what I consider a hobbyist machine. So in a class above machines like the Shapeoko, Longmill, X-Carve, and WorkBee…probably in the same class as the Stepcraft M-Series machines. So it is not a $20,000USD machine, but probably comparable to a machine in the $3,000USD to $5,000USD range.
Second, he uses log slices with an undulating edge as his stock and carves parallel to that edge. It creates a nice effect with the grain following the outer contour of the piece, but he had to first model that contour, and then he had to align and position the router with respect to that model. This is not a beginner task.
This is also a two-sided milling, and you can see from Jeff’s difficulty list, this is at the bottom (i.e. more difficult) of the list.
Also for as beefier as his CNC machine is over an MPCNC, he does not have a huge amount of height for his stock. A comment indicated that he has only 80mm, but looking at the bowl, it looks more like 100mm. So even this more capable machine has relatively short Z working height.