I want to air out a laundry list of features and benefits people would want in this next version. Possibly to inspire Ryan to motor towards something that fits in with the ecosystem we have now. The Primo was made before the LR4 (before the LR3, IIRC) and before the jackpot.
Now. The LR4 is very good. The break even size between MPCNC and LR was 24". But I don’t think that rule of thumb holds up anymore. The LR could be used down to very manageable sizes and still rock. The reusable table and potential full sheet capability makes the LR4 great for the middle of a workshop or a portable table in a garage.
So what space is there for the MPCNC? I am going to argue that the best place is a “desktop” version on a workbench, in a basement, or even in an office. If the MPCNC could let go of trying to be big and instead focus on being small, inexpensive, and performant, it would find a good home in a lot of people’s work flows.
This would be my list of features for a machine like that:
Enclosed
Dust collection
Air assist
Very rigid (to make it easier to use)
Small footprint (to fit mostly on a counter or desk)
Good visibility and lighting.
Router or Laser Diode
I would give up these features to get there:
Large size
Adjustable size (this feels like heresy, but let’s talk about a fixed size CNC)
Flexability in parts (if there is one easy to source vac hose, then let’s all just buy the same one).
I would like to keep:
Jackpot/FluidNC
Low Cost (this means a lot of things when some desktop machines are $2k-$3k).
Self source option
Wood, Aluminum, Plastic capability. At least for sheets.
Community supported and DIY attitudes.
I would like to imagine that having a set machine size and starting from relatively similar parts, we could start getting more detailed on the user manual. Especially if we had some conservative starting point for tuning and the machine was a little over built to keep things easier. I would love to have something really polished and inviting for people who are regular woodworkers and don’t like to tinker. But that doesn’t feel realistic. I like the yellow brick road. A machine that was fit and lean and repeatable could still be a value add to this group, I think.
You are basicly putting into words what I’ve been thinking for a while now! When we (eventually) will get the LR4 running at full sheet size, my small-ish Primo at home wouldn’t be very interesting for larger woodworking projects. But I’d still like a small CNC at home, for running the laser, making signs and inlays, some careful metal cutting, and some basic cutting/plotting/etc. We all need at least a few kinds of CNC in our lives, don’t we?
I think, for me at least, the key thing would be to identify the defining things that would make the MPCNC configuration preferable to just building a small LR4.
What is the key advantage that we see the MPCNC having? Is there a specific niche that it excels at, a specific aspect of the mechanical configuration that suits some tasks better or is it just that it’s different/unique and it’s nice to have a 2nd option? Are those benefits significant enough to be worth the time/effort input when it could be expended towards making something that doesn’t have a significant amount of functionality overlap with the LR4?
Would that same time be better spent making/showcasing a few specific ‘yellow brick road’ configurations of LR4 to make it easier to follow a given recipe without having to make decisions. Or time spent doing new things to ‘push the envelope’ a bit? Even if that is really just cleaning up some existing 3rd party mods and making them official ‘supported’ configurations. Small stubby LR4 milling steel. Modified/beefed up LR4 core for 80mm spindles. Customisations for alternative tools such as plasma cutters, lasers, ATC support etc. 3D Touch probe for stock location and profiling for engraving etc. Tool height probe. An official enclosure.
Are there any other things that could go along with the ecosystem to solve people’s existing problems, things like the USSA 3d printed tooling etc. Dust extraction/cyclone stuff like some of the motor driven centrifugal separators, etc.
Personally, I can’t see a situation where I would re-build my MPCNC in a significantly improved fashion vs making an LR4 in the same format so I’d be fascinated to hear where people think the best value lies.
I got five nema 23s from a friendly dude at a maker faire who liked my Primo a lot. I’d love to have something utilizing their potential, but I guess that would raise the price class quite a lot…
With multiple build stages and stops? Ideally Yellow brick road’s 1st stop is “Bootstrap Mode”, and uses the machine to progress towards more advanced stages (for people that want/need more)?
Yeah, I think that illustrates my first pass thinking about this. Is this a new MPCNC because there hasn’t been one for a while and we think there ‘should’ be, or is it because there’s something specific about the MPCNC configuration itself, because some of this could equally apply to the LR4. Personally I think that the analysis-paralysis aspect of building one of these things is often overlooked because those of us who have been through it kinda don’t see that anymore. The LR4 docs are a huge step forward in that fashion. I feel that it could still be improved by having a set ‘recipe’ to follow and one of those recipes being for a small stubby LR4 could be an alternative.
That seems like a very bottom-up requirement, though? Is there something specific about having bigger/higher torque motors that would suit the MPCNC format? We’re already not running the NEMA17s at anywhere near their full capability…
Yeah yeah, I know I know… I just want to make use of them, without needing to step out of v1e territory! I didn’t mean it as a serious argument, more a thought by myself.
My only thoughts after just building an MPCNC are the following:
why is there not a standard case for the controller board as part of the original release? Much more attention should be placed here for any final design to facilitate cable management, cooling of the drivers, etc.
it’s obvious that there’s a lot of backlash with the Z leadscrew design. I recommend updating the BOM to use the most cost effective anti-backlash device found from the community.
There’s no integrated cable management in the original design, this needs to be addressed in any future updates. Leaving someone new like me to find the “best solution” was troublesome. Wiring and strain relief from the motors and endstops should be included in the design by default.
I’ll end this on a positive note, the design is pretty solid and well supported by this community even though it’s a few years old. Usually the excitement fades after a year or two when something new and shiny comes out somewhere else. I really appreciate the community support and documentation. This has been a fun learning experience for me to get into machining with very little investment. I’d also like to add that I wouldn’t want to build a new version of the MPCNC vs building a LR4 once I have more covered space and that building an MPCNC was only because I wanted to be able to move a 3x3 table into a shed after using it.
So, for a “Desktop” style machine it will need to target smaller and quieter. Imagine something maybe in the 4"x12"x24" working volume, enclosable, with a replaceable toolhead (laser, spindle, knife, pen, needle, etc.) Options for dust collection.
In this size you aren’t going to fit a trim router, so finding a cheap but good spindle seems to be the hard part.
Every perception is important and I’m glad you raised this concern. I asked for a laundry list.
But I do disagree that this is a problem. At least from a technical side. The gantry is heavy and it is pushing down on that coupler. Hard. It doesn’t need to be an anti backlash nut. That is for situations where there is backlash. This has none because of the direction of gravity.
Anti backlash nuts do add friction. Which would make it harder for the Z motors to lift the gantry.
That said. We talked for a long time about how endstops were not technically necessary and the thing that ended up being the solution was to just assume everyone wanted them. Needed or not. People were happier to not argue about it and the kits were not really any more expensive or complicated. So a fix that wouldn’t add friction, but sidestep the debate would be ideal.
This is what has me so excited. Can we reduce rhe footprint enough that there is 6-8" on each X,Y around the work area? Is that possible? If the footprint was a usable 12-18", with workholding outside of that. I think you could have a very useful desktop machine.
It would probably look like a mashup between an MP3DP V5 (coreXY, linear rails), and some kind of spindle that can move up and down at the point where the extruder would otherwise be.
No flying Z, obviously.
Edit: Or maybe like a a baby version of one of my favorite machines ever, the now defunct Diabase H3. (Multi tool head, multi axis. Could additive and subtractive machine)
Yes, folks, that’s an endmill on a baby spindle in that image along with multiple extruders.
This is the area that I would look. A few years ago, when I built my first MPCNC, I remember feeling really good about the machine with one exception. The size of the machine versus the actual cuttable area was a shock. If there is a way to make more of the area useable that would be amazing.
I’d also say Z-capabilities are the biggest advantage of the MPCNC over the LR4, and I don’t see that changing.
I also stare at the MP3DP and wonder if its possible to focus on a laser or smaller tool and bring some ideas from that world to CNC (using an adjustable Z-bed, core XY, etc.) For example, if the spoilboard/bed worked up and down, the stability of the tool could remain fairly constant.
That probably WAY overthinking this question. But it’s all in fun.
The only reason why I brought up the backlash is I watched as the tool was moving and cutting along the X axis and I saw the toolhead vibrate up and down. The only place I could find movement was the 2 mm of backlash caused by the z leadscrew.
The mpcnc is the gateway drug primarily because it is cheap and accessible with easily sourced conduit, bearings, and even self-printed parts at a price point that is rather low. I think if you are championing feedback for a replacement, it should continue to fill that entry level accessibility price and capability. Beyond that it needs to outperform the 3018 that is easy to get for very little. Right now it is larger so it isnt a direct competitor because the other is only so big. If you only make this smaller, then it needs to have some other value proposition.
I will reiterate what several have already said here in an attempt to add emphasis: work area vs. machine size. I feel like so many “Your Builds” posts I have seen involve OP saying “I have a small shop so space is limited”. I think something like a mini mpcnc that has set dimensions, with an enclosed design, would really enable people who don’t have a dedicated workspace to build a machine or at least be able to dedicate a small amount of space to that machine. Having built both an MPCNC and a LR3/LR4, the Lowrider just feels like it uses the area that it takes up so much better than the MPCNC.
The number 1 draw (at least for me, and I’m sure others) of any V1E design has been low cost/easy to acquire components. The idea of buying a shapeoko/onefinity/etc. was never on my radar, and I could never justify the cost of a $2k-5k machine as a hobby, but being able to use old 3d printer parts and hardware store components to make something that I would say is comparable to any of those machines is what sold me. A new MPCNC I think would have to keep with that trend in order to ensure the value is there.
In case Ryan decides to invest his time in a new MPCNC design I would like that in addition to the suggestions indicated so far by Jeffeb and others he could include the option of a fourth rotary axis trying to find a way to integrate some commercial device in the MPCNC Firmware. Anyway I am already very grateful to Ryan for his projects.
I have been around here off and on for several years. The mpcnc was a fun build for me to learn a ton. I haven’t done any upgrades to the burly I built and it is great for what it is.
I have started making folding knives and the MPCNC has been a great gateway to cnc and producing the scales. The steel and titanium are laser cut by vendors beyond my capabilities.
One of my problems, I need to tackle my skew issue. I think this is something that can be solved once I dig into testing and shimming the end stops.
I would like to be able to make these wooden and micarta scales with enough accuracy and precision to mill one side, flip them over and mill the other side.
Right now I mill the inside on the MPCNC and use a bench top router on the outer bevel. There are many options if I can get make both sides on the MPCNC
Does that mean ball screws? Aluminum plates? I looked into a small printnc and wasn’t sure it was worth it.
Skew as in X/Y not being square, so if you try to cut a square does it come out as a parallelogram? Are you running the dual-endstop firmware? That’s the solution for skew on the MPCNC. Either way, that might be worth a separate post to chase down if it’s something that you’re having an issue with.
Ballscrews and metal vs plastic, linear rails vs bearings on tube etc, all of those considerations are rigidity based, i.e. force applied vs amount the tool deflects, and they all go together hand in hand. If you’re able to machine the material successfully at all then that means your existing force vs deflection isn’t too bad and if your tolerances are acceptable then it’s perfectly fine. Any improvements in that would likely only allow for faster cutting, really.
