Proper frame for pnumatic cylinders

Hello all,

I’m just a preemie here, ready to birth a cnc machine for my small business. But before I sink 1500 on a setup and laptop I want to make sure I hit the mark for my needs

I am an upholsterer and want to kick my business into a little automation for pattern cutting. Digitizing my files will help keep clutter down, a cnc machine will automate cutting and I have an idea for solutions to my perforation problems.

In auto manufacturing there are many variations in perforated materials. Size of holes, spacing of holes and special patterns make finding and buying the correct material for replacing worn panels difficult. I want to solve this and avoid the headaches associated with tracking, or in many cases failing to track down, these particular materials.

I plan on attaching a pnumatic cylinder with a leather hole puncher to my machine. My concerns have to do with the rigidity needed for this set up.

Does anyone have any experience with such applications?

Would the mpcnc or low rider be rigid enough to handle punching 200 - 1000+ holes in leather.

Any advice or leads to info would be great, thanks!

That depends a lot on how heavy a pneumatic leather punch is, and how much torque it generates as it’s punching. As long as the punch doesn’t weigh more than a router, the gantry can sling it around easily enough. Making a wild guess, but I’ll assume you want to go with a LR, since the MPCNC really tops out at around 3’x3’, and the LR is effectively 4-5’x[however long you can get a table or belts]. Which is actually good, since the gantry is supported on two sides, symmetrically, and any force from the punch can be absorbed evenly. The MPCNC supports the tool from one “corner”, so the potentially sharp, violent forces would be twisting the gantry every time.

Again, it really comes down to the weight of the punch, and how much force the gantry has to exert/resist to punch through the leather. For example, you may not want the usual spiral-cut connectors you find on most z-axis led screws, as the continued forces from the punch might stretch them out. You may need to use a different pitch threading on your lead screws to prevent the punch from pushing the gantry up too easily. Or use beefier motors to handle a heavier gantry, and/or resist the force of the punch.

Your limitation will be the amount of downward cutting force that you need.

The MPCNC has a smaller work area, but can be large enough for most perforated panels that you may need. The advantage of it is that it does not rely on gravity to keep the gantry “on the table” and can withstand upwards pressure on the rails. The actual Z axis, however, mostly relies on gravity to keep the core on the lead screw, though you can glue that down some.

The LowRider relies on gravity to keep the machine on the table. It rides on wheels. There’s also the nut for the Z screws involved here.

None of this is a problem if the pneumatic cylinder never exerts enough downwards pressure to lift itself and the machine core off of the table, of course.

I would not, for example, want to try to mount any kind of a press to this or any other CNC router type machine. but I have used a leather punch, and I don’t think that the necessary forces should exceed that, so long as you keep things reasonable.

It’s probably easier to manage tool changes on the LowRider, and also probably easier to place your material on it as well. The design where the machine is supported across the 2 bars is also easier to manage, I think.

The low rider and the mpcnc depend on gravity to hold the Z down. On the mpcnc, there is backlash, at least in the leadscrew, but also in other parts of the machine. Pushing the gantry up is not what it was designed for. But we have been surprised by new applications before.

Is it possible to cut the holes with a drag knife?

No experience with it but might the bottom/dual roller style of the LR1 help with this?

Thanks all for replies, and others feel free to continue to chime in if this subject interests you.

KCummins: 3x3 cutting area would be enough, 2x2 also, if 3x3 isn’t rigid enough. Not sure about torque yet but I imagine 60 to 90 psi pushing an 4 oz weight? weight of punch and cylindar is likely far less than a router

Jeffeb3:The holes are as small, like ball point pen tip small, so drag knife is no good, but I would love a sonic drag knife for cutting out patterns if anyone knows of such modules. Or maybe solutions to hold soft goods while cutting to keep them from moving.

I wonder what parts have backlash (slippage and movement when upward force is applied?) And if these parts can be secured. I see there are bearings all around the pipes, can they be adjusted to reduce movement or is position fixed?

I would like to know more about what parts backlash if anyone has experience with this.

I’m starting to formulate some ideas based on everyone’s insight. Since I have 0 experience with cnc I’m having to imagine what is possible and use you guys to correct my assumptions. So please continue jump in with more wisdom if something I imagine doesn’t jive, and thank you for your time and patience.

1st: the weakest link

I used to skateboard, so I am well aware of what a bearing can take.

My experience with hole punches makes me think the printed parts can handle the striking.

I imagine if screws start coming loose i can use thread lock

what I need to know is what that force will do to the pipe or conduit, how punch will affect pipe near the joints vs the center.

I may have to buy a cylinder and rig it to some piping at diff locations to see what will happen. This will also give me some input on psi force necessary for cutting.

I wonder if anyone has ever wiggled the pipes or accidentally pushed the z down hard with router turned off and seen the pipes bend. This experience would also give me insight into backlash, where it happens and how much movement there is.

What pipe would be the most rigid?

I think 2 ft sq cutting area would be plenty so that might help keep things rigid.

2nd: frame for z axis

Mpcnc seems to be the best option since it is not fully reliant on gravity, but as you guys pointed out, z axis seems to be a week point.

Can I avoid the z axis all together?

What if I mount to the frame that holds the z axis. I think the stroke of the cylinder doesn’t need to fully extend, which means I can position it to stroke past my cutting board and the impact would stop it so I would not need a z adjustment. This would let me have a knife and hole punch at the same time! I could program it to punch all the holes then cut it out!

Stroking past the cutting board could also help with possible issues if pipe moves more when punching in center of cutting area vs near the ends. If the pipe flexes, lets say 1 or 2 mm in center and less near end then stroking 4 or 5 mm past cutting board will keep the flex from making weaker punches that dont cut through in the middle. Perhaps more would be needed for slippage from other parts of machine?

Can anyone push up on their mpcnc z axis housing to give me an idea of how much play and how much force was needed to get to the play?

Thanks if you read all that XD

Do you mean it locks on to the plywood with wheels on top and bottom?

Mostly the Z screws. But the biggest thing there is that the nuts are set up to rely on gravity to keep them in place. Honestly, I don’t think that this would be a huge problem unless you’re looking for something that is going to punch through a dozen layers at once. If you’re going to just let the machine punch one hole at a time, and one layer, it’s probably less force than we get when plunging the tool when we do what we do, since end mills don’t actually work very well as drills.

Crashing into the bed happens. The design shouldn’t bend tube when this happens, it’s part of the reason that the Z nuts can come loose the way they do, and they should in order to protect the rest of the machine.

2’ square working area is actually a pretty rigid machine.

If you want a dedicated punch machine, then it doesn’t need a Z axis, but you still will want it for a drag knife to cut patterns. (Even if you end up dragging a sharpie to draw the patterns to cut later.)

It’s a weak point, sure, but what I’m hearing for the detail of what you want to do, it’s not likely to be much of a problem. Pretty much the minimum for a MPCNC is about 3" of Z. Less isn’t really going to help you, and you aren’t likely to ever need more, even with a number of different tools, including being able to carve foam. You could probably also (for example) cut door cards, though 2’ isn’t quite enough for that.

Mine lifts relatively easily, most of the downward force is the weight of the router and the Z axis moving parts, but it’s a sustained lift. A momentary lift like a punch is a different kind of lift. It works against inertia as well as the actual mass of the gantry. If you get a problem, you can add about 1kg of mass around the punch, which an MPCNC with 1" DOM or stainless steel tube can easily carry around and it will remain rock steady.

For this, what we may end up with is a problem with generating the program to make the punch do what you want.

For that, there is code to use a laser. I think that setting up the code to do momentary laser bursts, and use that to either fire a solenoid or open/close a valve for pneumatic pressure would manage, then you just need to run the code to do a dot pattern with the laser. The code for a drag knife isn’t difficult at all.

For cutting the pattern, lots of code generators support “holding tabs” so you leave little bits of the leather holding it to the rest of the piece in strategic locations, so that it keeps in place. Aside from that, a vacuum table would probably work really well, where you use a perforated sheet as the cutting board surface, and run a vacuum through it. Air pressure holds the work piece in place.


Correct, I never owned one but I think your scenario is why the second set of wheels was keep the unit from lifting itself off the table. It turned out that it wasn’t needed and they were removed but your use case may be an exception to that.

I think this answers my questions. The weight idea is perfect in case there are any issues. The hammer I use for punching by hand is way lighter so I imagine if there are problems this would help, but as you mentioned, the gantry weight will probably be sufficient.

I might try to put the knife on the z axis, and rig the punch on the frame that holds the z. I’m thinking I can find a cylinder with the reach that I need.

Now for programming. I really appreciate the idea of laser bursts. I was planning on using a valve with a dual action cylinder.

I’m pretty sure I’ll be doing lots and lots of playing around and learning with software. The machine looks super simple to build, if the electrical hardware comes with instructions it will make my life easier, but I was super not looking forward to the learning curve with the software. Thank you so much for the laser burst idea, that seems like something that is within my reach. Gonna get a cylinder and do some experiments with a punch and some conduit I have laying around.

Thanks soooooo much. This really helped me wrap my head around what I can do with this awesome affordable machine!

Since the work area you need is not all that big, reading through most of this, I think it could work. I am really interested to see what you have planned for a pneumatic punch. You can add a lot of mass to the gantry of the MPCNC to help deliver a good massive pop. The weight you add only means you need to move (more accurately, accelerate) the machine a bit slower between punches.

I think if you punch hard and do not actually deliver much Z+ force after hitting the surface, it will be okay. Meaning, try not to jack up the machine with your punch, after making the hole. Try to hit near the end of your stoke or hit fast with little pressure.

Hmmm, this is fascinating. Maybe we are all over thinking it.

@Chrisa410 what about using an automatic punch similar to one of these? This wouldn’t need any unique gcode since you could treat it like a drilling operation. I’ve only used automatic center punches so I’m not sure how well the one I linked works but it would simplify your setup.

We probably are overthinking it. I have 0 experience with cnc so I just wanted to make sure I had some basic understanding of what this frame can take. the more I talk about it the more I’m convinced the forces at play are no problem for this machine. Also i was thinking 3/4 pipe ( maybe from older versions) but the primo seems to be 1 inch pipe so it makes me feel like this is way within tolerances for my application.

That’s a neat tool, but I’m guessing, even with air punch this will be a lengthy operation. I would like to get a hole per second if possible. Something like this seems like it will take a while to make each hole. With pnumatic all I have to do is travel and stop before striking, with this thing I have to go down then up for every hole.

Yeah, that’s going to be time consuming going one hole at a time. You may be able to make a custom head with multiple punches to speed it up but that will also increase the force. Still sounds like a fun challenge. Good luck!

Here is a link to the thread discribing the vacuum table I made for upholstery cutting.


That’s a cool table. I love how you dont need holes. I will likely be gluing down to start with. I already glue back of panels for sew foam and punching holes will need a solid cutting board. I use super 77 from 3m but may look at other mist sprayers that clean easier. Whet did you spray before?

It would be a good experiment. I think we would learn a lot about the backlash constraint we mostly take for granted. We may also learn more about how impact affects the machine. I don’t really know how the hole puncher works, or what order of magnitude the trouble is. But even a small hammer delivers a lot of impact.

But yeah, overthinking is my speciality. :slight_smile:

Scotch brand light duty spray mount, it says it is initially repositionable. It wore off after a few and I didn’t like cleaning up the face up side. It doesn’t sound like that will be a problem for you.

1 Like

Super cool. MDF being that porous was a surprise to me.