Looking for some input to my design.

I'm planning to build my own table here soon. 5x10, Aluminum gantry, 4.5 NM Nema 34 motors (undecided on steppers vs hybrids vs servos), dual double T5-25mm timing belt X axis (doubled on each side) protected under table, cable track protected under table, timing belt and ball screw Y axis, ball screw Z axis protected inside gantry, replaceable water table pan, welded steel table, V bearings on v guides on top of rails and v bearings directly on stainless steel strap on bottom, motors mounted low for counter balance and protection. For control I plan to use LinuxCNC on a dedicated computer.

I don't have a ton of experience with CNC plasma but with my math it seems this should do 300 ipm and slower speeds well if I get the gear ratio in the belt drive correct. I have a XP45 with a machine torch and it shows a max of ~250 IPM for 16 ga. The ball screw will be preloaded against the gantry tube so that should minimize any whipping. The timing belt setup for the length of the x axis will sit against the tube of the table and have a second belt interlocking with it to minimize stretch and belt movement. Not that the gantry will be able to stretch the equivalent of 4" wide T5 belt.

What is everyone thoughts on my design? I'm open to constructive criticism and have no plans to manufacture it more than once (only for me).

The rendering is fairly basic since I do not know how to draw timing belts and I mostly did it to figure out the needed material lengths. I plan to mill lightening holes in the gantry side plates once I have all the limit switches mounted and timing belt covers made.

Thanks in advance.

https://youtu.be/X9CMCi_hZ1M

Hi Phil - Why double belts? a 25mm belt is a strong stiff belt. Plasmas have no contact loads so a 25mm belt is more than adequate? Plus one will be slightly different to the other and they will conflict in some way.

Add up your cost for the V bearing set up, you may find round rails are cheaper and more reliable once you add in all the bits to get a V bearing system working....I usually model a belt without the teeth so its simpler and initially in two overlapping (like two U's) halves. In this way when you change the centres the "belt" changes with it. Once you nail down the design then you can convert to a "single belt" if you wish... what CAD are you using? cheers Peter

Apparently my belt idea is already patented. The double belt makes it more protected and easier to mount. I drew the table fairly quickly in Sketchup. It's such a simple program it gets on my nerves sometimes, but it's all I have on my home computer. I'll probably pay for a Fusion360 license soon.https://www.makerstore.com.au/wp-con.../Everman-1.png

Quote:

---

Originally Posted by peteeng

Hi Phil - Why double belts? a 25mm belt is a strong stiff belt. Plasmas have no contact loads so a 25mm belt is more than adequate? Plus one will be slightly different to the other and they will conflict in some way.

Add up your cost for the V bearing set up, you may find round rails are cheaper and more reliable once you add in all the bits to get a V bearing system working....I usually model a belt without the teeth so its simpler and initially in two overlapping (like two U's) halves. In this way when you change the centres the "belt" changes with it. Once you nail down the design then you can convert to a "single belt" if you wish... what CAD are you using? cheers Peter

---

Just in case you haven't already run across these:
A commercial version
https://www.bell-everman.com/product...elt-linear-sbl
and the designer's thread on cnczone
https://www.cnczone.com/forums/linea...570-forum.html

It looks simple enough but I've spotted very few builders implementing the idea.

Hi Phil - The everman system has issues I believe. its not as simple as it seems. Plus the cost of two belts puts you in ballscrew costing territory. You can use Fusion for free for personal use. Use the belt on its edge to keep it clean or make it a "high rail" design even cleaner. My first machine used a 16mm belt and it cut plastic, timber and some aluminium, my small machines use a 10mm belt and it cuts plywood and timber easily as well so will drive a plasma easily. You won't need NEMA34s, 23s will be fine and be faster plus the whole system will be cheaper. If you want more accuracy find 400 step motors. I've used these on my small machine and they work really well and they are NEMA17's cheers Peter

I was unable to find 10 foot long (3000mm+) ball screws reasonable hence the reasoning for the belts. If they don't work out I should be able to replace them with gear racks easily.

I am also confused as why a Nema 34 won't work for the axis on this machine. You hint at them being slower. The cost difference between 23s and 34s is negligible. I normally overbuild things but am having a hard time finding data as to why too large of Nema motor is bad other than a little weight.

The only reason I'm building my own machine is my last attempt at ordering a machine cost me 6K which I may or may not get back and nothing on the market has both the features and software I now wish to use. I do however appreciate your input.

Quote:

---

Originally Posted by peteeng

Hi Phil - The everman system has issues I believe. its not as simple as it seems. Plus the cost of two belts puts you in ballscrew costing territory. You can use Fusion for free for personal use. Use the belt on its edge to keep it clean or make it a "high rail" design even cleaner. My first machine used a 16mm belt and it cut plastic, timber and some aluminium, my small machines use a 10mm belt and it cuts plywood and timber easily as well so will drive a plasma easily. You won't need NEMA34s, 23s will be fine and be faster plus the whole system will be cheaper. If you want more accuracy find 400 step motors. I've used these on my small machine and they work really well and they are NEMA17's cheers Peter

---

Hi Phil - They will work but they are overkill for your purpose. 1) they will cost more 2) the drivers will cost more due to the greater Amps capacity plus the required larger power source, bigger wires everything is bigger and more $$$ 3) The rotational inertia is more and the inductance is more both of these slow down N34s if speed is required. Do not select a motor until you have designed your mechanics so you know the weights so then you can calculate the actual size required IMO. :) If you use a U drive on your X axis then you have to carry two heavy motors back and forth , more weight. For a plasma I suggest you do not use a U drive but use a continuous belt so the motors are stationary, less weight to accelerate. Peter

The Servobelt is clever bur no doubt there are many ways to get it wrong, some considerations and pitfalls are mentioned in the thread I linked earlier. They have a different design for vertical orientation. I've never seen one running upside-down.
Rack and pinion would be the same price or cheaper according to someone in that thread who did some rough costing.

Quote:

---

Originally Posted by cyclestart

The Servobelt is clever bur no doubt there are many ways to get it wrong, some considerations and pitfalls are mentioned in the thread I linked earlier. They have a different design for vertical orientation. I've never seen one running upside-down.
Rack and pinion would be the same price or cheaper according to someone in that thread who did some rough costing.

---

It seems helical cut gear track is the way and just fine for tolerances. I should probably just go that way. It will probably wear faster and cost a little more in the long run.

The only complaint I've heard on ball screws is they are slow, but that seems easily corrected with proper gearing / ratios.

Quote:

---

Originally Posted by peteeng

Hi Phil - They will work but they are overkill for your purpose. 1) they will cost more 2) the drivers will cost more due to the greater Amps capacity plus the required larger power source, bigger wires everything is bigger and more $$$ 3) The rotational inertia is more and the inductance is more both of these slow down N34s if speed is required. Do not select a motor until you have designed your mechanics so you know the weights so then you can calculate the actual size required IMO. :) If you use a U drive on your X axis then you have to carry two heavy motors back and forth , more weight. For a plasma I suggest you do not use a U drive but use a continuous belt so the motors are stationary, less weight to accelerate. Peter

---

I'd love to move the X axis motors off the gantry since they are a majority of the weight but have not though of a clean accurate way to do it. I'd love to have ball screws but they get expensive and difficult to transport at 3000mm+ plus they can really whip at that length unless you move the motor drive to the nut and leave the screw stationary.

I'm not sure why the inertia issue can't be corrected with gear ratios since the 34s have much more torque. The cost is of minimal importance to me in all honesty.

Hi Phil - High rail design gets the belts or drive away from the much and you eliminate columns. But this being a big machine means you probably want to side load it? cheers Peter

Yes. Loading and unloading with my forklift is important. Once of the reasons the gantry rail is lower than the height of the actual table. The actual cutting surface is a little lower (.5" or so) just to allow water cooling without it overflowing on the sides.

Quote:

---

Originally Posted by peteeng

Hi Phil - High rail design gets the belts or drive away from the much and you eliminate columns. But this being a big machine means you probably want to side load it? cheers Peter

---

This belt on belt design uses a stationary belt and a short looped belt riding on the stationary belt like a bulldozer. It seems it would work in any orientation. I'm seriously considering it for a 24" x 36" 4 axis CNC router I'm designing. I'm not sure why he has the 2 looped sections engaged with the stationary belt instead of the longer section currently at the bottom.


https://www.youtube.com/watch?v=NJZ7YVHRwAM&t=19s

Hi - The wire - This sort of drive is used on large printers and plotters. Seems to work in that application. Peter

Hi The wire - The reason the "two" belts are used is that the eccentrics move these apart to take up the slack in the stationary belt. Belts do not match when you interlock them. They are designed to snug into a round pulley which means when they are straight they have quite a big gap in them. This is like using two nuts on a drive screw and pushing them out to take out the backlash. One approach is to use an ATL belt which has snugger fit. But this is more expensive then a std belt so your into R&P budget by them. By the time you add the long belt, pulleys and fittings you may as well use an R&P, same cost and less fiddle. or a small ballscrew with travelling nut...but long screws get damaged in transport...at 24" x 36" I'd use a std U drive or a continuous belt system if you use a belt. My 1/2 sheet router uses 16mm belt and its 60" long works great....Peter

Lots to read through, and I probably missed half of it.

At first, I misunderstood what was being said about the "double belts" and I thought Peter was telling you to drive it with one belt underneath, and I was like, whaaat? But now I see what you are talking about. Double belts on each axis. That's what I get for skimming through the posts.

I don't have any experience in building machines with belts, although I do own a commercial linear actuator that is belt driven. My initial impression is that single 25mm wide belts would be fine.

Your other option is rack and pinion.

Yes, you could use ballscrews with a very high lead (like 50mm) that are geared down, it's just overkill and unnecessary for what you want to do. Also, not owning a plasma CNC, I am not sure how the dust might affect them.

Don't bother worrying about saving a bit of weight here and there. It doesn't matter if your motors move with the gantry or not. What acceleration do you need? High end commercial lasers might be capable of 2 to 6G, but my understanding is that 0.15 to 0.3G would probably be fine for a plasma cutter. Perhaps someone who owns one can correct me if this statement is not correct.

Nema 34 motors would not be slower than 23's, it depends on the characteristics of the motor, not the frame size. So a really large Nema 34 (which typically have high inductance meaning a quick drop off in available torque with RPM) might be a worse choice than a given Nema 23. But a low inductance Nema 34 might out perform them both. Really you have to look at the torque vs RPM graph for each motor, or simply choose a Nema 34 of low inductance.

I am considering using some JMC servos for one of my projects, and geared properly, an integrated driver 400W JMC servo might cost less than a Nema 34 and good driver, and that would probably do you just fine. But you do need to consider such things when choosing your gearing.

For a typical Nema 34 motor, regardless of if you use a belt drive or rack and pinion, you'll want around (10 - 32)mm equivalent lead, meaning that for each turn of the motor, your machine moves that amount. If you look at the Avid CNC (at cncrouterparts.com), it uses belt driven gear reduction on it's rack and pinion to achieve around 25mm linear travel per revolution. If you look at a pulley with a diameter of 25mm, it has a circumference of 2x3.14x25 = too darn much. So you need to adjust your gearing and pulley sizes accordingly. For servos it's different, and depending on the size of servo, more gear reduction is needed.

The commercial belt driven actuator I have has built in gear reduction, but not all of them do. That's another option, to look for a couple of belt driven linear actuators for your long axis. I have come across a couple at great prices on eBay but I don't have any links saved. They have the linear rails built in. But be careful as they come in all different kinds and you'd need to pick the right ones for you, with the right linear rails, and gearing, if you can find them at a good price.

The first CNC machine I made used IVT guides from PBC Linear, hardened ground edges, and V-Groove bearings with eccentric nuts, along with the gear reduction from CNC router parts and rack and pinion from Moore Gear. I think that was before they offered their pro system with square linear rails and I wanted something better than the rolled plate bearing system. This kind of setup, similar to what you have drawn, if I understand your description correctly, would work just fine for a plasma cutter, but the extra expense is not justified when you could find some square type linear rails and bearings for the same or less money. There are many examples of the IVT guides being used for plasma machines on the forums. I also have a bunch of plasma cut parts from shop droids, I believe I got their gear reduction carriages, that I never used as I decided to go with the wider IVT rails and made my own carriages from 1" aluminum plate. At one time people were direct driving the motion and then realized that you really need some gear reduction, so you will see a few different variations as you search.

It's too bad you didn't live in Ontario Canada or I could make you a good deal on most of your parts. I don't have space for them where I live now and they are sitting in a barn owned by a family member. My original machine had a cutting area of 12' x 5'. They would not be easy or cheap to ship plus there is some rust that needs to be cleaned up.

I wouldn't bother using a ballscrew on one axis and something else on the other, although you could do this, I think it would be better to keep what you do consistent on X and Y.