2 steppers 1 axis. Why not?

[Blackened]

Yes, all that is true, but perhaps you have not read the whole thread... we are talking about two motors with one home switch only versus one motor, two screws and a belt in between. Two motors are going to be misaligned in no time if you don't align them on regular basis. One motor, two screws and a belt will never get misaligned once it is aligned when you start it up for the first time.
.

To clarify, each side of the gantry will be driven by it's own belt. If 2 steppers are used, they will simply drive the individual belts directly, if a single stepper is used, there would need to be drive shaft/s, pillow blocks etc, which in turn drive the 2 parallel belts.

Screws would be too slow. I expect I'll use screws for the z axis though.

[ger21]

You can get high lead screws that would be very fast, like these.
https://www.roton.com/family/torqspl...-nuts-7060998/

[Blackened]

It's an incredibly simple solution that works, and requires no switches or additional components. For this application, it would work perfectly fine.

Given the smoothieboard can't home the 2 slaved axis independently, this does seem like the best option. I was envisioning a removable jig though. Couple of contact points between each side of the gantry and the end of travel, manually press the gantry to the jig ( motors off), then home to the limit switch. But maybe I'm over thinking it. Either way, I've settled on 2 steppers.

To be investigated further though, is a possible way to decouple the 2 axis via an M code and a multi-pole relay or the like. Then home the axis, then re-couple. The smoothie config seems to disallow that due to certain settings for the slave needing to be removed from the config file. This would mean there's no step/mm settings for the control to use during homing. Might not work at all. But regardless it seems that 2 steppers will be workable. I can try and finesse a more automatic solution in the future.

[Blackened]

You can get high lead screws that would be very fast, like these.
https://www.roton.com/family/torqspl...-nuts-7060998/

Thanks Gerry

You should see the ones on the industrial laser I use at work. I haven't measure the pitch but it's something like 20-30mm. Huge travel per revolution. I like the belt idea due to low mass and a *perceived* higher feed. I haven't seen any examples of feedrates achieved with those type of screws, and I haven't done the sums to compare them. Certainly not suggesting they don't work, just that I'm more confident that belts will get my end result. And probably cheaper? This is a hobby build so I won't be recouping costs with an income.

edit: had a closer look (actually opened my eyes lol) and note that those screws are 1 inch per revolution. They look very different from the ones I'm used to, mainly as there's only one start thread, with most of the screw just bare and round. I'm going to measure the ones at work next time I have the covers off. Curious now

[NIC 77]

I think the high end commercial lasers have accelerations around 4G or something ridiculous like that. Not really comparable to a CNC router. I've seen a few examples of 6G on Youtube.

I'm guessing that for light components where you want high accelerations, belt drives are a good choice.

For the Z axis are you planning to have the bed move? I assume so as you mentioned screws (more than one). Makes sense so you can accelerate the gantry as fast as possible without the extra weight and offset from a Z hanging below.

I've been looking at "2.x laser" on YouTube. Looks like lots of people are using this design.



Correct me if I'm wrong, but that's a single stepper running both belts, and lots of people are happy with this, and it seems really fast (in other vids). Forget what I said earlier. Just copy that, the design appears to be open source.

[Blackened]

I think the high end commercial lasers have accelerations around 4G or something ridiculous like that. Not really comparable to a CNC router. I've seen a few examples of 6G on Youtube.

I'm guessing that for light components where you want high accelerations, belt drives are a good choice.

For the Z axis are you planning to have the bed move? I assume so as you mentioned screws (more than one). Makes sense so you can accelerate the gantry as fast as possible without the extra weight and offset from a Z hanging below.

Correct me if I'm wrong, but that's a single stepper running both belts, and lots of people are happy with this, and it seems really fast (in other vids). Forget what I said earlier. Just copy that, the design appears to be open source.

Yep, screws to drive the bad as z axis. It's much simpler. Given it's a CO2 laser with tube and mirrors, it's much harder to lift the entire assembly. And having a moving z axis that is part of the gantry would be much heavier.

And I have actually settled on the 2 stepper design. Have a look at my gantry test:


I've made a few changes since, with at least one more change on the gantry to come once a few small parts turn up in the mail. I haven't calibrated the mm/step or anything yet. I was purely testing the concept to see if it wasn't totally crap lol. "I'm an operator, not a designer Jim!"

I'd like the gantry to move along the x axis at roughly the same acceleration and speed. I think I'll have a better chance at that with 2 steppers and no driveshaft or other restrictions to motion.

[109jb]

Missed steps is a fact if you have two steppers connected in parallel. The only way you can avoid that from from causing issues is by having TWO home switches, one for EACH screw and have a proper homing routine to sync the screws, or if your motors use some other electronic synchronisation.

Wrong. What you are talking about IS NOT missed steps, but a misalignment during power-up. Homing switches only operate during homing and missed steps only happen during operation. A homing switch can't prevent missed steps in any way shape or form. A missed step is a commanded step that is commanded and not executed. As stated before, on a laser that does not have cutting forces and only moves the fixed weight of the machine itself, missed steps would mean either undersized steppers, accelerations or feed rates too high for what is there, or electrical interference. All of which are design problems.

As for power-up misalignment, the amount of misalignment on a twin belt driven axis will only be as much as the mechanics of the axis will allow. For example, an axis with linear guides can only be misaligned as much as the linear guides will allow because as soon as you power-off the machine, the steppers are free (except for a small amount detent resistance) and the axis will relax to align to the linear guides. Now power back up and the steppers will lock in that position within 1 step. If this amount of misalignment is not acceptable then either fixed stops as mentioned, or individual homing switches can get the axis into better alignment, but only as accurately as the stops or switches are adjusted. If the stops or switches are not aligned properly they can actually skew the axis rather than straighten it.

[NIC 77]

It will be interesting to see if the unsupported round shafts are suitable for high speeds / accelerations. Typically they are the worst thing you could do for a router gantry, for a laser, I don't honestly know. Could be awesome. I really don't know.

Looking at the 2.x videos on youtube, some people used V groove bearings and some upgraded to square linear rails.

I agree that the belt drive is best for this.

For the speed of the gantry, well, it depends, one larger stepper with a design like the 2.x, geared properly so it is spinning at lower rpm's while the gantry is moving fast with the belts could work pretty awesome. Although it is probably a bit more complicated to build. I honestly didn't think about the idea of using a stepper with the shaft coming out of the top as well, and I pictured in my head a convoluted assortment of belts and gears tying both sides together. You'd need a driver that could handle more current for the larger motor.

So what will you do about the gantry homing and the two motors? I don't think the smoothieboard can do it? Are you going to dump it?

Interestingly enough, you may have the same problem with your Z axis, as that is driven by multiple screws.

The quality of the workmanship of what you've done so far looks nice.

109jb, you are correct. On my current CNC build, I am making it so the mechanics won't allow much misalignment, at least that's the theory. I may add some homing to it as well, but not sure if I can do that and keep the 4th axis with the current board that I have. I may need to upgrade that in the future. Just trying to get it built, that's a detail I can fine tune later.

IMO, it doesn't become an issue if you have good mechanics unless you go with large spans. For something like a laser or a printer where the mechanics are more spartan and less rigid then perhaps the issue becomes more important.

[Blackened]

Nic 77,

Yeah I think I'll just play it by ear and see what the reality of the misalignment works out to be. As Gerry mentioned, simply pushing the axis up against a hard stop prior to homing might be all that's required. Or maybe I'll make some sort or removable spacing jig that achieves the same result. And I've seen youtube videos of the same thing for the z axis. Just jog it down until it hits the hard limit prior to homing.

Thanks for the compliment, yeah having access to a 5kw laser at work has make the build go easier. I have to design based on the tools I have, which is the laser, and hand tools. We are temporarily in a rental so I don't even have room to setup my drill press And the laser doesn't cut edges at a perfect right angle to the surface, so that requires some consideration.

There's been a few mentions of open source designs etc, but that would defeat part of the purpose of this build, which is the challenge of doing as much of this myself as possible. I expect I'll make mistakes! If the 12mm rod guides aren't rigid enough I know I can reduce the accel/decel settings and still get a working machine.

I was looking at steppers with dual shafts, but they constrain the design a lot. Whereas 2 separate steppers driving belts directly can be swapped out for a variety of different types with minimal alterations. Even to go from nema 17 to 23 for example would just be a mounting bracket, couple of holes, new pulley due to ID, and probably a slightly longer belt. But to swap out a dual shaft stepper for something else, then I might have to get drive shafts shortened, or remade due to different diameter, new couplings etc..on top of the other changes. So I'm planning ahead for the possibility that I get it wrong lol.

Oh and re: shaft deflection, you don't cut at those high feeds, and etching photos is done in a series of straight lines, so it's mainly a problem during rapids, which isn't much of a problem lol

[A_Camera]

Wrong. What you are talking about IS NOT missed steps, but a misalignment during power-up. Homing switches only operate during homing and missed steps only happen during operation. A homing switch can't prevent missed steps in any way shape or form. A missed step is a commanded step that is commanded and not executed. As stated before, on a laser that does not have cutting forces and only moves the fixed weight of the machine itself, missed steps would mean either undersized steppers, accelerations or feed rates too high for what is there, or electrical interference. All of which are design problems.

Of course missed steps won't be solved by any amount of switches, but to correct misalignment caused by missed steps, extra steps or whatever else you need to square the axis for. That can be done in a few ways, as Gerry pointed out, slowly jogging into accurately aligned hard stops is a budget version one, two home switches is another. Steps can be missed (or extra steps added) because of electrical interference also, and as far as I know, laser heads generate plenty of electrical noise. Of course, all that is a design problem, but if things are perfectly designed you would not need to have a home switch at all because your axes would be perfectly homed and aligned every time. As I mentioned, missed steps can be caused by many things, mechanical as well as electrical problems. The definition is too narrow if it is related only to commands not executed, that is almost never a problem today. Commands can be executed and the motor still may miss steps. Yes, home switches won't solve missing steps, but if you home before each job it is not likely that you will notice any issues in the works you have done. Also, please don't take "missed steps" literally. When you get alignment problems you don't initially know if they are caused by missed steps, extra steps, mechanical or electrical installation or whatever else that might cause misalignment. Only after you have analyzed the issue you have the answer about that.

As for power-up misalignment, the amount of misalignment on a twin belt driven axis will only be as much as the mechanics of the axis will allow. For example, an axis with linear guides can only be misaligned as much as the linear guides will allow because as soon as you power-off the machine, the steppers are free (except for a small amount detent resistance) and the axis will relax to align to the linear guides. Now power back up and the steppers will lock in that position within 1 step. If this amount of misalignment is not acceptable then either fixed stops as mentioned, or individual homing switches can get the axis into better alignment, but only as accurately as the stops or switches are adjusted. If the stops or switches are not aligned properly they can actually skew the axis rather than straighten it.

Of course. But if you start off with a slight misalignment every time then you will get binding issues which might cause stalling or other issues. That's why two home switches, or hard stops as Gerry mentioned, together with homing before each job is necessary if you have two steppers driving two screws and have no position feedback from the motors.

[A_Camera]

It will be interesting to see if the unsupported round shafts are suitable for high speeds / accelerations. Typically they are the worst thing you could do for a router gantry, for a laser, I don't honestly know. Could be awesome. I really don't know.

I think for this sort of application, laser or 3D printer, it is OK to use unsupported rods. Speed and acceleration wise it is definitely not a problem. My current 3D printer uses it and the speed/acceleration is actually amazing considering the tiny NEMA17 motors and 24V PSU. However, the 3D printer I am going to build will have dual ball screws for Z and dual square linear rails on all three axes. Maybe it's overkill but that's fine by me.

Looking at the 2.x videos on youtube, some people used V groove bearings and some upgraded to square linear rails.

I agree that the belt drive is best for this.

For the speed of the gantry, well, it depends, one larger stepper with a design like the 2.x, geared properly so it is spinning at lower rpm's while the gantry is moving fast with the belts could work pretty awesome. Although it is probably a bit more complicated to build. I honestly didn't think about the idea of using a stepper with the shaft coming out of the top as well, and I pictured in my head a convoluted assortment of belts and gears tying both sides together. You'd need a driver that could handle more current for the larger motor.

So what will you do about the gantry homing and the two motors? I don't think the smoothieboard can do it? Are you going to dump it?

Interestingly enough, you may have the same problem with your Z axis, as that is driven by multiple screws.

The quality of the workmanship of what you've done so far looks nice.

109jb, you are correct. On my current CNC build, I am making it so the mechanics won't allow much misalignment, at least that's the theory. I may add some homing to it as well, but not sure if I can do that and keep the 4th axis with the current board that I have. I may need to upgrade that in the future. Just trying to get it built, that's a detail I can fine tune later.

IMO, it doesn't become an issue if you have good mechanics unless you go with large spans. For something like a laser or a printer where the mechanics are more spartan and less rigid then perhaps the issue becomes more important.

I don't think we can compare CNC, laser engraver and 3D printers as being equal, especially CNC are having a bit different loads and requirements. I would not build a CNC the same as the weight and the load on especially the Z, but on all the other axes as well, is totally different. The round rails on any machine are pretty easy to bend a more or less, depending on the fixture and the diameter, but that is not a big issues for 3D printers and laser since there is practically no force. The light weight makes very high speed and acceleration/deceleration possible to reach at low cost, which is not the same (or much more expensive to reach) with a CNC. Backlash on Z in 3D printer is really not a big issue because the weight of the Z is always keeping the nuts at the lowest possible screw level. In a CNC it would be a disaster to have dual Z screws with backlash. But binding caused by misalignment or whatever else mentioned before, can be a bigger issue in a light laser or 3D printer. It can however be an issue also if you have a large CNC, so just because heavy and high quality material is used that's not the same as immunity against binding. I think if I had a dual screw CNC I'd definitely have dual home switches as well.