CNC advice please

[user 43593]

I can't argue because I don't have any data either

I cannot fault you for that. When the holidays are over, perhaps a mechanical engineer may see this segment and help out. Meanwhile, Cheers!

 

[user 43593]

As an aside, this is my dream milling machine...

 

[bilboburgler]

I'll do the maths later but I did the maths 40 years ago when I was asked to design a piston ring turning machine (in open profile not closed) since piston rings are (or were) mainly cast iron and very very precise this machine was pretty big and I needed to drive a very stiff little carriage holding the cutter blade. The rotary inertia was the critical calculation and made up far more than the simple weight inertia. Back then (big stepping motors were a thing of the future) so you needed large control motors with optical feedback so there was a danger that every thing just got bigger and bigger, but we found a compromise. I suspect that is why the Chinese version of the Open Source has stuck with 10mm ballscrews rather than 16mm ballscrews rather than their nonsense explanation which frankly left me bemused.

Anyway, thanks for this additional discussion. Today is going to be a fine weather day so I'll be out, but the storms come back tomorrow so I can have a serious read.

Rotary inertia is to the radius to the fourth from memory (you'll find google tells you to the square but remember that mass is also to the square and square x square is to the fourth)

Just one word about steel. Youngs Modulus for steel is 210, it varies a bit with alloy variation but unless you are buying some amazing very expensive stuff it is 210. Youngs Modulus is the number that physics uses to tell you how stiff a piece of steel is. Hardness is something else. A harder piece of steel is not any stiffer than a softer piece of steel (excluding alloy alternatives as above), hardness nearly always comes from treatments again nearly always heat based and or chemical based (see also alloys).

 

[worn thumbs]

Those who have been with this thread since its inception will understand why I feel a bit vindicated.If anybody is contemplating self build this may be of interest: SFU1605 BallScrew with Motor Mount Bracket +BK/BF12 End Support +Ballnut Housing | eBay

 

[user 43593]

I'll do the maths later... &c.

Oh goody! A real engineer. Looking forward to your explanations for the thick ones among us.

 

[user 43593]

Those who have been with this thread since its inception will understand why I feel a bit vindicated.If anybody is contemplating self build this may be of interest: SFU1605 BallScrew with Motor Mount Bracket +BK/BF12 End Support +Ballnut Housing | eBay

I would like to consider self build... maybe later in 2024

 

[sploo]

I'll do the maths later but I did the maths 40 years ago when I was asked to design a piston ring turning machine (in open profile not closed) since piston rings are (or were) mainly cast iron and very very precise this machine was pretty big and I needed to drive a very stiff little carriage holding the cutter blade. The rotary inertia was the critical calculation and made up far more than the simple weight inertia. Back then (big stepping motors were a thing of the future) so you needed large control motors with optical feedback so there was a danger that every thing just got bigger and bigger, but we found a compromise. I suspect that is why the Chinese version of the Open Source has stuck with 10mm ballscrews rather than 16mm ballscrews rather than their nonsense explanation which frankly left me bemused.

Anyway, thanks for this additional discussion. Today is going to be a fine weather day so I'll be out, but the storms come back tomorrow so I can have a serious read.

Rotary inertia is to the radius to the fourth from memory (you'll find google tells you to the square but remember that mass is also to the square and square x square is to the fourth)

Just one word about steel. Youngs Modulus for steel is 210, it varies a bit with alloy variation but unless you are buying some amazing very expensive stuff it is 210. Youngs Modulus is the number that physics uses to tell you how stiff a piece of steel is. Hardness is something else. A harder piece of steel is not any stiffer than a softer piece of steel (excluding alloy alternatives as above), hardness nearly always comes from treatments again nearly always heat based and or chemical based (see also alloys).

That was the issue for me: I can look up the formula for calculating inertia of the various parts of a system, but it doesn't mean I'd understand the true impact of the values, or whether I've considered everything relevant to a gantry style CNC machine. There are people on the CNCZone forum who I've seen put some of these sorts of values into posts, so I suspect there are people "in the know" there.

 

[bilboburgler]

I've assumed that you have four axis of movements, the two parallel ones Y and A bolted onto a bench, they drive a bridge axis with a single X axis in it and Z offers a relatively short vertical axis.

from memory and after Christmas cake I think it works like this
it uses Newton's second law of movement and there is a rotational version of this so
The torque=ball screw pitchxefficiency of slidexmass movedxgravity/2pi + Inertia of the ballscrew x the angular acceleration. The other things that have to be taken into account is the chance of whirl of the ball screw and the Euler collapse of the ball screw. I'm guessing you can reduce the chance of Euler collapse by pretensioning the ballscrew so it never sees compression. Removing whirl will probably be by ensuring you rotate concentrically and avoid any critical speed of rotation. (don't go too fast)

Now some designers have gone with 10mm diameter and 10 pitch while others have gone with 16mm diameter and 5 mm pitch. The second moment of area "I" is a calculation to the fourth of the diameter so by choosing the 10mm diameter the inertia load of the ballscrew will be 6 or 7 times less than the 16mm one, but the slideway mass (which for the Y and A axis is going to the whole X axis plus Z axis and router) will be 2.5 times more, somewhere along the design/negotiation there will be a real benefit in using a small ballscrew. You would need to see what the torque/rotational speed of the axis motors graph is, the good news is that for Y and A axis you get two motors, each of which have a ballscrew to turn but they get to share the solid mass as mentioned above.

The X axis only has the one ball screw, the mass of the Z carriage and the router to move around so if your axis motor and ballscrew works for the Y and A it will be fine for the X.

Now I guess that brings us to the question, which should be longer the Y/A axis or the X axis my gut continues to think the Y/A because you can sub-support the Y/A structure with a large stiff table while the X is only ever be a simple bridge supported at each end. Thoughts?

 

[user 43593]

from memory and after Christmas cake I think it works like this...

I see. Thank you. I can see that I made a fundamental error in wanting to determine the inertia of a belt on its own. I thought an inertial moment could be determined from the mass of the belt relating to its mass and its inherent resistance to movement based purely upon its mass. I misundertsood the term inertia. Inertia will derive from the stepper motor being used to move the belt. Thank you for your explanation. I may not be much wiser but I now think I have a better grasp of the subject.

 

[BucksDad]

Lot of discussion on self build here... has anyone got a YouTube link or forum link to a definitive build guide for the recommended options - heavy frame, linear rails, water cooled spindle with VFD as I understand it?

 

[Ollie78]

Lot of discussion on self build here... has anyone got a YouTube link or forum link to a definitive build guide for the recommended options - heavy frame, linear rails, water cooled spindle with VFD as I understand it?

Trouble is there are so many options. So definitive is a bit tricky.
There are many ways to build a machine but when you look at a lot of different ones certain common methods become clear.

https://www.cnczone.com/forums/diy-cnc-router-table-machines/And
https://www.mycncuk.com/forums/253-DIY-Router-Build-Logs
Ollie

 

[bilboburgler]

I've been playing with ideas during the day and calculating why some suppliers use 1604 ballscrews and other 1210.

Then I found these guys LEAD 1515, HGH15, SFU1604 Ball Screw

 

[Ollie78]

I've been playing with ideas during the day and calculating why some suppliers use 1604 ballscrews and other 1210.

Then I found these guys LEAD 1515, HGH15, SFU1604 Ball Screw

It is likely a function of matching the screw nicely to the torque and speed of the stepper or servo. So could be chosen in either order. If you find a good deal on screws of a certain type then get motors that match or vise versa.

 

[bilboburgler]

It is likely a function of matching the screw nicely to the torque and speed of the stepper or servo. So could be chosen in either order. If you find a good deal on screws of a certain type then get motors that match or vise versa.

it is a bit more complicated than that, it is a negotiation between cost, critical ballscrew rotational speed, top acceleration required, X and Z mass being driven, axis motor characteristics, quality of ballscrew set up. This is born out when you read home-builder's blogs as they keep running into these variables. For a private individual remaking is just another cost. For a Maker Space that would be club money I was spending.

 

[worn thumbs]

I think you are over thinking the hardware side of things.The loads are extremely varied and you need to take account of the force required to push the cutter through the job.You also need to take account of the frequent need for all three axes to move simultaneously.If you get it wrong the machine will perhaps stall and that isn't good but with a decent ballscrew and a fairly modest NEMA 23 stepper you will be generating surprising levels of force.Before I upgraded my phone some time back I had an app that would calculate the force applied and from memory it was something like 800lbs-which sort of tells us which country the developer was based in.The other element to consider is the power supply to the steppers and the voltage it delivers.Being unfamiliar with the world of such things,I simply bought a package deal from Long's Motor on ebay and it has never given any concern.I know it operates at 36Volts and the motors are matched to it.I don't get too concerned about maximum speed or acceleration because for my tiny work envelope there isn't the space for a great deal of acceleration or ultimate speed.I have used commercial machines that had a rapid rate of 38 metres/minute and I will only see a small percentage of that.

Should anybody doubt that all three axes may need to move at once,I would ask you to consider the situation where you select a ramp entry cut that isn't aligned with one of the orthographic axes.As you may have guessed,I use ramp entry cuts for most jobs.I know that the weak link in my system is the router that turns the cutter and one day it may well be upgraded as may the entire machine.

 

[Ollie78]

it is a bit more complicated than that, it is a negotiation between cost, critical ballscrew rotational speed, top acceleration required, X and Z mass being driven, axis motor characteristics, quality of ballscrew set up. This is born out when you read home-builder's blogs as they keep running into these variables. For a private individual remaking is just another cost. For a Maker Space that would be club money I was spending.

I agree with worn thumbs about the overthinking.
I suspect that the commonly used ballscrews and stepper motors are likely to be fine for what you are going to need.
Look at the specs on well built machines for sale and that others have built that are the size you want to build and use this as a guide.
Most common is 16mm c7 ballscrews on most smaller machines.
I think my machine was made of whatever they could find as I have 25mm ballscrews on 2 axis and 16mm on z with nema 34 motors on all. Also has some 15mm hiwin rails and some 20mm.

 

[sploo]

My info is pretty dated; but I recall that it was common to oversize stepper motors in order to (hopefully) avoid any chance of lost steps under normal/expected loads. However, there are now quite a few closed loop stepper systems available (I assume that they're just a stepper + encoder; such as AMT Modular Encoders | CUI Devices. In the same way that my machine has servos (which are effectively just DC motors with encoders).

In terms of sizing components though; there are so many commercial and DIY models around that I assume it should be possible to find a machine that's close to what you're after, and take the specs from that (on the assumption that it's probably appropriately designed).

 

[worn thumbs]

I don't know enough about closed loop steppers to give an informed comment but on either www.mycncuk.com or www.cnczone.com there was a post a few months ago that said you might as well stick with open loop steppers.If you have been aggressive enough with feed rates and cut depths to lose steps,how will you know?My own experience is that you are more likely to snap a cutter and I've only done that once and it was a tiny cutter.
As for sizing steppers,you need to be certain that you are comparing like with like.there is a big difference between a stepper coupled directly to a ballscrew and a stepper driving a toothed belt reduction gear system.I need to be convinced that a NEMA 34 is needed for a modest sized machine since I used to use a couple of commercial machines that had NEMA 23's but one had a planetary gearbox in the system and the other had a toothed belt system.One was an 8'X4' and the other a 3M X 2M.The larger output shaft from a NEMA 34 is probably a nice thing to have.I have a feeling that a "weak link" in the drive system is probably a comfort if it can act as a mechanical fuse and break before serious damage occurs.

 

[bilboburgler]

Thanks, overthinking is what a Professional Engineer does. Yes, I've looked at different machines in the same "class" and noting any issues to do with the quality of the cut etc. I too think the H15 slideway will be good enough. In the past I've used 50mm and 30mm ball screws on aluminium and cast iron industrial machine tools and we don't need them here, if, if we pre-tension the ballscrews otherwise they will flex and or go into Euler or whirl excessively either way I'm going 1605 or 1210 and I'm leaning towards 12.

In terms of cutter driver I'm leaning towards Milling motors, from AMB, Suhner or Mafell as that reduces the weight of the Z axis significantly, it also reduces the diameter of the cutter driver (43mm) plays 60 or 80mm which will reduce the torsion load on the X axis, while simplifying the cable tidy system and further reduces the weight and complexity, no water, no leaks.

What I'm hoping is to find some C aluminium channel rather than just 40x80 (any ideas). Ah, self solved ebay.co.uk

CNC-PLANET​

I also aim to maximise the chances of success with the initial design but allow for upgrades, eg add on a 4th axis, have drop out T plates on the bed to allow product to come up from below and to the be space for the 4th axis. I'm going to try a rough BOM costing next and then see about CAD work with the key components.

We may end up just buying one of the standard solutions, but Maker Space people like to make so this is my make project ;-) I couldn't have gotten to this stage without all your sage advice, thank you.

 

[Ollie78]

For your gantry you could do what I have which is two sections of 45 × 90mm extrusion bolted together in an L shape. This is very strong and stiff. I have one linear rail on top and one under, the L faces to the back and the ballscrew runs in the space, if that makes sense, keeps the ballscrew out of the way.

For your 4th axis plan I have actually commissioned a machine to be made for me that incorporates the 4th axis on the end.
The gantry is designed to pass over the end of the bed where the 4th axis is set up.
This will also give me a place to vertically clamp workpieces to joint the ends of them.
This means you dont have to disturb the bed to do rotary work or the ends of table legs etc.
I can pm you a drawing if you want.

Ollie