[GuyUlm]

I've been planning this in my head on and off for quite a while. I started this rough draft CAD drawing and the more I research materials and parts, the more I realize I need help getting the best bang for my buck.

I'm planning to build the frame using 1.75" x 4.5" aluminum extruded tubing (anodized storefront material I get as scrap from my job). I want to get it working well as a CNC router to work various materials, up to and including aluminum. I plan to be able to replace the spoil board with a shallow water tray so I can change out the router for a plasma torch.

I have tentatively planned to have the X and Y axis operated by two of these "STEPPERONLINE High Torque Nema 23 CNC Stepper Motor 114mm 425oz.in/3Nm" with this controller "STEPPERONLINE CNC Stepper Motor Driver 1.0-4.2A 20-50VDC 1/128 Micro-Step Resolutions for Nema 17 and 23 Stepper Motor".

I've got an Amazon list of the parts I am seriously considering, https://www.amazon.com/hz/wishlist/l...?ref_=wl_share .

First of all, can I assume that my frame will be sufficiently rigid?

Is my single motor driving two ball screws to move the gantry via belt a good design?

I'd really like to get a conversation going here so that I can start acquiring the necessary parts and begin my build. I welcome constructive criticism and discussion.

(now to see if I can figure out how to imbed images)

[joeavaerage]

Hi,

First of all, can I assume that my frame will be sufficiently rigid?

No. It would be OK for plasma and maybe wood with a router but aluminum......no.

If you used the same material (1.75x4.75) but made it only 18inch x 18inch then you might get it rigid enough to do aluminum. Cutting metals takes a great deal more
rigidity than either a plasma or a wood router.

Is my single motor driving two ball screws to move the gantry via belt a good design?

It doable, but for the cost of another motor and driver I think it a poor choice.

I have tentatively planned to have the X and Y axis operated by two of these "STEPPERONLINE High Torque Nema 23 CNC Stepper Motor 114mm 425oz.in/3Nm" with this controller "STEPPERONLINE CNC Stepper Motor Driver 1.0-4.2A 20-50VDC 1/128 Micro-Step Resolutions for Nema 17 and 23 Stepper Motor".

By in large that sounds OK, but I would up the drivers to 80VDC capable and use an 80VDC power supply. A 50VDC supply and driver is going to be slow, and especially with plasma you'll need speed,
go for 80VDC.

Craig

[peteeng]

Hi Guy - Combining plasma and routing is not a good idea. Try to find a commercial one that does this or even a hobby one and study it well. Each application has speed and rigidity requirements that are quite different. Plus plasma is dirty and timber cutting needs to be clean. So think it through a bit more and find a good example to copy or at least have a muse over. Peter

[GuyUlm]

Thank you so much for the feedback!

When I say that I want to work with aluminum, I don't mean precision machining. I'm referring more to cutting material that might be 1/8" thick for making templates and custom plates for work. I understand the rigidity required for milling a billet into something usable.

I thought that if I went with the single motor for the gantry, I'd eliminate any issues with syncing two different motors. With the motor I listed, wouldn't it have sufficient torque? Is syncing two motors easier than what I've read?

[joeavaerage]

Hi,

When I say that I want to work with aluminum, I don't mean precision machining. I'm referring more to cutting material that might be 1/8" thick for making templates and custom plates for work. I understand the rigidity required for milling a billet into something usable.

You're not the first person to suggest that because its only light work it does not need to be rigid......and its just not correct. When a rotating tool engages a piece of aluminum with sufficient force to 'peel' off a chip
there will be a reactive force. The machine needs to contain that force WITHOUT DEFLECTING beyond your accuracy tolerance. It does not matter whether that aluminum is part of a billet or a 1/8th sheet,
the reactive force is the same, and the machine must contain that force no matter the material.

I thought that if I went with the single motor for the gantry, I'd eliminate any issues with syncing two different motors. With the motor I listed, wouldn't it have sufficient torque? Is syncing two motors easier than what I've read?

99% of all gantries have two motors, Synching two motors is possible, which is not to say the same as easy. I suspect your one motor will not have enough torque, so you'd be up for two motors anyway. If you are contemplating using
some of the same aluminum section for the gantry, you are going to be severely disappointed, its just too flexible.

Read some of peteeng build threads and you'll see how seriously you need to consider rigidity.

Craig

[GuyUlm]

Thank you for the response.

I understand that plasma requires very little rigidity, other than to accommodate speed. But I'm not sure I follow the dirty vs. clean comparison. My design would make it a process to change from one to another. One plasma-related concern I do have is in reference to the roller guides. Wouldn't they be prone to getting spatter on them and GREATLY reduce their performance or kill them entirely? I do know that there are interference issues with plasma cutters that have high-frequency arc start.

[GuyUlm]

Point taken on the aluminum rigidity.

Is significant torque sacrificed in belt material?

If in doubt, should motors drive screws at 1:1 ratio or better to change the ratio to give a torque advantage? Should they just be direct drive where possible?

My plan for the gantry is to use two pieces of 1.5" x 3" ribbed extrusion and some right-angle pieces of 3/16" aluminum extrusion.

I'll be looking around for designs on the forum. I haven't figured out how to find any galleries yet.

I really appreciate your feedback.

[peteeng]

Hi Guy - a 4x5' machine is getting big for a single drive. There is no mechanical issue having the central motor and two belts. By the time you make mounts for pulleys, buy belts etc your probably at the same cost as two motors. Steppers are cheap and CNC software is set up for the extra axis anyway. Plasma is a dirty process, have a look at some of the machines and you will see. This spatter and muck will contaminate your timber work. As I said find a machine that does both and see how/if they work. Cutting 3mm aluminium or 25mm aluminium requires the same rigidity. There are small routers that do this but they cut say at 1mm deep and do many many runs to get there. Not very efficient and costly in bits.

How thick is the 1.75x4.5"? you could rivet or bond them together to improve stiffness. Plus you need to publish a sketch or model of what your going to do. A correctly sized belt does not lose torque. Plus a plasma needs rigidity to cope with cornering loads. There's heaps of plasma discussion about wobbly cutting heads especially at right angle cuts. Again find a machine that does what you think you need and copy it. Peter

screw drives are sized by their critical vibrations. You need 16mm or 20mm screws for this size machine. The rpm of the screw defines the speed it vibrates at so you have to stay under that speed. Since a stepper runs out of puff around 600rpm thats your upper limit for vibration. Plasmas last time I looked require high speeds and I think you may want to look at belts for the machine vs ballscrews or rack and pinion. Peter

[joeavaerage]

Hi Guy,
may I suggest an experiment:

Get a length, say 2m, of the aluminum material that you want to use and have access to. Put each end on a brick or something so that its suspended a few inches above a flat floor.
Now place a 1kg weight in the centre of it and measure how much it deflects.

A 1kg weight is a 10N force and is small beer in comparison to cutting forces but the deflection you measure will be about as good accuracy as you might expect if you use
this material. If it deflects less than 1mm I'd be surprised, if it deflected by 2mm to 3mm I'd say....yes that sounds about right. Can you afford to have your machine deflect by that
much in response to a very modest force? Ideally it would deflect 0.01mm or less.....but well that would be ideal, and we don't live in an ideal world.

Craig

[GuyUlm]

How thick is the 1.75x4.5"?

- It's 1/8" (3mm).

you could rivet or bond them together to improve stiffness.

- I plan to use aluminum channel in a way that, in the industry in which I work, is referred to as a shear block. It allows for holes to be prepped in the material and a countersink with the corresponding holes slightly offset so that when the beveled-head screws are sunk into place, it creates a degree of clamping tension that pulls the cleanly cut end firmly into the adjacent part.

Plus you need to publish a sketch or model of what your going to do.

- Are my drawings in the original post above not displaying?

screw drives are sized by their critical vibrations. You need 16mm or 20mm screws for this size machine. The rpm of the screw defines the speed it vibrates at so you have to stay under that speed.

- I'm not sure if the Amazon link in my original post works, but the screws I have currently selected are 16mm (SFU1605). The one on the gantry is 1200mm and the pair moving the gantry are 1500mm. I'm exploring bumping the 1500mm screws up to SFU2005 because of their length.

[joeavaerage]

Hi,
just run a 3mm thick 6061 section of 1.75inch x 4.5inch through FEA for1.6mm deflection of a 1m cantilever with a 100N load in 'broad direction' and 0.35mm deflection
with the same 100N load in the 'narrow direction'.

This results in a 1m cantilever having a stiffness of 0.0625N/um in one direction and about 0.286N/um in the other direction.

Provided you do not make the lengths excessive this would be marginal for a wood router or other soft materials but fall well short of the stiffness you would need to cut aluminum effectively.

Craig

[GuyUlm]

Based on your previous response suggesting an experiment, I'd guess the tubing I plan to use is considerably rigid. I plan to test my theory possibly tomorrow.

I don't plan to have any piece of my structure longer than the two side rails at 72" (~183cm). Those won't actually be an unsupported span as the legs coming off the bottom will be inset from each end by ~30cm, leaving the actual span being just over 102cm

Are the images I initially posted not visible?

Guy

[joeavaerage]

Hi,

Are the images I initially posted not visible?

No they are not.

I'd guess the tubing I plan to use is considerably rigid. I plan to test my theory possibly tomorrow.

I'm sorry to say but I believe your aluminum sections are rather too flexible. If you imagine the legs are going to hold the rails straight then the rails are too damned flexible.
It should be the other way around, the rails should be that stiff that the legs can attach to the rails without altering the straightness of the rails in any way.
The rails down either side of the machine need to be UBER stiff, and I think your aluminum sections are marginal at best.

By all means proceed with the experiment, it should give results not disimliar to the Fusion FEA analysis.

Craig

[burs]

Hi,

Are the images I initially posted not visible?

No they are not.

They are now. Sorry.

[peteeng]

Hi Guy - Since your getting the material for free and you seem to know how to bolt things together properly develop the design as far as you can. CAD work is free but once you get to the real world things become difficult. With twin gantry designs most people eventually connect the two together with a brace along the back. Your columns will need stiffening on the outside as well. The ballscrew on the gantry will be able to push with say 150kgf and this will bend your columns sideways... What size and tooth spec belt do you envisage? belts need to be tight to be good so you will need a tensioner. Using your hand to tension the motor then doing up screws is not enough. Especially if its a 16 or 20mm or bigger belt. I'd suggest some sort of 10mm bolt or 8mm to pull the belt up will be needed. Good belt tension is at abpout the limit od the strength of the motor shaft so pick one with a big shaft diameter,10mm is good don't use 6mm or 1/4" you will fatigue and break them quickly . The small dia motors are fine for direct drive thru a coupling. Peter

re" bolting things together . Using a csk is fine to create firm connections but machines need a little alignment and that sort of connection is too firm. Just use a couple of extra bolts and they will be fine in a snug hole. Say a 6 or 8mm bolt usually uses a 0.5mm oversize hole. So make then 0.3 or 0.25mm oversize eg 8.25mm or 6.25mm. This gives you a small amount of movement to align things. Friction in an aluminium connection is higher than a steel connection. I use a wicking loctite to set parts once the position is final. This means you don't have to pull the connection apart to set it. Peter loctite 290

[peteeng]

Hi Guy - Regarding screw pitch. 5mm is too small you need 10mm or 20mm pitch. 5mm will give you tonnes of force but slow speed. at 500rpm this is 2500mm/min too slow for a plasma and about right for cutting timber and aluminium,. So you would have no top end left and the screw will be spinning really fast. Then there's 20mm dia. At 20mm you will be using more than half the torque to spin the screw,. all the more reason for one motor per screw. Ballscrews are very high inertias compared to the motor inertia, size them too big and your machine runs really slow... Same with N23 vs N34 motors. The N34s have lots of inertia and a smaller long N23 will spin faster with good torque left over.... Peter

you do need to research plasma operational speeds. If you use the wrong screws you may not be able to use the plasma!!

[GuyUlm]

Hi Guy,
may I suggest an experiment:

Get a length, say 2m, of the aluminum material that you want to use and have access to. Put each end on a brick or something so that its suspended a few inches above a flat floor.
Now place a 1kg weight in the centre of it and measure how much it deflects.

The piece I used is 72" (~183cm) and has clips installed at the ends, shortening the span by nearly 10cm. The only readily available weight that I had was a piece of thick-walled steel pipe that weighed 7.71kg. Using what I had to measure, it appeared to deflect approximately .75mm.

This is with the rail laying flat instead of on edge as I plan to have it. I'm sure the different orientation would yield less than half the .75mm deflection.

[GuyUlm]

I'm sorry to say but I believe your aluminum sections are rather too flexible. If you imagine the legs are going to hold the rails straight then the rails are too damned flexible.
It should be the other way around, the rails should be that stiff that the legs can attach to the rails without altering the straightness of the rails in any way.

I wasn't implying that the legs would stiffen the side rails, only that because the legs are attached to the rail it shortens the unsupported span. Deflection in this context is directly related to unsupported spans. The material I am using comes in ~7.3m stock lengths. If I had only each end of a stock length blocked up and applied several kg, there would be visible deflection if the material were lying flat. But if I moved the blocks in 1 meter from either end, the unsupported span would be reduced as would the degree of deflection. Agreed?

[joeavaerage]

Hi,
that weight is about 70N force. The span is about 1700mm. The deflection is about .75mm, approx half of what I calculated for a 1m cantilever....so given that your experiment
is supported at both ends then I'd expect half the deflection. Within a 'broad brushstroke' then the FEA analysis compares favourably with your experiment.

The question is that if your machine when operating experiences a similar 70N force as a result of cutting forces and/or axis acceleration it will deflect by 0.75mm.
Do you think that is adequate? Lets say you are cutting a straight line and yet there is some disturbance that equates to a 70N force, the line will no longer be straight
but have a 0.75mm deviation???

This is the problem with long lengths of aluminum section....they just deflect too much for a given force with subsequent inaccuracy of the work piece.

Deflection is inversely proportional to the cube of length, so if you made it half as long its deflection would reduce eightfold, so from 0.75mm to 0.1mm. Would 0.1mm be a better accuracy?
As the size of the machine goes up the standard of the parts required goes up eight times for a doubling in size.

If you want to build a rigid and accurate machine with modest materials build small. Aside from anything else it will save you a ton of money. Even more importantly you'll learn
a huge amount from your first build that by the time you go to design your second it will be vastly better because of your experience.

Craig

[GuyUlm]

With twin gantry designs most people eventually connect the two together with a brace along the back. Your columns will need stiffening on the outside as well. The ballscrew on the gantry will be able to push with say 150kgf and this will bend your columns sideways.

The columns I have for the gantry are going to be made of 2 each end 4"x4" aluminum angle that is 3/16" thick. I really can't imagine the force being great enough for the gantry to deflect. If it does, I'll fabricate reinforcements.

What size and tooth spec belt do you envisage?

I'm not sure, what is ideal?
I actually have access to belt material. I work on automatic doors for a living and sliding doors are almost always belt-driven. Therefore I also have access to drive pulleys, tensioners, and idler pulleys.

re" bolting things together . Using a csk is fine to create firm connections but machines need a little alignment and that sort of connection is too firm. Just use a couple of extra bolts and they will be fine in a snug hole. Say a 6 or 8mm bolt usually uses a 0.5mm oversize hole. So make then 0.3 or 0.25mm oversize eg 8.25mm or 6.25mm. This gives you a small amount of movement to align things. Friction in an aluminium connection is higher than a steel connection. I use a wicking loctite to set parts once the position is final. This means you don't have to pull the connection apart to set it. Peter loctite 290

I had to look up what "csk" meant. I am only planning to use that method of fastening when building the structure. Everything else will likely be some kind of machine thread bolts/screws. I completely understand your point about using slightly oversized holes and then clamping force. Also very good to know about the wicking locktite. Not something I knew existed.