Designed a router with high Z axis.

Hi all!

I am designing a new cnc router that is designed mostly for larger 3d foam objects and some soft wood projects. I build a small table top router once and I am not sure if I have designed this router well.
I am doubting if I need some updates on my design or some "downgrades" if something is overkill.
I would love to get some feedback so this project would not be a disappointment when I finish the build.

Specifications:

Travel (Work Area): X Axis 1220mm (48") / Y Axis 1220mm (48") / Z Axis 430mm (~17")
Workable Material Height: 405mm (~16") without 4th axis, when rotary axis added to router: 800mm (31.5")
Calculated max cutting speed: 9525 mm/min (375ipm)
Calculated max rapids: 17018 mm/min (670ipm)

Important materials:
Frame: 80x 80x 3mm (3x 3x 3/32") square steel tubes
Ball screws (X&Y): diameter 20mm (0.79") pitch 20mm(0.79") length 1400mm (55")
Linear rails: hgr20 with two carriages per rail
Stepper: 4x nema34 12N (1700 oz.)
Drivers: HBS86H
Spindle: 2.2kw water-cooled Chinese one.

The biggest issue you are going to have is with the extrusion the spindle is mounted to. When it's all the way down, it will have a lot of flex. May be OK for foam, but not much else. At a minimum, add some side panels, maybe 100mm deep and 10mm thick? Or a much bigger extrusion.

If it were me, I'd tuck all the motors out of the way by flipping them and using a belt to drive the screws.

Hi Jeff - Your table for foam and timber is quite good, but as Gerry says you will need to work on the Z it needs to be much stiffer then designed otherwise your tool will skip around in timber.

Most of the time with this sort of design you are using it with the Z extended so its not ideal most of the time, Think about having "pits" or bolsters for most of the time so you are using the machine with a short Z. Only use the table or pit for when you actually need the high Z.

You have mounted the rails in the middle of sections, this is the least stiff position. Move them over to an edge to improve local stiffness. Plus for the tubes the rails are mounted on use very thick sections, the "frame" ones can be quite thin but anything that has load transfer has to be thicker then you think. 1) for local stiffness 2) for vibration 3) to minimise distortion when you weld it 4) for thread depth. Ideally you need to have the frame stress relived and the rail lands machined...Peter

with a frame like this it is logical to sub divide sections evenly but unfortunately in terms of vibration this is a poor strategy because you have just created a natural freq amplifier. So sub divide sections unevenly and put in some triangles its harder to get triangles excited then square shapes...

You won't need NEMA 34s and N34 will be slower then N23s... You have base your rapids on getting the stepper to 850rpm this is unlikely as torque is running out at that speed. Do you have a torque chart for the motor to prove you have usable torque at that speed? A 20mm ballscrew has lots of inertia and this will be the dominant factor in getting up to speed at a reasonable acceleration.

N34 have very large inertias and large inductance, both of these go against accelerations and high speeds. I just looked up 3 N34s and they don't have torque past 650rpm so have a think about that...

Foam and timber usually require fast speeds (unless your happy with long cycle times. Because of the high Z I'm guessing you want to do curvy high things more then sheet work ) and maybe you should consider cheap servos vs steppers. Full torque to 3000rpm is a great thing... leave some steel out and put the $$$ into servos.

Thank you (and ger21) for the clear feedback.

I am think to raise the table in some way so the z axis will not be extended if not needed but most of the time it will be used for larger objects so I need to think how I will improve stiffness. I think I would replace the extrusion for solid steel plates.

I will place the rails to a edge. How thick the steel needs to be for the frame? And I will redesign the frame with more unevenly spaced sections and triangles.

I added the torque curve for the NEMA 34s, the red line is for the ones I would like to use (closed loop stepper).
I also searched for servo motors but I find it really hard to find a calculator how large of a servo I need for decent speed. As you mentioned, I am looking for fast speeds because of the 3d application.

Hi Jeff:
1) steel thickness at least the same or thicker then the nut required. Plus I'd go even thicker so its stiffer can't be too thick
2) The calculation for steppers and servos is exactly the same. The torque required is in two parts a) for cutting which depends on your material and tooling and b) your acceleration required. Acceleration is the usual hurdle and your 20mm ballscrews and N34 motors will have lots of inertia. Unless you have an acceleration target you can't calculate the torque required to turn the system.... Peter

Whenever somebody comes up with a scheme for a tall 3 axis machine,I have to wonder if they have visualised the factors that might limit the usefulness of the machine.You will be unable to machine a plane into the vertical sides of a large body at a low Z height because of the diameter of the spindle body and associated clamps.You will be unable to do the same thing at a depth from the highest flat surface that exceeds the tool length and you will be unable to machine a planar slope with an angle exceeding the angle between the tool tip and either the spindle body of the corner of the backplate.I hope the attached image is helpful.

Attachment 457380

Even with a 5 axis machine I have had to tilt the head 35 degrees to get down the side of a project.

Routalot, for foam applications there are special large bits. That is one of the reasons why I designed the router with enough space in the middle.

For example, the manufacturer Streamline Automation with the brand FROG3D build a mill designed for larger foam applications and does also sell large foam bits.
When I have build the machine I will test the bits from Ricocnc

I wish you success with your project.I also think you will need to increase the lateral separation of the bearings carrying the Z axis.As peteeng points out you will need fairly high feed rates for foam and the leverage imposed by super long cutters will be racking the Z axis in two planes.Incidentally,80% of the length of the cutting flutes will never do any work.If you could find alternatives with a similar overall length but with only a third of the flute length and a waisted section above the flutes,they may well cost you less in the first instance and less to sharpen.

Quote:

---

Originally Posted by routalot

I wish you success with your project.I also think you will need to increase the lateral separation of the bearings carrying the Z axis.As peteeng points out you will need fairly high feed rates for foam and the leverage imposed by super long cutters will be racking the Z axis in two planes.Incidentally,80% of the length of the cutting flutes will never do any work.If you could find alternatives with a similar overall length but with only a third of the flute length and a waisted section above the flutes,they may well cost you less in the first instance and less to sharpen.

---

If he is only cutting foam then almost anything will work so he will be fine, with what he has not perfect, so if he needs to upgrade the Z axis at any time it is easy to do, to give you an idea of all that is needed to mill foam have attached a snip of a foam router

Looking at your attached CNC jpgs, I have some suggestions:

- Avoid the motor mounted vertically above the Z leadscrew. At that height the top-heavy axis will sway like the Tower of Pisa (I know this from experience, my first CNC was had a 2ft tall Z-axis)
- Instead mount motor behind the Z-axis upward and use 2x gears to drive the screw.
- The spindle looks mounted to wood against rails in your designs. Replace that with aluminum. It will be lighter and thinner at that size. Goal is to reduce weight and bulk on Z-axis as much as possible.
- Keep in mind the extra weight. You may need a stronger/thicker ballscrew and bottom reinforcement on the screw to prevent the axis falling.
- Allow the 4x pillow blocks to be optionally closer or farther apart vertically. Easily adjusted with multiple mount holes on the spindle mounting plate. This let's you trade off stability vs cut travel.
- The most stable design is when the 16" tool tip just barely clears the bottom of Z-axis structures in the fully up position. That way the work is as close to the machine as possible.

Instead of mounting the Z axis rails to the Y axis plate, mount the rails to the Z axis/spindle plate and put the bearing blocks on the Y axis plate.

The way you have drawn it will always have low rigidity, even when the Z axis is fully retracted. If you put the bearing blocks on the Y plate and have the rails on the spindle/Z plate then when the Z axis is retracted you will have the tool very close to the bearing blocks, meaning MUCH higher rigidity.

It seems like a little wasted effort if the plan is to make a welded steel frame but then use a flimsy long aluminum t-slot Z axis.

If it's just for foam you may as well save a lot of time and make the whole thing out of aluminum t-slot extrusions.

One thing I learnt (the hard way) is that the weakest link will dictate performance. I.e. The stiffest frame in the world won't matter if the z axis is flimsy.

How long are the end mills (max) that you'd use for foam?

I built my machine with a high Z so it might accommodate a 4th axis in future but the usable Z axis travel on a 3 axis machine is likely less than 6 inches. Less than 3" if stiffness is required.

That design would be nice for a 5 axis machine though. If it's just for foam you might even get away with using a cheap Chinese 4th and 5th axis.