[tantemay]

I have a problem with circular orbits and diagonal orbits. If the machine is milling a straight line parallel to the X or Y axis, the surface is very good.
However, as soon as a path with G2 / G3 instructions comes up, or Digonal, the surface has facets.
The machine is checked mechanically and OK. It doesn't matter whether it's POM or aluminum, the milling pattern is the same.
I tested different CAM software, always the same result.
My suspicion was that the curve consists of many short G1 instructions, but as I said, there are G2 / 3 instructions that are used in the CAM. There is no difference between G64 / 61.
Now I only suspect two things. Either control software that may convert the G2 / 3 instruction into G1 instructions before milling. Or the cogging moments of the steppers.

[PlanetCNC]

Unless transformations or some other "path changing" options are enabled, software does not interpolate arcs as lines.

[tantemay]

Thank you for your prompt reply. I did not activate any path change options on the control. Respectively. I don't know where to do it.
Then only the steppers are suspected.

[tantemay]

To see in the photo. Straight milling track on the left.
On the right is an oblique milling path.
How do you get the facets away?

[dazp1976]

When you say facets. Are you talking about all those little vertical lines on the cutting edge?
Had that on my sieg. Couldn't really get rid. I put it down to table vibration due to poor rigidity and light weight.

The only way I figure I could have got it stiffer is if I could fit longer double nuts on the C7 ballscrews I had. The singles wobble a bit unless you can fit oversize balls.
Also, using thrust bearings did not help matters either.

[tantemay]

Yes, I mean these lines.
The machine used has double ball nuts.
If it were due to some stiffness, the problem would also be with a straight milling surface. But as you can see it's OK.

[tantemay]

Sorry wrong .jpg

The upper milling path is X0 Y0 to X20 Y2, i.e. an oblique path. This has facets or levels. How do you get that away?
The lower one is X0 Y0 to X20 Y0, i.e. a straight path. This is OK for me.

[PCW_MESA]

Do you have dovetail ways? If so, it might be a stick/slip related problem.

If you have a high resolution indicator, you might try making a series of very small
moves and check how closely the actual machine motion matches the commanded position.

[tantemay]

The machine has 25 linear rails.
On the long axis (650mm) two ball screw drives with double ball nut. On the portal axis (520mm) is a spindle with a double ball screw.
Unfortunately I only have a 1/100 fine probe. With this I can not find any deviation of the instructing and the actual position. The deviation is less than 1/100.
Likewise the concentricity of the milling spindle. The concentricity error is less than 1/100.

[PCW_MESA]

The deviation may be less than 1/100 of a mm.
If these are step drives, are they direct drive, and what
is the ball screw pitch?

As an example with direct drive 5mm/turn ball screws, a full step
(with a normal 200 step/turn motor) is 1/40 of a mm

[tantemay]

The 1.8 ° stepper motors are flanged directly. No straps.
The step resolution is 4000 / 5mm
So 4000 steps per revolution. The spindles have a 5mm pitch

[blowlamp]

I think you'll be getting a kind of 'staircase' pattern because despite having your motors set to micro-step @ 4000 over 5mm, in practice you will probably find they can not dynamically hold these micro-steps precisely and so show up on the part.

[dazp1976]

Are you using 3D program that produces hundreds of small movements?

Could it be done with just a couple of radius arcs instead or is it a bit more of a technical radius?

[tantemay]

The 1.8 ° stepper motors are flanged directly. No straps.
The step resolution is 4000 / 5mm
So 4000 steps per revolution. The spindles have a 5mm pitch

Yes exactly. I already thought that. So that are the 200 locking moments of the stepper. (I don't know how to write that in English)
But why is it just me who has the problem? Or do I complain about something that is normal for others.

[tantemay]

Are you using 3D program that produces hundreds of small movements?

Could it be done with just a couple of radius arcs instead or is it a bit more of a technical radius?

It occurs with a normal G2 or G3 instruction.

[blowlamp]

Full steps and half steps are quite 'rigid', but when you try to go much finer between steps things get a bit 'springy' - they don't lock as tightly as full steps.
You could use finer pitch screws or gear-down the ones you have by using timing belts to the point where you need fewer micro-steps.

[tantemay]

I have already tested lower step resolution, but it has remained the same.
They are close-loop steppers.

[maki]

have you checked the tolerance under Settings-movement-blend for xyz? take a look to manual at page 340 + ...
other thing is in my mind, if you using some other program for generating G-code, there is probably some adjustement hov long shoud the lines-cutves be between points, or in other words,
when the straight line shoud be used between 2 points (the shortest distance between 2 points) and when the arc is implemented. Part of it is a PP for generating G-code.
Hope it will help
ps
DXF can be interpreted in many many short lines, or can be an arcs also, depends of version dxf export - generated...

[tantemay]

Hi. Yes, my last idea was also the blend.
But with the default settings it did not get better even after activating. It wasn't until I changed the parameters.
What values have you set?

[maki]

I still have a defoult settings, 0,5 , but lower than 0,02 is nonsence I think.
What about other question about PP? Or, DXF export etc...