3d milling question

For me if it moves in height uniformly it shouldn’t change anything. On the other hand, if it veils or twists, yes.
I don’t know about the plate but if exact repeatability is not assured then yes we should probably leave the machine on.
I screwed my plate back on and added screws then started the milling with the next cutter and we will see if the problem persists or if that is where it came from.
I prefer to do the test on this failed model than on the next one.

Hi Fred you mentioned that you stopped mailing because of the noise does that mean your machine is outside sometimes it would can expand which will make the bit appear lower

The machine is in a garage which is not thermally or sound-insulated and there is an apartment above.
So if I overdo it I’ll end up not being able to do anything.
So after a certain hour I stop the noise.

Expand yes but that much?
Knowing that there has not been a big variation in temperature these days.

When you probe Z again are you using a spot that has not been milled or altered since the first time you probed Z height? Also, do you turn the block upside down so the flat side sits against the sport you are probing? I can probe Z height repeatedly and the result is within 0.1mm, a 1mm deviation is not typical.

I deliberately leave the reference corner intact in order to always have the same reference.
I generally do prob z with the plate right side up (meaning of prob xyz) but it is possible that I have also done it on occasion with the plate “upside down”
Do you think it could have an impact?

Above all, I think that my problem comes from poor fixing of the work board and that the board has twisted.
I will know tomorrow evening, if when I start milling again the problem persists or not (the plate can no longer move).
Otherwise it was my fault, if so the problem comes from elsewhere.

But I must admit that this is my first two-color milling and that I made mistakes in preparing the board to work on, so before blaming the machine I will re-prepare a board without these errors.

Good afternoon Fred,

In principle it should not, unless it has grooves from unclean milling, but that should be inaccuracies nowhere near the observed range, or if it isn’t milled plane parallel.

It’s easy to get error by this. Fixing the workpiece is a matter of its own and often a cause of inaccuracies.

Track it down by exclusion like Sherlock Holmes :slight_smile:

Very ambitious project!

This thread seems to be bouncing around a bit, but I’d like to share my thoughts on the matter of shutting down the controller versus leaving it on.

When it comes to intricate carving work like inlays and 3D reliefs, I’ve experimented with shutting down or rebooting the controller and attempting to return to the EXACT same spot at a later time. I must admit, achieving accuracy within 0.5mm to 1mm (0.0197" - 0.0394") has proven to be a significant challenge. I believe one of the primary reasons for this difficulty is the process of homing the machine. Even when using limit switches, let alone stall homing, attaining sub-millimeter precision is a tall order.

The closest I’ve seen this level of precision is with one of my 3D printers, which employs an inductive probe for Z-axis bed leveling, achieving a 0.01mm (0.0004") precision and approximately 0.005mm (0.0002") accuracy. Nevertheless, I still find myself fine-tuning the Z offset by about 0.005mm back and forth every dozen or so prints. Even with such a high level of precision, it still remains a formidable challenge to power down the controller and expect the printer to home itself within 0.5mm accuracy and seamlessly pick up where it left off.

As a result, I’ve adopted the practice of leaving my CNC machine on when doing multi-step detail work. The primary concern with leaving the machine on is the constant energization of the stepper motors and their drivers, which may potentially reduce their lifespan. However, I believe this is only a genuine concern over a time frame of weeks or months of continuous power-on operation.

3 Likes

By far the best upgrade to my Onefinity X50 was adding proximity sensors for homing, initially this was using the Onefinity Buildbotics controller. The second best upgrade was moving to a controller that could auto square the Y axis. This provided the precision in the homing process that allows me to stop, power off, and restart a job or to recover a job that lost its work offsets.

2 Likes

Hey all,

but what this thread is about is not powering the controller off or not, but about probing repeatibility. If you re-probe Z, it makes no difference whether the controller was shut off or not in between.

Leaving the controller on is just a matter if you want or need to proceed a program later without reprobing.

And additionally, possibly what the problem boils down to here is actually workpiece attachment

1 Like

It’s better tonight but it’s not there yet.
The cutter is smaller so the offset is shallower.
I don’t understand.
The tray no longer moves; there are 5 screws over 72 cm to hold it.
The offset is just in front of a screw.

Hey Fred,

what did you do between the two passes? Did you change something in workpiece attachment (tighten screws etc.)?

Hi Aiph5u.

Between late Saturday and this evening, nothing at all.

Between the offset of 0.5 mm and that of the photo.
The difference is the cutter used.
For the big shift it’s the one 314.032.8D
For tonight’s one it’s the one
314.016.8D
Cutter details here:
https://www.fraisertools.com/fr/meche-a-spirale-pour-gravure-2d-3d.html

Hey Fred,

thanks for the cutter specification.

Changing the cutter means you re-probed Z, right?

Where did you probe Z?

Yes absolutely I reprobe z for the new cutter in the same place as for the previous cutter.

The new one works with a depth of cut half as great as the previous one which could explain a smaller offset.

I use smaller and smaller cutters to get the details, I still have one later (I don’t know yet if I’m going to use it or not)

Hey Fred,

Are you already milling the final surface height?

Of course you can run a pass with a 1.6 mm radius bit and then the next pass with a 0.8 mm radius bit.

But if you run the exactly same program, or a program that mills down to the exactly same depth on the workpiece, and probing is accurate, the result is that it should not mill deeper. That should be the expected behaviour if your program mills to the same depth on the workpiece as the previous pass.

Or do you run passes that mill deeper than the previous one?

Hey Fred,

I think in this situation I would run a systematic test with a new workpiece. I would test probing Z accuracy of the touch plate when switching between different bits, i.e. increasingly pointier, smaller radius bits.

For this, first I would use one program that mills a workpiece to a specific depth to have a flat surface.

Then I would first use no program for a first test but jog around after reprobing, thereby taking care not to move Z.

When using the XYZ touch plate for probing Z after bit change, you should always have the same workpiece height after milling around, no matter what bit it is.

Keep in mind wood moves naturally. (Swelling and shrinking) Unless you have constant environmental control, your piece may move due to humidity levels, temperature changes and stress relief in the piece of wood. I use to work in a climate controlled environment for that very reason. Now, i know metal and wood are different…but the environment will still impact it.

I leave my controller on till I’m finished a project! It might run 24/7 for a week or more, depending what might distract me. I also wtite down the x,y, and Z numbers in case of power disruption.

Hey Aiph5u.

Yes, I will try to plan something to validate the precision when I have finished this work.
Otherwise I could contact onefinity and ask them for their opinion on this.

Hi gamebird Pete

Indeed, as I talked about previously, these movements could be a reason. > But the different boards were glued according to the rules of the art (which > does not prevent movements) but I have just checked nothing to report.

I just got back and started work again.
The controller was left on last night : no offset

Forgive me if this has already been said, but can’t you just power off the machine, and then re-zero when you turn it back on without homing?