Zero on Surface of stock?

[bric...@gmail.com]

I'm looking to use my V2 as a very precise traditional 3 axis CNC, as this machine is just the ticket for micro-milling.

To that end, I'd like to set the Z-Zero to the surface of the flat bar stock I have mounted in a fixture. The surface is prefaced, and exactly parallel to the B-plate.

In Fusion 360, instead of the "Magic .8331", I've set my WCS to the surface of my stock (just like we do with every other CNC machine), and have generated a valid toolpath.

Now the trouble comes in: My tool length offsets (TLOs) are natively set from my machine's built-in B plate offset (in my case .8331").

This does not help my tools, as they tend to want to cut in the wrong places - either too high, or heaven help us, too deep.

I've resorted to loading each tool op as a separate program, and as it just starts to run, I stop the program in the Web interface, and manually jog the tip of the tool to the surface of my stock. I then set zero the G54 Z Offset, and rerun the program. This works, but I have to do it at every tool change which is a drag.

Is there any way to set the Z zero ONCE, and then have each measured tool length offset be an automatic offset value from this new zero?

BTW - My stock is of non-uniform thickness. Sometimes it is .495" and other times .496" or .498", so I can't just change it in some Linux CNC server var.

- Perplexed in Peoria

[Bob Vawter]

My machine is currently on-order, but I've been doing some research to plan out how to run it.  I might be completely and utterly wrong, but it looks like it might be possible to use a non-machine-WCS in your g-code.

The latest version of the Fusion 360 post has a new option to "Use TCP mode" (Tool Center Point) which wraps your part geometry in a G60 mode.  If I'm understanding this https://github.com/PocketNC/Settings/tree/master/features/five_axis_kinematics correctly, that should make the machine movements relative to your WCS zero by working backwards into machine-coordinate space.

If I post out a sample part with TCP enabled and use a part- or stock-relative WCS, I get what looks like nearly the right toolpath in the simulator.  I say nearly, because it looks like it's using the machine zero as there's no way to set a WCS zero in the simulator.

[bric...@gmail.com]

Looks interesting, Bob. You may have something. From a python.py file for that machinekit build:

# to disable five axis kinematics:
# G60.1
# You can now disable tool length offset with a G49
# or issue commands such as G53 Z0 to ignore tool length offset

However, I don't want to lose my TLOs. My cutters have different flute lengths, and I don't think I can get them all set into the collets at the same "stickout".

For kicks, I'll try it. I'll likely need to add plastic stop-rings to each cutter set to a uniform depth before mounting them in the collets.

[Bob Vawter]

Right above that comment you posted, it says that tool length offset must be enabled for G60 mode using a G43. That pair of codes occurs whenever the post turns on TCP mode. In my test program I see

G60
G43 H2

Where H2 corresponds to tool-number 2 in my program.

Total conjecture on my part, but it seems that correct TLOs are mandatory, but that as long as your Fusion tool-library numbers line up with the TLO table in the machine, you'd be ok without having to set every tool to exactly the same length.

I'm eager to hear how your experiment goes (and to get my machine :-)

[ripf...@gmail.com]

What I've been doing to quickly work around that is to set my tool number to zero and post, this, so far, makes it so the pocketnc doesn't use the tool length option when machining and you can use your g54 offset normally.