Twin Spindle / Twin Turret turning center

rexo wrote: Two linuxcnc is probably the way but i can't find solution how to interconnect it and sending data from one to second. The only one function i found is linuxcncrsh for commanding linuxcnc over network.

I think that this would be best done with a custom setup where the controllers toggle dedicated GPIO pins that are read/written by G-code.

So, start with a list of what each controller needs to know about the other.

For example I would dedicate a digital pin that would tell the secondary spindle to sync speed exactly with the primary one, and then make the primary spindle angle available to the secondary one to use as a target for a position PID.

But, first of all, I would suggest making it work as a single-controller XZUW machine and use it for a while. That will help to realise what you need the machine to do, and give you a grounding in how to do it.

I can also imagine a set up that combines both, with a master control and daughter, where the master could control both sides as XZUW, or enable the daughter to execute synchronized code independently.

I think the hardest and most important part would be building the nessisary custom user interface capable of interacting with the daughter control. Hasn't there been some work on remote interfaces? This would seem to be the perfect use for such development.

Now i have working config, i am using 2 axis (X, Z) and 4 joints, for correction i use U and W column
basic tool correction is only in X,Z axis, in the U,W is only difference from X,Z
value is read from motion.tooloffset.u (w) and transferred to custom kinematics.
This solution is working, i have one program for 2 spindles and two tool turrets.
But i have one problem.
When tool and correction is loaded, then 3 and 4 axis is move to corrected position, but very slow.
This is due to pid, i don't known why. Axis is moving correctly, only correction moves is slow.

Could you post your config files?

here is config files

I try modify stepgen to position mode without pid and now it's working fine, offset moves is fast.

Well thats odd. PID is preferred


I see P for joint 2 was 100 should be 1000

Can you make a video of the lathe in action - very curious to see!

I don't think that a change in the tool offset should _ever_ cause an axis movement until the next programmed move.

What you describe could happen if the pid pins were connected to the wrong inputs and outputs, but the HAL looks conventional enough.

But, anyway: If an axis moves with just a change in tool offset (rather than just the displayed position changing) then something is wrong somewhere.