Yes unfortunately no one on either of those threads posted back to say if it was fixed or not. Unhelpful 
Wish there was a way to use two controllers at once. One for pure G2 arcs and the other for the Z axis in addition to the 0-10V analog control and 24V TTL signals to control the laser source.
A machine cannot serve two masters.
Unfortunately FSCut controllers are about $4K. They actually use a PCI card installed in a computer for their setup (not to mention servo drives)
Not that I’d feel like reverse engineering the current controller to figure out the signals to control the source in addition to the height sensing amplifier.
Gotta be some way of salvaging the situation.
What did you change since the last cut to get those flats?
Nothing at all, they were there before actually in the photo with the two circular discs.
Changed several settings with the same results. At least on a positive note -> have started to figure out how to reduce corners from being vaporized by using the minimum power setting.
Is there a way to use an oscilloscope to probe the X and Y drivers and log the data to confirm that the drivers aren’t being sent pulses to make the straight lines? At least that would rule out electrical/software issues and point purely to either belt tension or other gremlins.
Are your X and Y parameters exactly the same?
Without knowing the controller, I’m just guessing.
Yes sir, identical for X and Y.
I wonder if you’d be able to see an issue of you simulated the cut of one full-work-area-sized circle.
Could you swap the step/dir signals for X and Y? If the flats stays on the same axis, you could guess it’s a mechanical thing.
Wonder if this could be done just attaching a ballpoint pen on the head and using it as a dumb plotter. Would reduce waste in metal and O2 at least!
I found this little nugget of info that nearly looks identical to the cuts in the pics I posted earlier, found at 2 minutes 48 seconds: However that doesn’t explain the source of why the Pro might have backlash in it and how to solve it without a software bandaid.
1 Like
This is really cool. I know that LightBurn is held in wide regard (and have played with it a little) but didn’t realize it was this advanced. Thanks for sharing!
Given the low backlash nature of timing belts and pulleys, and assuming you have good tension all around, have you checked your grub screws for good seating and tightness?
That was the first thing we did on unboxing the Shapeoko Pro was tighten the pulley grub screws. Glad we did since they were loose! 
Should you be able to pinch the belt below the pulley above the bearing idlers in any way, shape or form? What is a good amount or none at all?
Unrelated, but here is an interesting (not in a good way) feedback loop of the height sensing in action. Note that isn’t caused by air going through the nozzle or any outside forces. I stopped the loop by holding the plate down. I initiated it by lifting the plate up and dropping it!
How are you doing height sensing and how tightly does it have to be controlled? Sheet metal is such a pain for flatness, at least on the plasma the arc voltage sensing gives a pretty nice control loop. I’ve only really seen vision and lidar based autofocus for hobby grade lasers.
The industry uses these RF amplified capacitive sensing via a ceramic insert in the end of the fiber head. In earlier pics you can see the amplifier mounted midway on the left side of the head with a coaxial cable going to the business end.
I actually have it on the least sensitive setting and I wish it could go even lower.
You can also jog to the plate manually with height sense disabled and use a feeler gauge to get the nozzle gap right. I made a cut that way to see if it would work. I’m not sure if it would work though going across the entire working envelope of the slats due to any cumulative tolerance issues. For the 20x20mm cut though it was nice to not see the head ‘hunting’ for the plate.
1 Like
I stumbled on this old post on CNCZone that confirms my thoughts:
Going to give this a try today.
https://www.cnczone.com/forums/diy-cnc-router-table-machines/322998-cnc-post1967716.html#post1967716
Works fine, as long as the belt pulleys screws are tightened with both motors “on” to make sure they are stepping in sync. Belt does not need to be super tight since it will not influence the accuracy.
1 Like
That’s is almost certainly backlash. My first 3D printer made circles like that. It had a very similar layout to the Shapeoko. X had two steppers. One on either side. The GT2 pulley on one of the X steppers did not have the grub screw properly tightened. It would wiggle from one side of the flat on the stepper’s D shaft to the other. This meant that that side would lag behind the other. This is why one of the flats on you circle is higher than the other.
Unfortunately one of the pulleys is totally seized to to the motor trying to remove it, so now we will need at least one new pulley and one new motor.
Is there an emergency overnight way to get parts @WillAdams ? I’d prefer a shaft clamping pulley instead of set screws but they probably don’t even exist in this size.
Done cc’d sales as well.