The Leadshine site does not have that driver (⸘‽), but it does have a CS1-D507S driver that is obviously different. If you have a link to the manual for your CL1-507-3LS, that will be helpful, as all my searches come up empty.
The switch settings on the CS1-D507S driver do not match what’s shown on the side of the CL1-507-3LS driver on the Thunder Laser site:
A photo of the data plate on the side of your driver (opposite the heatsink) will confirm that picture matches your driver.
Assuming your drivers match the CL1-507:
- SW1-4 = ON OFF ON OFF = 4000 step/rev
- SW5 = ON = rotate clockwise
- SW6 = OFF = enable auto tuning
- SW7 = OFF = Pulse/Dir
- SW8-9 = OFF OFF = default time constant
- SW10 = ON = 500 kHz BW
- SW11 = OFF = pendant
- SW12 = OFF = closed loop control mode
Assuming the CL1 functions are similar enough to the CS1, I can leap to some conclusions.
The CS1-D507S manual says nothing about auto tuning beyond the switch description, but it generally means the driver will adjust its internal settings to improve the step response. It is not clear what that would mean in this situation, but I’d turn it off (SW6 = ON) and see what happens to the 60 mm/s 100 LPI test.
Other than that, nothing sticks out.
Now, based on some doodling …
As you observe, slower scan speeds show a damped oscillation, rather than what looked like a relaxing offset at higher speeds. Running some numbers on the 60 mm/s 100 LPI scan suggests the oscillation has a period of 25 µs → 40 Hz.
Assuming your KH7050 machine looks like that, it’s essentially identical to mine, but with a box sticking out for the anode end of the 100 W = much longer tube.
Quite by coincidence, I recently ran a set of similar tests on my (simple stepper motor) machine and came up with a gantry oscillation with a 10 ms period → 100 Hz.
That’s Close Enough™ to the frequency of your machine to suggest my initial assumption of a slow servo loop filter was wrong. The initial speed of 500 mm/s you showed smears those oscillations across 8× the distance with the initial offset corresponding to the initial position error at the start of the line, which comes from the phase of the oscillation; the oscillation is so stretched we didn’t see the damped cycles.
Based on your series of tests, I think we’re seeing a simple mechanical oscillation of the entire gantry after a step change in Y axis position, rather than anything horribly wrong in the driver.
Bottom line: try disabling auto-tune (on both drivers) and see what happens. Most likely, it’ll look about the same, which will confirm that’s not the problem; whereupon re-enable it as before.
The RDC6442 manual has a section for Cutting Parameters containing an Idle Move Acceleration value with this comment:
idle stroke speed and idle stroke acceleration can be set higher to reduce the working time of the whole figure, but if they are set too high, it may cause the jarring of track, so comprehensive
consideration should be given to the setting.
There is a similar Y Axle Acceleration setting in the Scanning Parameters section, without any description. I think the Y axis values affect the speed / acceleration of the line-to-line motion.
Experimenting with those values will be instructive, as reducing the Scanning settings should move the whole gantry more gently from line to line. That’s the next part of my experiments here, so you’re sailing into uncharted waters on my map. 