I did the math but in my simpler understanding if we are configuring for 12800 steps and it outputs 51200 we have to tell Lightburn that 51200 steps is a revolution. Because we are note configuring the machine, we are telling Lightburn how the machine is set.
I still may be wrong.
I don’t get that. If a setting in the rotary setup window of 12,800 outputs 51,200 to the microstep driver, then a setting of 51200 would output 204,800 to the microstep driver. Neither is the case. A setting of 12800 in the rotary setup window outputs 12800 pulses to the microstep driver.
Close, but we have to configure the microstep driver for 12800. Or if we configure the software for 51200, we have to configure the microstep driver for 51200. (Better have a pretty dam good microstep driver, not one of those $11.00 ones) Or have a 4:1 reduction.
Here is an easy test (Caveat: For galvo rotary BJJCZ board or similar) Set the rotary steps in the rotary setup window for 3200 steps per rotation (microsteps). Hit test. See how far the chuck or wheel rotates. Set the rotary steps in the rotary setup window to 6400 steps per rotation. Hit test. See if it rotates twice as far (Or half as far) I think It will rotate twice as far. (I could go down to the shop and try it but about my bed time!) If the microstep driver is set for 3200, it will make 2 rotations, stretching the art out 200%, because the BJJCZ board is still telling the laser the rotary moved 360 degrees, but the microstep driver is telling the stepper motor to move 720 degrees. If software is set for 12800 and microstep driver is set for 3200, it will rotate 4 times and stretch the art out 400%
I have a 6:1 reduction on mine, microstep driver set for 6400 and rotary setup window set for 38400 to get one revolution. If I connected it to a direct drive, changing nothing else, it would rotate 6 times.
It was a long day and maybe I´m in tilt mode, now I also have to test, You planted the seeds of doubt in my mind. 
Guys, I tried reducing steps per revolution. For example halving them to 6400 only led to erratic movements when I tried to engrave a 2mm wide (correct) and 10 mm high (wrong result) object.
Not only the wrong height, the rotary did not move consistently, even wrnt back and forth, jumping multiple lines of engraving.
I am more and more convinced the problem is located on the mainboard. So do not waste time.
And for the math problem:
12800 steps per revolution should be 0.02812° per pulse. To reduce the too big dimension of the output, only reducing the number of steps could lead to a correct result. (Which it does not in my case, because the movement becomes inconsistent and unpredictable)
QUESTION: When set for 6400, what did the “TEST” button do?
It could also be the microstep driver, I have bought several over the last couple years and 1 was DOA. I have one in the $90.00 range in my galvo, and $20-$25 in my projects, but they do go as cheap as a couple bucks. And even the good ones can fail.
IF you had only posted you have erratic movement and nothing else most of us would say bad wiring/ loose wiring/ reversed two wires, etc. between the driver and the stepper motor.
It also may be “Less than precise” descriptions of what we are talking about and results we are expecting.
Interesting thought.
The very first time I plugged in a rotary chuck it moved erratically and made quite a noise. At first I thought the rotary was defective until I took a closer look and realized I didn’t seat the plug in correctly.
But I don’t think that would have caused it revolve 4 times
I think I was thinking I was in DIY mode ( who knows), like the one o is “manufacturing the device” instead of using the final product wit it specs. Altough the math may not be right because I was imposing a lower deg/step value that may not be real for the involved system. 