AWC-608 & PPI

[sshwarts]

I've never been overly impressed with the performance of the AWC-608 when using PPI (with vector in particular). I always suspected that they implemented it differently than some of the other implementations described here on buildlog.

I decided to put my scope on it to examine the pulse length on the TTL line with various PPI settings. I was hopeful that it stayed constant with different feed rates, but it's not. Here's what I measured. Maybe someone can help me figure out the math behind their implementation. The pulse width is the width of the portion of the square wave that is low. The pulse train appears to be symmetrical, so I'd guess they just determined the time of the move and width of pulse they'd need to get the desired PPI.

100 PPI
Feedrate Pulse Width (ms)

7 21.6
8 19
9 17
10 15.2

200 PPI (it halved which I expected)

7 10.8
8 9.2
9 8.4
10 7.6

800 PPI

7 5.6
8 4.8
9 4.4
10 4

1000 PPI

10 2.4

At a Feedrate of 1 mm/sec and 100 PPI, pulse width was 152 ms

dirktheeng original postings about PPI, he found that a 2-3ms pulse width gave the best overall performance which if true, points to specific PPI and feedrates.

[terjeio]

Pulses for 100 & 200 DPI have a duty cycle of approximately 60% based on your measurements, the formula then seems to be 25.4 / speed / DPI * 3 / 5. Your measurements for 800 & 1000 PPI does not make sense, they imply a duty cycle of more than 100% if the DPI setting is to be trusted.

25.4mm / 10mm/s / 800 PPI * 3 / 5 = 0.001905 (~ 1.9ms) per pulse, less than your 4ms measurement. 4ms pulse width is close to 400 PPI @ 10mm/s with a 60% duty cycle.

Is it not possible to set the duty cycle or pulse width as a separate parameter?

[sshwarts]

Thanks for the response. I might rerun some of those measurements to double check. Unfortunately there does not to my knowledge seem to be any way to adjust the pulse width or duty cycle.

To make a generalization, I'd have to say that the AWC608 and I assume here, the newer model, don't really do PPI at least as described here, or by some of the other laser manufacturers. If correct, it doesn't surprise me, I know from some Marco comments that the Chinese programmer's don't understand the need/desire for PPI control.

I might investigate some ways to add "true" PPI by using the first pulse width and the laser firing output. While I can't count per se, I do know what pulse widths the controller thinks it supposed to be outputting.

Scott