Hi there. I’m using the fill mode and scan angle to achieve a certain look. As I understand, fill mode at multiples of 90 degrees is handled in hardware, and thus the overscan is calculated on the machine. However, it seems that other scan angles lack overscan. Here’s a screenshot of LB’s preview to show what I mean -
Presumably, LB guesses the overscan based on the machine config. But, is there not a way for LB to force overscan when using the scan angle feature, considering it isn’t being done in hardware on the controller anyways? It’s really quite limiting at the moment, as the material I’m engraving really benefits from the overscan, but it doesn’t seem that I can combine it with the scan angle.
Hopefully I’m not just missing something obvious. Thanks in advance for the help.
Yeah, I can do that, but it proves a huge pain. At that point I’d sooner ditch it. But, since Trocen doesn’t seem to do it internally, I’m basically requesting it get added to Lightburn. I think the logic already exists for gcode machines…
It may well be there is not, because turning the laser off precisely at the boundary requires two motion commands with differing power.
The controller surely does its usual lookahead to splice the commands together, but perhaps includes deceleration before and acceleration after the common point to meet the cornering motion limits, thus producing visible defects at the boundary.
I do know that Ruida controllers have three classes of motion commands:
Idle with the laser off (rapid motion)
Cut with the laser on (low power = “engrave”)
Engrave over a prescribed distance
How it handles the junction between a Cut with nonzero power (to “engrave”) and a Cut with zero power (to decelerate) remains a mystery.
You can take a look at the G-Code for various engravings to see what happens. The note in the G-Code Clustering section suggests that’s how engraving on a diagonal gets done.
Things that seem obviously trivial to those of us with no knowledge of how the firmware works may not be possible given its limitations, which may explain why folks who actually know what they’re doing haven’t done the obvious thing.
That’s fair. Hopefully it’s something that can be looked into though. We have two Trocen machines so I’d be willing to troubleshoot and test this sort of thing if necessary.
