Roadmap for raster clustering

rabbit

Hi,
Are there any plans when raster clustering will be implemented?
I have a fairly big CO2 laser with duet3 (sbc). Now I have the issue that greyscale engraving is overloading the raspberry pi.
Best,

T3P3Tony

@rabbit Its on our list to consider (and has been for a while) I can't put a target release on it yet.

o_lampe

@rabbit
I'm using a Raspberry 400 with 4GB Ram on my CO2 laser. I'm using Meerk40t software which is capable of writing a gcode file. I guess, it's much easier to send only gcode lines instead of calculating everything on the fly and send it over?

It also has an option to run from PC and send life data to a remote server on the Pi. That's probably the smartest way.

droftarts

@o_lampe raster clustering is described here https://forum.cohesion3d.com/t/improved-raster-speed-firmware-update-for-cohesion3d-boards/850/8
It's calculated in the laser GCode generation software, and sent to the laser controller in a longer form of G1 command. This allows the controller firmware to use more of it's GCode buffer for look ahead, rather than a lot of individual G1 commands. It's most useful for laser engraving/rastering, rather than cutting.

From this post, later in the thread https://forum.cohesion3d.com/t/improved-raster-speed-firmware-update-for-cohesion3d-boards/850/13:

The simple explanation is that you send one G1 move and send one or more S values, separated by colons, to be evenly distributed along the G1 move, like this:
G1 X0.5 S1:0:0.5:0.75:0:0.2

From what I understand, it's implemented in Lightburn (software developer @LightBurn is also on the forum here), not sure about other laser control/gcode generation software.

Ian

dc42

@o_lampe I've added this to the work list for RRF 3.5.

dc42

@o_lampe this feature may make it into RRF 3.5 experimentally because it can use some of the code I am adding to support another feature. Will you be able to beta test it? Which Duet do you have in your laser machine?

o_lampe

@dc42 It wasn't me who started the thread. But to answer the question, I still have the original controller in my K40.
Does the new feature require a Duet3 compatible MCU? Buying a Mini5+ and a Lightburn license is out off the scope for me...

dc42

@o_lampe it does not require a Duet 3, it should work also on Duet 2.

dc42

@rabbit this facility will be implemented in RRF3.5beta3 which we hope to release later this week.

rabbit

Hi,
That are great news.
I have to build an air filtration system first, but then I can beta test it.
I use the duet 3 board without SBC.
Best,
Vincent

bg_86

@dc42 I resorted to turning on clustering because I was desperate to resolve a speed issue which I detailed in my first post here earlier. Ran the clustered gcode for 3 minutes and saw immediate speed increases but if you need me to do any testing I have a corexy machine I built out of 1/4 aluminum plate with 2A nema17s that flies. Raster engraving preliminary test and subsequent print cancel showed excellent results. Not sure what you want tested.

dc42

@bg_86 thanks for the feedback. It would be good to know how well it works at high speeds, in particular whether the registration of pixels printed on adjacent rows is accurate enough.

bg_86

@dc42 My 10W diode laser doesn't deliver enough photons at 400mm/s to get any real definition that you can visually analyze well on the one rastered image test I did tonight. I'm not sure where your concerns might lie in regards to adjacent row alignment and testing. Are you looking for a heavy clustered line from a rastered image with lots of on/off commands in a single row or will a basic vector outline image with less clusters work? Also, if I do happen to see streaks that extend beyond the point where the laser should be commanded to stop, could the speed of the 5khz LaserTree 10W laser to shut off the beam be muddying the waters of the test? I did a quick test of a 5" wide rastered image and I want to say I see streaks extending but I figured I would reach out with a little more guidance on this testing before I provide any kind of seat of the pants test results.

So far though? Looking good!

o_lampe

@bg_86 said in Roadmap for raster clustering:

I want to say I see streaks extending

Do you use air assist that blows still hot wood particles in one direction?
I only use air assist when I cut through material, because the dust can disappear below then.
Laser Diodes often don't use air assist (beside the cooling fan for the laser), but I thought I mention it.

bg_86

@o_lampe I do, but the air was off for late night testing in the garage. Clever thinking tho