The problem, unless I’m mistaken, is from where the fiber originates, such as the galvo and the angle it strikes the material (incident angle). Further away from the center of the galvo, the higher the angle. If you change the galvo distance from the material that angle changes.
I have worked with lots of lenses, I don’t claim to know them all, but when the industry uses specific types of lenses, there is usually a reason for it.
I doubt we are the only people that’s aware of this issue.
I think a telecentric lens will produce what we want, but it’s coverage is directly related to it’s size and that makes it difficult to use in most applications, such as with a galvo of any usable size. I understand they can
I think the only real fix is to use a gantry machine where the beam is always perpendicular to the material.
I feel like we are overcomplicating the problem… ever deep engrave solution on the market (EZCad 2 with and EZCad3) both handle deep engrave the same way… They slice the art, and then lower the z at a set rate based on a number of passes…
Will this give you machinest level quality for cutting and drilling… no, but that is true for all galvo heads (geometry is tough to get around)…
Also if you think “I want to optimally cut the current surface” then the answer is pretty easy “lower z X steps per pass” is the way to keep the focal point on the current surface. I don’t think anyone here is asking for 3d following (which to be fair is an advanced feature of EZCad3 and some other industrial softwares… but those are like mark on the fly (marking a moving object) features which are not in Lightburns wheel house either…
I go back to my original thought which is if we could get a “ping this output and wait X MS” would give most of us looking for Z control more than enough functionality to get to where we want to go right now.
Josh I agree with what you are saying, just to add ezcad 3 has the ability to use a .stl mesh it does nothing more than slice it in the same way a resin printer software would which then converts the slices to b/w images, pretty much the same/similar output of depth maps. Ezcad3 is not that advanced for full 3d the best you will get from it is 2.5d.
Ezcad 3 can’t deal with undercuts, the same as a CnC can’t deal with them either.
I would like to see a simple solution as well based on slices lower the z as we go.
Hopefully this conversation grows some legs… and more people get involved in the discuss at least and see the merit of having a feature like this added to an already very nice piece of software.
Dynamic autofocus looks pretty cool, but does not fix occlusion. It can track curved surfaces but not do depth cuts.
One other note I had was that if the bottom of the cut becomes irregular, that basically spoils the cut. The depth of focus is pretty shallow, and there may be reflection instead of absorption at low angles of incidence. When I was trying to cut a hole in a Lowes ceramic tile, I kept running into what were probably impurities of some sort of filler that was much harder to cut. Once the depth around it eroded away, the area under that part turned into a very well-defined cone widening under the speck that didn’t cut. This was only corrected when the beam got out of focus and then only heated the material instead of cutting, it vitrified the ceramic into a liquid which leveled itself. Nice, but probably not very practical.
So, if you did make a “laser cnc mill” with the galvo set to just wobble in place and moved the stock, attempting a single pass would end up creating a slope down the cut and it can’t focus on all of it. Sounds like it would need to be many passes because a single pass depth would have to be much smaller than the depth of focus.
But, any inconsistency would get you in trouble again, as it would tend to make itself worse quickly.
I suspect that the wobble is going to run into that on curves, even with a very low pass depth. The inside is going to get more exposure than the outside, slanting the bottom. Maybe dynamic autofocus would adjust for that, I don’t know. Seems unlikely.
I’m playing with the concept further, trying to cut ~1/4" stainless in a straight line that is coincident with the center of the beam field so there is no occlusion.
It’s slow so far. I probably could be going a LOT faster if I had the right settings. I’m messing with q-pulse, freq, speed, and wobble width/step and the results are all over the place. Or, just forget the wobble and do a line cut.
The shower of sparks coming out of the cut is of some use gauging the effect, but it’s deceptive in that a wider cut can throw a large volume of sparks while gaining very little depth
It’s mildly interesting in that I have no other way to cut this thickness of stainless in my garage. But I’ve been at it like 30 min trying to get through ~180mm wide flat of it. Learning experiment.
The need for automated Z control is clear. I don’t know what rate to set to drop it yet, but this is like 10x its depth of focus so any cutting by this method would need constant manual input otherwise
I tried a lot of different different settings on a 1911 slide, about 5-6 mm thick. I tried 2 different things, using wobble and offsetting the line .125 and using fill. Rectangle About 1/8" wide and 3/8" tall right under the lens so figured 25mm of cut. Never got better than about 1-2mm/hour. Tried fast and slow, High and low freq, High and low Q. Get tired of waiting. Try something different next time. Maybe cut some slots and use feeler gauges to measure depth.
I don’t think we ever got a full on response form Lightburn engineers, is there any kind of signal we can “listen for” via IO pins or API for a layer completion with the EzCad 2 controllers?
I am guessing, given the marginal (but nice) feature enhancements in LB 2.0 that there was a lot of refactoring and technical debt payoff to warrant calling it LB 2.0 not just LB 1.8… Once they get through the beta/RC/release rush I may spark this question back up as there seems to be lots of ask, lots of speculation, but very little engagement from the actual engineering teams.
yes I have voted. and again, if you look back at the position I have been taking its never been “z control” which would require writing logic for steppers or other things, and far more limited to “pause per layer” and “ping an IO pin” per layer… Which was posted more than 3 weeks after Colin’s post and as such I have no reason to believe its been materially considered…
the EZCad2 lite controllers don’t have 2 controllable axis so it makes since that its a lot of code for a limited market to do complete Z axis control… however the far more limited ask may be acceptable for those of us willing to do the work beyond that…
I don’t know how Lightburn would/could use these as I know little about these controllers.
To remap any of the pins or other things that involve the controller, I don’t know if they have enough information to even begin software development. I know I can select which input I use, but don’t remember any output can do what.
If they could detect it, wouldn’t it be just as useful in a Z axes control?
I can’t speak for them, but all of the people I’ve dealt with there are on top of what people want. Unless you mute a thread, you get notified every time someone posts or clicks an icon, such as a heart.
They do miss stuff now and then, but I figure they’re human, so I give’em a break
One of the issues, I understand, with the EZCad3 is that the communications from the pc to the fiber is encrypted.
The IO Pins are there and can be called by Lightburn with the EZCad2 controllers…
Also if this Opensource project is right, then Lightburn is sending “one command” at a time Bryce Schroeder / Balor Galvo Laser CLI · GitLab… As such they would know when they have finished a layer… That is not to say there isn’t an optimized way to talk to the controller, and that LB isn’t using that optimized way over a call by call way… however, it does look like (from a quick code review) that it is “possible” to know exactly what the galvo head is doing…
I totally respect this is a priorities issue, and as I said earlier today given the 2.0 release is right now, there are bigger fish to fry like making 2.0 stable enough for the next N years of improvements.
I also don’t want this thread to auto close and the 1 comment we got from LB engineers over a month ago is all there is to the discussion… Until someone new comes along and strikes up a new conversation and its gets treated as a “new request” all over again
take a look at galvoplotter too; It credits balor as an original source/inspiration. It does have a way to check the status (in galvo/controller.py). My approach is to send a layer at a time to the spooler and wait for idle before moving the Z axis and firing off another layer.
I am pleased -and surprised- to say there seems to be a ready-made solution for the Galvo Z control problem!
I am currently into buying a new machine from an AliBaba supplier and they have showed me a video of an item being marked at three different height levels. In between the Z axis automatically re-adjusts the focus.
The supplier claims it works with an EZCad2 board ánd works with LightBurn as well! It can’t be too expensive because he has upgraded my machine for free.
It works by plugging in the Z-plug (you do need a motorized Z-tower) into the Rotary socket. (I have not seen the device itself, only the marking on the object). Apparently you can set it up for all kinds of uses, like 2.5d engraving, adjusting focus for deeper cuts etc.
I hope my machine arrives in a month or so. When I get it I will shure test it and post a video or something on the LightBurn Forum
Sounds interesting. Thought about that before it would take some experimenting but everything’s already there. I was thinking using repeat marking with the movement set to a very small amount.
Well, about sharing the video. The supplier has let me know he would appreciate it if I could share the video, but we are in a little bit of a dispute regarding the machine I am ordering. If everything goes well, my machine should be ready for transport in a few days. I just want to make shure everything is being shipped on the agreed terms before sharing the video.
Please understand that it’s not that I don’t want to help out the community, but I really want to be shure that everything checks out with the supplier.
Long story short; I WILL share the video but please have a little patience
