My comment was more toward others. I am like you and really like to know how things work (regardless of whether I own one or not).
I love reading through the information you have shared on your github wiki. I’m hopeful that it will lead to having the various protocols documented enough that someone will be able to write a “driver” that would talk to the machine without the cloud.
Also, it appears that glowforge is using quite a bit of GPL and LGPL licensed code, and should have all their code available to avoid being in violation of those licenses. I hope they release the firmware source soon.
The protocol between the device and the cloud is well understood, documented, and open sourced Python modules exist (they haven’t been publicly released, yet).
The structure of the motion files sent to the device from the cloud is well understood and documented (you can thank @palmercr for that), and development of open sourced library code for them is underway.
The only parts missing are a user interface (other than command line), a “G-code to GF-PULS” converter, and some glue-code to put it all together.
The “G-code to GF-PULS” converter will not be a small undertaking, but is completely doable.
If the beam had any significant divergence the focus would change as the head moved horizontally because the beam travels much further on the right. This implies the distance from the laser to the lens has no effect and the focus is therefore a fixed distance below it.
I don’t have my glowforge yet so I can’t comment whether the missing focus range actually makes a functional difference in cutting or engraving. However, if it does even 0.015-0.03 would be unfortunate because the missing upper end happens to be 0.5"… an extremely common dimensional size in the US. I think more likely to see it during engraves in materials like metal and glass rather than woods and plastics
You mentioned the pass-thru slot is not working on the Pro?
Good to hear. You might try searching for a project called LightBurn. I know the guy coding that, and given the GF_PULS info he could possibly make something that produced that instead of needing to go from G-code to it. He does already support some other specific later drivers boards, and a g-code output form as well.
Perhaps if you guys have something that can take GF_PULS data and feed the machine, then something could be worked out to interface between his code and yours?
I am already using LightBurn to “compose” stuff and then export to SVG to load into the GFUI. It’s quite nice, and FAST. Still being developed, so some things are not yet there, but he’s updating regularly.
The pass though slots are there, and you can use them for larger materials. However, there is no software support for aligning things if you wanted to engrave/cut larger areas, this is something that is supposed to be a feature.
As in, you put large material in though the slot, do an engrave/cut pass on the portion that is accessible to the laser, then slide the material further in and do another engrave/cut pass that is all lined up with the first one automatically by the software.
This can be done manually with jigs to varying degrees of success/acceptability.
I have used my pass through slot many times to do jobs on a portion of a larger piece of material, so it has that benefit.
Yeah that’s a huge deal. Software alignment to me was the main feature of the pass thru slot vs every other machine with a pass thru and where the pricing difference between the basic vs pro could be justified.
I’ve seen problems on forum with alignment vs cameras so I guess maybe they have to get that out of the way first. I’m assuming they still plan to do it.
the good news is that lots of dimensional plastics as well as wood composites use “nominal” dimensions: the actual size is often at the bottom of the tolerance range, or -10% on most plastics…
similar on plywoods…
Even if that was consistently true (which I know for engineering plastics it’s often not), that’s not the really the problem for the most part because you can height of the material.
The problem is it’s supposed to have continuous real-time focusing, so 0.5" focusing within the same part.
Formally, no.
It’s documented in the Python code that’s being written to read and write the file type.
It’s actually pretty simple.
The only part that isn’t super well understood is the header, which contains some device settings that use abbreviations that haven’t been completed deciphered - yet.
I don’t know that the continuous auto-focus is really that feasible.
As @palmercr has documented, the focus step resolution is currently 0.7 mm. It’s likely that this would leave visible witness marks in any engravings when the focus shifted between steps.
I have a similar concern, that I am actively digging into right now, concerning the modulation of the laser’s power: Rather than vary the PWM signal to the power supply, they pulse the supply at full power using a dithering algorithm.
The result is very visible patterns in engravings.
I am working to see if this was done due to a limitation of the power supply, the way the cloud software works, a limitation caused by the PULS files they use, or something I haven’t thought of (probably the most likely).
I wonder if we’ll see some novel definitions here. Perhaps if your machine is the same as Scott’s, “focus on materials” means “to within a defined tolerance of 0.015in” or something like that.
for a while.