I am starting work on a new project creating a board that will act as an interface between the Glowforge’s proprietary hardware and common 3rd party CNC driver boards.
General idea is a replacement board for the existing control PCB to provide simple breakout terminals for the fans, steppers and laser controls. This would then be connected to any of the many available boards that run GRBL or Marlin or whatever. It will also have the option to install a small SoC board like a RPI Zero 2 W (or similar) to get access to the cameras, lights, etc.
I’m looking for input from the community to see what features and capabilities people are interested in.
Main progress will be tracked on the OpenGlow forum, to avoid any risks of running afoul of ToS. Though, I am happy to discuss ideas here - for those that are not interested in signing up over there.
Very cool to see that you’re still working on this stuff.
As one of the folks still running on borrowed time with a day-one machine, I am personally less interested in changing out the controller than being able to fix it when something finally goes. That something is likely to be the tube or power supply so I am prepared to have to say a sad goodbye when the day comes.
I took a bit of a detour, which ended up in the revival of the ForgeFIRM project (firmware that runs on factory Basic/Plus/Pro devices).
Anyhow, I have GRBL running on stock hardware. Working extremely well with LightBurn. It allows you to switch between GRBL and the Glowforge cloud service.. Best of both worlds, no?
WAUW! What is the status of this project? I am close to throwing my Glowforge out the window everytime I try to setup a larger batch of pieces. The coordinate system is constantly shiftning, when changing between engrave and cut, but if that only was the issue, you can compensate for it in the positioning of the graphics. how ever every f#cking time I ad somewhere between five or say ten pieces, the coordinate system shifts AGAIN?! I HATE that I have this unit, and not some plain simple XTOOL or other laser. That ‘cheaper’ Glowforge has surely made a discount feel terrible.
So PLEASE let ud know how to hack the son of a b…
The coordinates don’t shift unless you accidently hit the head or arm or something, so something else is going on. I might suggest you make a post of your own with more information, because I bet someone here can help you be less frustrated.
Reading through @evansd2 thread is definitely a good first step.
Literally I can turn my machine off, and come back the next day, hit print, and it’ll be aligned exactly - that is standard behaviour.
Reading the documentation, you definitely know where the bodies are buried on the design and specs. The section on commissioning was really interesting. It seemed hacky at first glance but the more I think of it the more I like the concept, you can allow for much larger manufacturing tolerances if you have a commissioning process. Glowforge must have an analogous process that they run in-house, they have referred to the idea that your machine is “calibrated” at the factory, this must be the same idea. Cool.
It’s only ran against a single machine - my 2017 basic (with a plus control board). I have zero reference to know exactly what the manufacturing tolerances are.
This is why users in the early days are asked to share the results of the process. That way I can gauge how much of that is really necessary. Only things that need to be there will remain.
It’s possible that the only real differences are the tubes (which can vary by age and use).
Even after the early work I did on these, I continue to find new and interesting nuances while working on ForgeFIRM.
Some decisions they made are certainly head scratching. But, on the whole, I can see (and appreciate) the ambition of the original engineering team. They were genuinely trying to do something different and innovating. In many ways, I think they achieved it.
The bodies you are referring to are present in every piece of hardware I have ever had the privilege to reverse engineer. They’re certainly present in every piece of hardware I have ever designed. It’s part of the process. Ghosts in the machine.
There are a lot of capabilities that have yet to be tapped. While not the most robust platform by today’s standards, there is plenty of potential lying in wait.
The beta version of ForgeFIRM already lets some of it loose:
Parts of the SoC that have long been dormant have been brought to life (graphic and video processing and streaming, for one).
Their brilliant use of other capabilities has been embraced and enhanced to enable localized motion planning with finer step resolution (Grbl shipping to the SDMA/EPIT ring for timely step generation at 32 micro steps)
Many, many sensors that you may not even have known where in there…
Yes, I know where ‘some’ of the bodies are buried. I also know where many of the capabilities are that were built into this hardware that have yet to make themselves known. (Sadly, though… some it may not be able to be set free because of lawyers, developers not wanting to get sued, and what not).
I have to say, I have actually enjoyed the last 4 months or so of work on this. I hope when you finally see it, you will too!
