Is my lid camera misaligned?

Sure, I’ll use the same example I’ve used before, that of a steel ruler. Unlike the GF one:

  1. This one’s markings start at the very edge of the material on 3 sides, and;
  2. It’s made of a material that the GF cannot cut (so we have no bleed to cut away like with timber).

Example: https://www.jetpens.com/Kokuyo-Stainless-Steel-Ruler-15-cm/pd/11291

Oh, and we want to fill the bed with them.

Looking forward to hearing how to best tackle that one without a known 0,0?

1 Like

I think the answer here is that we will have to build our own 0,0 jig. Its one of the main things I am wanting to accomplish first when I get my prod unit. This along with numeric positioning will allow me to skip making a jig for so many things, as it will basically be a corner in the top left to push everything into and measure against.

4 Likes

Is the goal to engrave an existing blank ruler? If so, I would create a file like this:


then put a full sheet of inexpensive material down on the bed, with magnets or tape to hold it in place (depending on the material). I’d load the file. I’d set the teal engrave step to ignore. I’d cut the red line. I’d pull out the center piece of material and lay down the rulers as shown in the ‘hole’. Then I’d set the red line to ignore and the teal lines to engrave.

We have other ideas for how to make this even easier in the future, but on the rare occasions I’ve needed more accuracy than the camera can provide, I’ve found this approach to be only an extra few minutes of effort.

14 Likes

It is.

I like this and if it’s all we’ve got right now I’ll take it…

A few downsides:

  1. There is a bit of waste on the bed dimension because we’d have to leave enough margin for rigidity of the jig;
  2. We’d have to go through the process EVERY time we want to engrave a set of rulers.
  3. There is material waste (cheap or not).
  4. We’d have to find a way to hold down the jig very securely - a few strong neodymiums would probably suffice;

But… It looks like it will provide one of the key functions we need. If you could improve it to allow for repeatability that would be a next level functionality for the machine.

Hopefully, we’ll get this feature one day:
8. Optical Alignment
The dual cameras align the laser head with the frame, with your design,and with your material. Glowforge realigns with every cut and engrave, adjusting timing and position, so every print comes out perfectly.

8 Likes

Thanks @takitus, however, the problem is that the GF can’t seem home to an absolute 0,0 position… relative to the frame / gantry. I think how it works is that the lid camera reads the head position (logo?) and then determines its origin point. Problem there (as I understand it at least) is that it happens every time a job is started. And if there had been any movement or play on the axes over time, the 0,0 could be in a slightly different place every time. However small this difference may be, it’s a problem when precise alignment over the bed area is required - in a repeatable manner like with the placement of a jig.

1 Like

Yeah, you are correct, and I am really familiar with the issue and have been a big proponent of a 0,0 for a while now. I previously proposed a fiducially marked corner piece that would serve this purpose by allowing the head camera to triangulate itself based on the 3 fiducials on that corner piece. (was on the thread you started)

12 Likes

I’ll likely do the same - if only because that’s what I’m comfortable with. But it’s funny - I was working on the Shopbot last night and moving it using the jog positioning function and thinking that it would be nice to be able to just drag and drop the project on the board :slight_smile:

It’s going to be all about specific use cases I expect but I’m getting more flexible.

1 Like

oh i totally agree, we were cutting out foam pieces for my GFs armor last night, and drag and drop was a lifesaver. I never would have been able to do that on the K40. Its definitely one of the reasons that pushed me in the direction of the GF, but I really am going to need precision placement for a number of projects, and lining it up to scores on the tape I put in the bed is really not what I want to be doing forever. It leaves too much margin of error.

You’ve got me thinking of using the back wall of the Glowforge to affix a jig to. I have been using the front and an edge of the crumb tray for pretty good alignment of a jig and using cardboard. Trying to do some rulers and having more exacting standards is making me rethink it. Some type of jig that I can register on the back side of the Glowforge, perhaps using a bunch of rare earth magnets to keep it fixed and use one of my acrylic framing squares as the corner jig.

2 Likes

i wonder if an interim step might be a grid overlay that you could snap things to within the gf ui. it would cover the full cuttable area, letting you consistently put things in the same place each time. i’m not sure if it’s easy to put material in the same place each time, though.

1 Like

Aye, im sure there will be some blueprints at some point. I am going to have to make one just for the fact that the crumb tray has so much play in it. It will probably be a situation where the crumb tray will sit in this large more snugly fitting piece that has 0 play. attaching to the back wall magnetically will probably be essential, and will provide additional support for the passthrough slot on pro units as well.

1 Like

numeric positioning would be a much better solution to this. the worst is snapping to a boundary that is just a fraction off of where you really want to be and cant get to it because of the ‘snap’.

i agree, i’m just thinking of solutions that would be relatively easy to implement and still mesh with the overall obfuscation theme that they seem to rely upon.

1 Like

I think this is great for a general use case.

For a specific use case, it would be awesome if you could combine this with the jig approach @dan has above.

If you can tell the Glowforge “This is a jig”, then you place fiducials it will etch for you. Then, next time you make those part, tell it “I’m using that jig”. It should then look for the fiducials, and automatically align to the parts. That way, it doesn’t matter if your jig is perfectly positioned.

FYI, this is what we do at work with the LPKF laser. (It also does not have a 0,0 position)

6 Likes

Its funny because by using some development tools you can do this in the UI right now to a degree, and it becomes apparent that the implementation of this feature would be incredibly simple. I know theyre trying to knock out the promised functionality now, but im hoping this is something that would be a quick add down the road. If they want to hide it a bit more they can always add an ‘advanced mode’ in the little gear menu that only has one item right now…

3 Likes

I wonder why they use that icon. That’s usually used for Settings. A refresh would seem to want the two-arrowed circle icon that is usually used for repeat or refresh.

3 Likes

The GF would need to use the head camera to find fiducials because the lid camera doesn’t seem accurate enough. We have never seen the head camera do anything other than measure the depth. However it needs to find edges very accurately for pass through and double sided mode, so once that is working I don’t see a need for a jig with fiducials.

3 Likes

I agree

1 Like

Yes, my point is if the head camera is working then jigs aren’t needed and jigs with fiducials need the head camera to be working. It would be very odd if GF got the head camera locating fiducials before they got it working to find edges, so I don’t see why they would ever need to implement this.

It’s going to have to find a lot more than edges to make pass-through work! It also has to sense how far material has been indexed between cuts, and that may not be evident from the material edges alone.

2 Likes