Glowforge replicator head!

Although if you look at something like a Kinect which uses a stochastic dot pattern it does a pretty amazing job of depth without an extraordinary resolution. Stochastic turns out to be better in most cases than regular patterns (kinect 1 was regular grid IIRC). You can make pretty small dots with the laser (because if something is smaller than the laser can focus, do you care? On what technology would you reproduce that?) and especially if you are willing to make multiple stochastic patterns (you can turn the pattern 90 degrees for instance to get around all sorts of defects in your scan). Of course I have no idea if that’s how the GF does scanning internally, but it’s a tried and true solution in many other applications.


Ok…now I’m hungry…


The one next to the chocolate covered one was stuffed with something savory…ham and cheese and herbs, I think. Let’s see and Eggo do that!


Which was what I asked at the beginning of the thread :slight_smile:

British is best! :smiley:

I’m not convinced that it is better but perhaps 1st hand use will provide that. I suspect that the need for limit switches depends at least a bit on what one intends to do with the GF. For some it be might be essential and for others of no concern at all.


Sooo cool!

So far for me it is not better. I have a conceptual idea of an iterim fix that I could potentially use to make this work. A virtual and digital jig of sorts. Hopefully I can get around to messing with it soon to see what kind of margin of error to expect. It will for the most point make interacting with the optical alignment part of the glowforge workflow unnecessary, and everything will have to be done in illustrator. I hate to have to do it that way, but I can enter numeric offsets, sizes, etc in illustrator.

Again this is only for engraves that need very specific placement. It might be something they add into the UI soon.


I’m really looking forward to your progress in this area. Dealing with a virtual jig seems like a sub-optimal solution to me but maybe it’s not as bad as I’m thinking.

1 Like

Agreed. Currently it might be the only way to get any repeatable accuracy.

So for making this vjig, what I do know right now, is:

  • The bed area in the UI is always the same. The image loaded from the camera can change with material depth, but is irrelevant if we ignore it.
  • A physical square of material on the bed shouldnt move too much if at all once placed if attached to the bed. Removing the bed will change this alignment.
  • The head always starts at roughly the same spot (should easily be sub-mm margin of error if that at all)

What I dont know right now:

  • When a full bed sized svg file is loaded into the UI will it always go to the same place

If a file always goes to the same spot when loaded (top left corner) I might be able to make it work.


So far sounds pretty promising. Like you, I really would like to specify the origin of a file with precision. Maybe it’s because I’m still dragging the history of how “I’ve always done it” but without a GF I can’t tell how I get repeatability without it.


Well theres a reason that is the case, and thats because it works repeatably. Right now theres just no way to do it that weve found, especially if you are scaling an image at all. Once you close the window youll never get it back to exactly the size or position it was previously at. If your browser or the app crashes, youre out of luck. If you take a piece out of the bed and realize, ‘hey i think I should engrave these holes a little deeper’, youre out of luck and just have to cut a whole new piece.

In a lot of the projects Ive made, especially the Bioshock one, there was a LOT of waste because I just couldnt go back and change a piece I had already made. I had to make a new one and hope it worked out, and if it didnt, I just had to make another new one. So for the ease of use in the workflow its really a tradeoff in material waste right now. Luckily this is software and there is a lot of room for change.

There was the advice of making a jig out of cardboard, but the problem is when you put a sheet of cardboard in, and cut out the perfect slot for your item, you still have to get that piece of cardboard out of the slot for your item to go into. so you have to weigh down the rest of the jig, and stab it from above and hope it doesnt shift anything on the way out. AND you have to do this EVERY TIME… you also need to cut with low enough power to keep the edges from burning/charring/burning back, but high enough power so that it goes all the way through so that when you try to pick it up, it dosnt shift the jig rendering it useless.

So theres definitely some exploratory work left to be done to find a good solution


All of those are use cases that happen to me (& likely every other laser user). Nothing ever cuts perfectly the first shot. Or even after you’ve run a test cut and tweaked the settings. Being able to re-do a part of what I’ve done already (your deeper holes example) is something I do as well. It’s the difference between needing to re-do the whole project or just a few more minutes with another pass in the machine.

Part of the reason I went CAD/CAM was for the precision on the second time through with something. Otherwise I’d be an artist with hand-tools and I’m way too old to be that :smile:


I’m reminded every time someone posts a “me too” post about one of my comments and gets more likes.


Well, this time, your post got 12 likes and my response only got 11.:stuck_out_tongue_winking_eye:


It should go without saying, but please keep us posted on additional testing or software improvements that get released to your GF for this capability. It’s looking like an issue for my typical use.

well @dan posted this little gem early this morning which could potentially be helpful in solving this issue:

Really cool of them to do that on the fly. Ill try to test tonight and see what happens. Of course will update here as well


Thanks - I hadn’t gotten to that one yet. That does sound encouraging and it will be interesting to see what you think of that approach after you’ve had a chance to try it out…


me too !