Examples of Working Alignment

That might work, too. The laser-created markings on the cardboard would have to be dark enough to get a good camera scan of them for subsequent re-printing.

It seems @Tom_A has a procedure that could be standardized, too. I’m not sure I understand the details of how he does it.

How about two of these…?

From…

3 Likes

I think it would be hard to measure offset distance with this image. It would be much better to have discrete targets.

3 Likes

I may just start with a big image of a ruler. Then I can compare the original ruler to the scanned ruler to the printed ruler to the re-scanned-and-re-printed ruler… top, bottom, left, right, and middle in both axes. I’d also like to see how “incorrect” material height affects those results, and maybe even offsetting the lid slightly.

1 Like

You wouldn’t be seeing alignment over the whole bed.

To reiterate, if someone is claiming to have good alignment, it needs to be demonstrated everywhere on the bed, not just under the lid camera.

You’re prolly right.

Maybe this grid…?
https://community.glowforge.com/uploads/short-url/3HkT1frkQLH5XWvErS913o34MZz.svg

From here…

2 Likes

I wouldn’t call it a procedure. I just did what I thought made sense.

When testing on cardboard I used bed-sized cardboard flattened with magnets. I put a raster of a ~1" x ~1" targeting reticle dead center (as visually aligned with the rulers in the GFUI), and in each corner, and roughly center of each edge, then sprinkled a few more throughout the bed as well just to see a little variance. YOU SHOULD EXPECT offset at the extremes. That’s guaranteed. The amount of that offset is likely to vary machine-to-machine. But dead center should yield near-0 offset since it’s directly below the camera.

When testing on Proofgrade, I didn’t want to waste any, so I used some :proofgrade: scrap I had and moved the, in my case 4", scrap around the bed, and then targetting it in various places similar to the cardboard and grabbing a shot with the Snipping Tool for each one.

Again, I don’t know that much of it matters. I think the real test is the center-bed test. Can you NAIL it? Does it hold up at >100% magnification? In my first machine I could hit a 1" x 1" piece of material with… well… laser accuracy.

That is simply not going to be possible at this time. Literally not possible. For the claim to be valid, center bed.

I think you’re helping explain why we need to agree on a test pattern in advance. :slight_smile: Otherwise this thread will be full of people nit-picking why any given set of marks isn’t a good test.

But to be clear, what I’m talking about doing would be kind of like this, which I think captures the whole bed:

3 Likes

I have been quite pleased with my alignment from day one.

3 Likes

I strongly disagree. If pass-through and double-sided cutting is going to work, the camera has to be able to align more than just in the center.

1 Like

How much of the bed does this image cover? What are the dimensions? Are those 1" squares?

If you follow that link above the image I spec out the entire test.

1 Like

Thanks for the link. So you’re seeing worst case just under 1/8" misalignment, right?

663

The lid camera won’t be performing that alignment… the head camera will.

That’s what we’re being told…

Correct, at the bed extremes on material that is not perfectly flat. I have noticed that even a very slight warp or variance in thickness will have a large effect at the edges of the bed. The closer to the center of the bed the less warping/thickness variations seem to matter.

1 Like

Another reason the edge alignment is important is the way people want to use their material. When cutting something out of a large piece of material, it is best to position it near the edge to maximize the area of remaining material for future cuts. It’s a huge waste to cut something out of the middle of a sheet.

1 Like

Any test pattern would need to allow space for magnets at all corners and near the center to hold down the material. Most of our material, even straight from the package is not perfectly flat. If it is not perfectly flat across the entire bed your tests are useless.

1 Like

Very true. Flat is important. I hope to see some examples of good alignment on flat sheets of material.

1 Like

I’m not saying it’s not important. I’m only saying it’s impossible today. So testing against it is useless at this time, aside from knowing exactly how far off your particular machine is on the extremes.