We're sorry, an unexpected error has occurred

Community Support
15

Don’t quite know how to explain this…increasing the LPI doesn’t seem to make much difference to the way an image is engraved once you get past a fairly low number. (Seems to be in the 195-225 LPI range.)

The higher resolutions on the LPI are burning over areas that have already been burned as the beam moves up a shorter distance. So the resulting engrave gets a little more charred, and a little deeper, but the coverage for the engraving is exactly the same as it is at about 195-225 LPI. (And I know someone can probably calculate the exact LPI number for full single coverage given the size of the spot of the beam…I’m not going to take the time to do it.)

So when you’re interested in creating a burn pattern, you can get the same pattern at the lower (195-225) LPI values that you do at 340 or higher. 340 LPI and higher just get darker and deeper, and take a heck of a lot more time. And generate a tremendous amount of additional data for the buffer to have to deal with.

So for larger surface area engraves, I will reduce the LPI to about 195 or 225, and then, if I absolutely want to have it be charred black, I can run the engrave part a second time without changing any settings after it finishes the first round of processing. It’s still faster than upping the LPI value. (That introduces a factor of four to the time required, if I recall what Dan said about it once. Might be wrong…don’t quote me.)

When you get into the draft settings for LPI (extremely low end - under 125) you will start to see gaps in the coverage, so you don’t want to go too low with it. But since no one likes to wait around forever for their engraving to finish, it makes sense to run it at the lowest LPI that gives you complete coverage.

The extremely high LPIs really do eat deeply into the material, and they’re useful if you are trying to actually carve out material. My thinking was that gets used during the 3D engraving process. (And I use them to carve out channels for seating butt joints.)

The laser will pick up tremendously small detail at 300 PPI. It really doesn’t improve the image resolution to go larger than that, because the beam is a fixed size. And too much pixel data can bog things down again.

So for larger surface area engraves, I start with 300 ppi on the raster image, and 225 LPI for the engrave, and generally the interface doesn’t have trouble accepting it. If it’s still too much, I’ll drop the LPI to 195 and try again.

Don’t know if that will help or not, but it generally works for me and even though I don’t do a lot of engraving myself, I’ve checked a lot of problem files that couldn’t be loaded until the LPI and PPI were dropped down to manageable levels.

9 Likes
16

Thanks so much for that thorough response! I’m really interested in how more passes will affect engraves, too, so I’m gonna go do some testing immediately.

3 Likes
17

They advertised 1000LPI but the machine uses metric belts so it can never get exact inch units.1354.66666667 is the first value > 1000 with 400 step motors, 30 tooth GT2 pulleys with the stepper drivers in x8 mode. They can go up to x32 mode but that would make the waveform file much bigger and reduce the maximum engrave time even more.

18

Don’t forget they advertised 1/2" variable focus but they use 11mm (.433) [I think that was actual, might be 12mm, can’t remember off hand]

Glowforge likes to round and redefine words.

2 Likes
19

They still are here.

“Completely Internal — Lens moves internally up and down inside the head to focus on materials up to 0.5” (13mm) thick”

Right now focus only goes up to .433". I wonder what the issue could be that is preventing them from enabling the full .5".

2 Likes
20

Yes I mentioned that in another thread. The GF works in metric units, they only get converted to inches or obscured in the GFUI. 0.433" is 11mm so they are 2mm short.

The Z motor has a limit switch so I don’t think they have uncertainty where it is like X & Y. I wonder if they simply mounted the head too low with respect to the crumb tray. If that is the case I can’t see it being fixed. If it isn’t the case I don’t understand why it isn’t fixed already. It is a very simple thing to drive the motor up and down 13mm from a limit switch. It can either move that far or it can’t.

21

Just wanted to add one technical detail since I’m not sure it was stated directly. I may have missed it since this topic went the usual way: we all pick a side and throw rocks at each other.

When you’re doing a raster engrave, the content of the image is irrelevant from a time/complexity perspective. A black filled square is the same as a black outline is the same as a an ornamental filigree of the same size. What does matter is the physical size and lines per inch, as that affects how big of a file is eventually sent to the machine.

I mention this because you talked about simplifying your design quite a bit. I may be misinterpreting what you meant, but just in case and for future reference, that shouldn’t be necessary. Printing at lower resolution for large images, however, may be.

Hopefully Glowforge will be able to remove this limitation soon. They clearly have a large backlog of features to get through.

2 Likes
22

And the speed!

The file is 10KB per second regardless of the content, so the total time of the job is limited by the 100MB buffer size at the moment to around 3 hours. There seems to be other size / complexity limits in the cloud as well though. For example I think large vector paths choke it but they don’t take long enough to fill the buffer.

1 Like
23

It does depend on the material. For things like wood that are being vaporized and leaving behind a charred and smoking path of destruction, sure, the damage tends to spread out even at low density. But as I posted in the thread on laser tile the other day, this material (mostly) just changes color where it’s hit by the beam with hardly any smearing out. Even at 450 LPI in this camera phone picture, you can make out the individual lines on the end.

IMG_8523
IMG_8523816×612 111 KB

I trust the math others have done as to the mechanical limitations of the machine: that’s not my strong suit. I just know that not everything is created equal when it comes to the minimum distinguishable LPI.

5 Likes
24

Fair enough. I also didn’t want to complicate things but there are of course multiple choke points. It’s possible to successfully produce a file that’s too big to fit in the machine’s buffer, but it’s also possible for the cloud backend to fail to generate the file by running out of time or resources due to complexity, or as we’ve seen, SVG features it can’t handle.

Actually, rereading your description from yesterday, it’s not clear that I was correct in my initial statement. You mentioned that one of the bits indicates whether it’s motion data or laser power. What happens if the power doesn’t need to change? Would a solid grey square require half the data as one that’s a different level on every “pixel”?

1 Like
25

The laser PWM value is only used once each time you set the power for an operation. If it is full power the power is set to an arbitrary value, presumably the correct value for 40W for your particular tube. If you use precision power then it is set to 127 to disable hardware PWM and the laser enable bit is used to manually PWM the laser to give the selected power.

So it only ever uses one bit per 100us sample regardless of the mode. It has more bits per pixel because a pixel lasts for many samples. It actually PWMs with an 8 sample cycle at 1.25kHz. That only gives 8 levels but it dithers to get more.

4 Likes
26

Yes the beam has a peak intensity in the centre and tails off away from it with a Gaussian distribution. So it doesn’t have a hard diameter. You can overlap the fringes of the Gaussian curves. I expect the optimum overlap is when the two fringes add up to the same intensity as the peak. You then have a residual ripple at twice the LPI.

3 Likes
27

Absolutely! I saw your testing on the laser tiles, (going to come in handy…thanks). But you bring up a good point that when you are using the laser to do something else besides engrave through ablation, the rules might be different, and testing will be required.

Each material has it’s own properties, it’s not a “one size fits all” proposition. In the case of the coating on those tiles, it looks like exposure to the beam is darkening the resin, and subsequent or lengthened exposure continues the darkening process. The 450 LPI definitely looks better. :slightly_smiling_face:

3 Likes
28

Wow, your design is stunning! I’m so sorry you are having trouble printing it.

There’s a problem with our software in handling very large engraves and it looks like that what’s happening here. Your message will help us improve our software for everyone.

I’m glad you were able to print by reducing the LPI. If you’d like to print it with a higher resolution, I’d suggest you divide your image into pieces and print them one at a time, as shown below.

Split your image into pieces

  1. Save your design as a PNG file
  2. Go to imagesplitter.net and upload your file
  3. Click on the “SPLIT IMAGE” tab and enter the number of rows and columns you want to split your image into
  4. Choose PNG
  5. Click the blue “SPLIT IMAGE” button. The software will split your image into the number of rows and columns you specify and automatically download a zipped file of the pieces.

Create a single file with all the pieces

  1. Locate the file on your computer (where your downloads go) and unzip it
  • Mac: Double click on the file to open it
  • Windows: Double click on the file to open it, then click “Extract All”
  1. Open a new file in Inkscape (used for this example, although other software will work too)
  2. Choose File > Import and select all the images from the zip file
  3. Line up the images up so they are seamless. In Inkscape, when you drag the images near each other, they will snap together. (If they don’t, go to View > Show/Hide > Snap Controls Bar and adjust the settings.)
  4. Select File > Save As and save the file as an Inkscape SVG file

Upload and Print

  1. Sign in to app.glowforge.com, click “Upload” and choose the file you saved. Each piece of artwork will import as a step in the app. Click on a step and choose “Ignore.” Ignore all the steps but one, and then press “Print.”
  2. When that print finishes, leave your material in place
  3. Set the step you printed to “Ignore” and print another step
  4. Continue ignoring and printing until you’ve finished

Please let me know if this helps!

2 Likes
29

This 14-step process is a TEMPORARY solution to a TEMPORARY problem that will be fixed, right?

2 Likes
30

Well, cutting-and-pasting-in-place on a separate layer in Photoshop was definitely much easier, but thanks for this! What I just learned from this: You can save a layered SVG and import it as multiple operations. Cool.

Still, some part of me almost wants to agree with the Mr. Negativity-and-Sarcasm guy - I’ve found I’ve had to use this method on all 3 of my most recent projects (even after reducing LPI and following other tips in this thread), just to get it to print. I’m actually still trying and failing on one. It’s pretty frustrating and time-consuming, and I hope GF can put a dent in this problem soon.

4 Likes
31

Middle name: Disappointed

3 Likes
32

Looking past who said it, yes, you should be able to drop a high res photo that all you have done to it is adjust the brightness and contrast. Then print it full sheet. Anything less is not an acceptable long-term.

At least with indexing and material flipping there is nothing there so we know it is not finished as it is not even started as far as the end user is concerned.

As much as I love the :glowforge: and where it looks to be headed, engraving leaves me with a quesy feeling.

33

Indeed - the current limitations are frustrating for me and the team at GFHQ as well. It’s high on our priority list.

9 Likes