First attempts at engraving image on slate very slow

Hi all,

I’ve recently started to try engraving images on to slate. This is the first time I have used images, as so far, everything I have done has been with vectors.
I am using an Atomstack Pro A20 (20W output).
I initially ran a test on some slate so I could find the power and speed settings. Once I did this, I looked online on the best ways to prep a file for engraving on slate.
Most people said to use Imag-R. I followed tutorials on how to prep the image using Imag-R (converting to greyscale, changing the size and DPI, increasing the contrast etc).
All the tutorials said that once you had finished the prep work, to bring it in to lightburn as an image, but to make sure to choose “pass-through” in the image mode.
I did this and Lightburn, which is normally quite accurate at timings, said it would take 48 mins. Nearly 2 and a half hours later and it was completed. This is on a 100mm x 100mm slate coaster.
I did some more investigating and a few people had the same issue, and found that they had to turn off pass-through and use “threshold” instead.
I did this, but have the exact same problem. The laser slows down so much and it takes just as long as before.
Does anyone have any idea what might be going on here? To my mind, I think the controller on the laser just can’t keep up with the constant on/off of the laser for this image, but it seems insanely slow.
I have attached a small video of how it runs. Bear in mind it should be running at 1875mm/min
https://www.dropbox.com/scl/fi/oqr2rjql4eul9r4zofjjo/SlowSlate.mp4?rlkey=41mq1frcivwa8cflv40x3ug6b&dl=0

Below you can see it runs much more slowly on dark areas

https://www.dropbox.com/scl/fi/9ncvdze46xkozy195jugo/SlowSlate2.mp4?rlkey=9ye2ho4a99g4tpuy58lcx8hch&dl=0

What is the DPI of the processed image and what DPI are you using in LightBurn?

Grayscale to begin with is going to be a very dense in terms of information. Then unnecessarily high DPI settings can further increase the data density. You could just be chocking the controllers ability to process the g-code.

The processed image is 254 DPI, which is the same as the layer is set to in Lightburn

I certainly understand what you are saying, but I’ve seen so many other people do the same thing on the same hardware without these issues. Either I’ve done something very wrong somewhere, or something is wrong somewhere else.

They didn’t post their video of the machine going slow… :crazy_face:

If the image is already dithered, you don’t want to dither it again.

It’s always helpful to tell us what you have, what you want, post the .lbrn2 file, if possible and show a photo of the results. The more you help us, the more we can help you.

Seems like a relatively high interval…

:smile_cat:

You’re saying that at 254 DPI with threshold you had the exact same performance as grayscale at 254 DPI with pass-through? That’s odd.

Are you able to upload the .lbrn file for review?

Picture on Slate 2nd attempt.lbrn2 (361.8 KB)
I’ve attached the lightburn file.
When I first ran the image, I ran it with “pass-through”. The second time, I turned that off and set it to Threshold. Both times I used the same image.
I just feel 2 and a half hours to do a single slate coaster is insane compared to others.

I don’t see anything unusual or excessive in the .lbrn file. I do see that the image mode is set to Jarvis. Did you also try with that?

Images will normally take longer than other burn types but I agree that two and a half hours seems unusually long.

How are you connected to the laser?

Can you take a screenshot of Edit->Device Settings?

The image mode is not set to Jarvis on my file, so I assume that was something lightburn changed when you opened it.

So far I have tried pass-through and threshold. Not Jarvis.

As far as Lightburn is concerned, I am connected via USB (Technically I am connected via USB on another laptop, and I share that USB connection to the network, allowing me to use my main PC in the room next door). I initially thought this might be the issue, but as you can see from my device settings, I am using buffered, so it shouldn’t be slowed down if the network shared USB isn’t running as fast as it should. I also tried running the file directly from the laptop that is connected to the laser and got the same speeds.

This would not be changed. It’s a part of the .lbrn definition.

Can you further elaborate on this? It’s not clear to me how this is being shared.

Is LightBurn running on the computer connected directly to the laser?

I can only tell you that Jarvis is not selected on my copy of the file and I changed nothing.

With regards to the USB connection. I have a laptop, in the same room as the laser. That laptop is physically connected to the laser via USB. I have software on that laptop that allows me to share it’s USB ports over my network. This allows me to fool my office PC in to thinking it is directly connected to the laser.

When I ran a test on the laptop that is directly connected to the laser, I used Lightburn. It is installed on the laptop, but only rarely used.

Try downloading the file that you uploaded. What does it show there?

What specific software are you using?

Can you take a screenshot of Edit->Device Settings on the laptop?

I downloaded the file I uploaded and weirdly enough is shows Jarvis. I am certain that I haven’t used Jarvis though so a little confused on that one. I re-opened the file I originally used on my last test and it was Threshold as expected.

The software I am using is USB Network Gate (USB over Ethernet: share USB port over Network (LAN and Internet))

The screenshot below is from the laptop that is directly connected to the laser

That’s comforting at least that the file is consistently Jarvis. Less comforting that you’re not sure how it got that way.

I’m not familiar with this specific package but it looks like this uses a client-server model. It’s possible that the very presence of this software on the server side of this impacts USB performance. Are you able to disable the server component or attach a computer to the laser that doesn’t have the software for reference?

I don’t see anything particularly alarming in the screenshot.

The other possibility is that your controller is particularly slow. If I’m remembering correctly your laser is based on a 32-bit controller so it should be relatively fast but I may be misremembering and it’s still using an Arduino Nano based 8-bit controller. I wouldn’t be surprised that grayscale performance would be slow but the threshold burning should be much quicker.

Try one thing, can you try burning to the machine using LaserGRBL?

Sorry for the slow reply, I just wanted to go through everything properly.
It seems you were right to suspect the USB sharing software. I’ve no idea why it is running slowly (possibly because the laptop that is doing the sharing is a slow and old laptop), but as soon as I completely shut down that software and ran the image directly from the laptop that is connected to the laser, it completed it within 45 mins.
Thanks very much for suggesting I test these things, I’m just gutted now as this USB sharing software was A: Expensive and B: The only solution that seemed to work for my situation/circumstances.

So it seems for now, if I want to do anything more complicated that simple vectors, I will have to use the old laptop that’s connected directly to the laser, unless I can find another solution in the future.
Thanks again though. I appreciate all your help :slight_smile:

Glad you have it sorted now. Too bad about the expensive software.

I know of a couple of other solutions.

Virtualhere is a popular commercial option:
Home | VirtualHere

USP/IP is a native Linux based offering. It offers a Windows client.

This fork offers is a dedicated Windows server and client but I haven’t used it directly:
GitHub - cezanne/usbip-win: USB/IP for Windows

This is a 5"x7" image on slate.
80W: Speed: 360mm/s Power: 26%.
Prep slate with a quick wipe with a damp cloth.
I use an external editor for images and let Lightburn ‘just’ do the laser part.
It took 18 minutes…

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.