I’m hoping someone here can shed some light on the issue I’m having. The backdrop and story.
I have a small home built diode laser I’m using to etch leatherette hat patches.
Machine Specifications -----
Controller: Arduino using a CNC Shield V3 with GRBL 1.1h
Drivers: TMC2208 Silent Stepper Drivers
Motors: 2 - Nema 17 1.8A@12VDC
Laser: AtomStack 5W Optical Output
Frame: 2020 and 2040 V-Slot Aluminum Extrusion (20"W x 18"D)
I plan on using it while out attending shows to do heat pressed hats on site. It works fine but I’m having an issue with the bi-directional scanning part of LightBurn. I hope the photos will explain my issue a bit better than words can.
I would have assumed the same steps in one direction would be the same in the opposite direction and laser would fire in the correct spot. Obviously that’s a bad assumption on my part.
Although it works in single direction mode it adds quite a bit of time onto a single patch burn. Is there a way to overcome this issue or am I stuck using single direction etching?
Because this is a diode laser with relatively slow mechanical motion, the problem is almost certainly mechanical backlash in the X axis.
Because you built it, you will find it difficult to believe there can be a mechanical problem and it’ll be even more difficult to find. Ask me how I know.
Anyhow, start from the X axis motor and work your way out to the laser head, looking for:
A loose setscrew letting a pulley move slightly on the shaft
A loose screw letting the frame shift slightly as the direction changes
A loose corner joint allowing frame flex
A loose bracket letting the laser head tilt just a little
Judging from the scale, the slop is under a millimeter, so feel for any play that will allow lost motion along the X axis, because your fingers may be more sensitive than your eyes.
If the shafts have flats, it’s possible for a loose setscrew to jam firmly in one direction and feel tight, while breaking free when the shaft turns the other direction and jamming against the other side of the flat. You must loosen the setscrew and tighten it while wiggling the shaft back and forth to settle it in the middle of the flat and snug it down; Loctite will become your friend.
This guide is for commercial Sculpfun lasers, but will give you an idea of where problems will hide and how to find them:
Thank you for your reply ednisley It’s appreciated.
Before I posted my question I did check the X axis belt, the gear tightness on the stepper as well as the tightness of the actual frame that the laser is running on. Nothing is loose. Then I checked the freedom of movement of the X gantry and checked for mechanical play without the belt tensioned. No issue. I also tried different patterns to see if there might have been an issue with the SVG file. Same problem. Guess I should have mentioned this as well in my first post. Sorry.
Just as an FYI, this isn’t my first attempt at building a laser frame. I have one I made that has a frame size of 22 inches in the Y axis and 36 inches in the X axis that I use with my 40 watt optical output Laser-Tree K40 laser for cutting wood and certain colours of cast acrylic.They all work just fine. I also built my own wood cutting CNC machine using Nema 23 steppers and the same GRBL board 3 years ago as well. No issues.
Not to downplay your reply but rather let you know that this isn’t my first attempt. But as I already mentioned, I really do appreciate your reply and hope we can maybe tackle this problem further.
I’ve read about the laser scan offsets but I thought they only mattered on higher speed machines where the laser reaction is not fast enough compared to when it’s fired. 7000mm/m (116.67mm/s) hardly seems fast. (Check the LightBurn file I uploaded) Then again, maybe I’m mistaken?
Well I ran the file as you suggested ednisley. Needless to say the result looked like a bad representation of a Picasso. But I did notice that almost all of the errors were happening in the X axis which was also my scan direction. And they were major.
I had this mounted on the gantry to use for focusing.
I removed it and replaced it with a LaserTree sliding focus plate I bought for my large format laser.
This is the end result of the patch run with the optimizations turned off and bidirectional scanning enabled after I changed it out. It’s as close to what I was after as I could expect.
Guessing there was just too much mass hanging out from the gantry plate so it would sort of swing? (If that made sense?)
Thank you to all of you for helping me out and pointing me in the correct direction.
That’s maybe the second-most-common problem with diode lasers, particularly ones “upgraded” with high-power heads: the original frame is rigid enough for the smaller heads and totally flexy while slinging something with the general aspect of a brick.
The *first-most-common" problem is a loose screw or two, which is why we generally start there.
But backlash is where you find it. Good to know you found it before having to tear the laser head apart to find a loose lens.
