Amost there- Eleksmaker A3 Pro

[troth530]

Jeff and John have been tuning me up quite a bit. But there is still quite a bit of work to get it dialed in.

Machine -
Eleksmaker A3 Pro with 2.5W laser module
Mana SE controller with PWM output
Added pulldown resistor on PWM output to eliminate firing during startup
TMC2130 upgraded drivers voltage set to 0.8V X and 1.6V for Y (still experimenting)
Air Assist
Fan for cooling controller
Upgraded optics (3 element AR coated lens at 3" focus (per Jeff's suggestions)
$30=255 Max spindle speed, RPM (255)
$31=0 Min spindle speed, RPM (0)
$32=1 Laser mode, boolean (1)
$120=500.000 X Acceleration, mm/sec^2 (800.000)
$121=500.000 Y Acceleration, mm/sec^2 (800.000)
Steps/mm is 80.00, will change to 40.00 when using 8 microsteps (later)
$130,131=5000mm/min max

To Do: Homing/Limit switches, Upgraded Gates belts

Currently I've loosened the belts to debug the canvasing effect.
The motors may be a little weak, as I think detents of the steppers are overcoming the holding force of the steppers.

I'm using PEP6 with Analog PWM, and 25% frame rate change and extended frame of 25 pixels. The material I'm testing with is scrap MDF from Home Depot that was in a shopping cart on its way to the dumpster when I asked how much and the guys generously said "Take it". The material is a very light material that burns very clean (not a lot of variation) and is perfect for testing. For this material I'm using 2800mm/min and 0.1mm pixel resolution on a 45Deg angle.

Focus, Focus, Focus
With the upgraded optics installed it took quite some time to find the focal point. I used a piece of black anodized al to prevent scattering and then found when the dot went from big to small. I was only able to get about a thread in, so I used gas pipe teflon to help it stay stable as regular teflon was too thin and it would just wiggle around. Then once I got it close by eye, I used a test burn to find the exact point (I guess I used 1/8 turn increments or less).

Flatness, Flatness, Flatness
I guess the focus depth of the lens is narrow because I needed a really flat stage. I got a table which was an old smog machine based and a scrap piece of marble from Habitat for Humanity both for $20. I still have one corner which is definitely off a little, which I'll be fixing today. I found I need to do test burns in each corner.

Still fighting backlash- new gates belts are on order. Under Special options there is Align X and Align Y that Jeff pointed me to that can do the job for sure, but I'm still messing with my belts, so I figured I concede some run time until I get a stable setup. To do so, I wrote a python program to only engrave in 1 direction and it would do a rapid movement of 5000mm/min to efficiently get back to the start point (reduce run time from 2x to 1.4x or so) as I felt it would take some time to tune in the backlash settings.

[troth530]

Apparently this was done with only 1 of the 2 y-motors plugged in as I just found out while trying to shim all corners.

[Picengravertoo]

Nice setup.

I use Kevlar reinforced timing belts on my CNC machines, but I have the MXL & XL pulleys instead.

Have you tried changing to these grbl settings?

$0=30 (step pulse, usec)
$1=255 (step idle delay, msec)
$11=0.005 (junction deviation, mm)

Jeff

[troth530]

Hi Jeff,
Yea, not bad for a $20 payment to Habitat for Humanity which is also tax deductible. Also it was a really worth the money since it only took about an hour to setup and gives a really flat surface. The material I'm using is warped, so I'm taping it down but still affects the results. I had a friend over today and we discussed why this MDF burns better (darker and wider range or tone) than birch, and we suspect there's something in the glue. Most of the MDF I've gotten in the past has been darker, which is not as good because I like the contrast and rich tones. I'll go by Home Depot to find out where they keep it on the shelves. I'd really be bummed if I can't find it. As I said, what I found were strips in a shopping cart slated for the dumpster.... crossing my fingers.

I'll look into those $ settings, I'm assuming those settings may be able to increase the holding time of the movement to help fix the canvas issue??? Those actually may be the last 3 settings I haven't tinkered with yet.
My current settings are $0=10, $1=25 and $11=0.010, for reference.

BTW, you can see in the picture I posted the white connector on the right hand side dangling out in space intended for the second Y motor. Oh well, I'll set up another burn and do some more testing tomorrow.

I have gates fiberglass reinforced GT2-6 belts arriving Wednesday. Belts and pulleys are tricky as they have to be paired and there are many types. I'll also bet the machines that you're using have stronger motors as well. That may be a limitation of my machine, but I still think there's a lot of user error involved, so I won't place full blame the machine just yet.

I appreciate all the help you've been giving. As I'm getting the setup dialed in it's been really helpful to have the benefit of your experience.
Thanks.

[Picengravertoo]

Here is our Shapeoko 2 laser engraver setup.

https://www.instructables.com/id/Shapeoko-2-Arduino-UNO-R3-grbl-9g-8bit-Raster-Phot/

There has been some modifications to it since. It has the J-Tech 3.8W laser kit mounted on it now.

[troth530]

Hi Jeff, my holding torque is 0.28Nm. Do you have the part number or specs on your motors. My guess is to have 2x to 4x holding torque than I do. LightObject which is within driving distance have 0.57Nm motors I'd be willing to try.
Thanks.

[Picengravertoo]

I'm not at the shop right now to confirm, but I believe this is the Shapeoko 2 stepper.

https://www.inventables.com/technologies/stepper-motor-nema-17

Here is the stepper driver I use that is plugged into a Genuine Arduino UNO R3.

https://www.inventables.com/technologies/gshield

[troth530]

Hi Jeff, that's a 62oz-in motor (or 0.43Nm) that you're likely using. Definitely a better motor and driver. Not sure if I'm on the right track, but willing to try a few things.

[Picengraver]

I strongly suspect your "canvas" effect is not from the stepper motors. I think it is possibly being caused by out of true pulleys. I bought some cheap pulleys from China several years ago, and had to throw them away they were so out of true. Jeff has a source of quality pulleys, 18 tooth as I remember. I have attached a simple test image that you might try. The lines are one pixel wide. Try engraving it in both horizontal and vertical directions and looks for backlash/alignment errors. Do not convert it from a bmp to a jpg file as that will introduce compression artifacts in the image.

And a Happy New Year to All,
John

[troth530]

Hi John, could be. The pitch of the lines of the canvas effect do not align with the detents. My stepper motors are 1.8deg which is 200 steps per revolution and the pulleys are 20 teeth at 2mm pitch, so that's 40mm/200steps or 0.5mm per detent. The "canvas" lines are more like 0.8mm. I definitely have a backlash issue which is why I only engrave in 1 direction. I'll start with new belt and new pulleys, as I have a new belt coming in tomorrow. Once I replace the belts and pulleys, I can then work on backlash and characterize again.