Just curious if you have tried an 11 X 11 bed level? Has to be done from a computer hooked to your SnapMaker then gcode it.
@gwfami I haven’t. Is there a guide for how to do this that you could direct me to?
Try this link.
You will have to have a computer hooked up to the SnapMaker via a USB cable. I use a terminal emulator to connect to the snapmaker, but you can also use Luban to connect to it and then enter the commands in the workspace console.
When I do it I enter the following commands.
G1029P11
This sets it to an 11X11 leveling grid. The P value is the grid points.
G1029A
This begins the auto leveling process. It checks each point 3 times, so it will take a little while.
The head will stop at the last point and require you to do the last point manually, just as you would normally do.
G1029S
This saves the leveling matrix.
G1029D0
Never use this one, but the link says it ends the leveling processes and turns the auto level on. Not sure if it’s necessary or not but doesn’t affect my printing at all.
M420V
This shows you the just created leveling matrix so you can see it has changed. This command is why I used an emulator, as Luban doesn’t allow me to copy the matrix from the console (using a Mac). The emulator allows for the copy command. I copy it and put it into a file for later stat analysis.
One thing that I discovered is that if I do the final z offset when leveling according the the instructions, that when I pull out the leveling card, the head drops a tiny bit. This could be part the problem that you are running into. What I’m doing now is to do the finally z offset when the card just begins to catch. If you do it so that it pulls in one direction and wrinkles in the other, as per instructions, it seems to cause the head to drop.
Here’s a long post that I started when I began to go bald from pulling my hair out.
I eventually had to get SM support involved and they sent me a new print module.
Hope this helps.
G1029D0 is indeed important. Running that synchronizes the planner, computes a final offset of all points (D0 offsets by 0, but that could be changed if necessary), recomputes the catmull subdivisions, and enables bed levelling saving a reboot.
If you need to move everything up or down by say, 1mm, G1029D1 would do that. I forget if that moves it up or down, some experimentation would be in order.
Thank you @gwfami and @brent113. I’ll need to go through everything in more detail after my deadline (Friday morning) when I can experiment in leisure. The good news is that I was able to print my project parts and I am so happy. It’s not perfect and I had to spend most time on just getting the SM2 recalibrated (…etc…) instead of adding more adaptations to the design, but I think it will work for now. I’ll know for sure Friday.
In the meantime, I’m not sure I understand fully why I would want to use the backend for bed leveling rather than the auto bed leveling app. I posted a picture of the auto leveling app in the controller further up in this thread based on a similar question someone had. It looks like a bigger matrix than 3x3. So, which matrix does the auto leveling use for A350s? 3x3 or something else? I couldn’t find information about the auto bed leveling method (ABL) for SM when I was going through the Teaching Tech calibration steps, which is one reason I chose to use STLs instead of the generated gcode on the website.
1 Like
You’re doing a 5x5 right now.
The bed is not perfectly flat, and some people have done some work to characterize the bed to find out theoretically what bed levelling grid would be required to perfectly compensate it. The theoretical answer is somewhere around a 70x70 grid as the bed is really really not flat.
Unfortunately the controller does not have enough memory to store a grid that size as well as no one having the patience to sample 4900 spots on the bed 3 times each, but it can store an 11x11 grid and 121 samples is more reasonable, so that’s what many people use. It more closely follows the hills and valleys of the bed and improves the first layer.
Perhaps it’s just old age or senility on my part. I must have done this as the catmull subdivisions are calculated and saved and autoleveling is on. Sorry about any confusion.
Oh, and the backend gcode approach is necessary as the app doesn’t have 11X11 as a choice. I think it only goes to 5X5.
It’s not strictly necessary as all of that would have been done already, or on the next machine boot. I think it’s good to plan on doing though.
Hello @gwfami and @brent113. I am finally getting back to where we left off about the bed leveling. We received the upgraded rails, print head, and power supply so I was waiting until we installed everything to revisit this.
For the 11x11, should I be doing this with the bed cool or heated (like the auto leveling done through the screen)?
I’ve done it both ways. Problem is that the G1029 command shuts off the bed and extruder heaters, so things may cool down before you are finished. My suggestion is to try it first without the heater on, but others may disagree.
1 Like
I calibrate without heat as well, only because it’s worked for me. Not that it’s better or anything
1 Like
I’ve had better success with turning heat on. (using touchscreen, warm to 10-15 over printing bed temp)
Most but not all seem to.
Everyone seems to have different experience.
Try both ways and see what works better for you.
Might want to search forum for “tramming” if there’s a definite difference where left or right is higher (although calibration is supposed to take care of this)
-S
if you think heat is the issue do a 5 x 5 heated calibration from the screen and see what that gets you…
Thank you everyone. I’ll try the 11x11 again once I can attach my other PC to the printer. Connecting via Luban/wifi won’t let me complete the steps on the controller screen. Having another issue with the controller that I’ll post in a new thread.