On achieving a perfect level

A thin insulating layer between the aluminum frame and the heating layer might help. You still have all the mounting screws that would be transferring heat. Maybe you could replace all the metal nuts/threads with plastic ones. That would slow down heat transfer. I was going to try replacing the aluminum frame with a epoxy/glass one that should be stiffer to start with plus it would do better with the heat.

Regards thickness, the Auto Leveling process is supposed to deal with that. You have a little wiggle room there but would run into problems if it is thicker than 3mm or so.

I think the aluminum frame is the the main problem. It warps with heat. It’s not all that stiff.

1 Like

I just met a user, Bjorn, who told me that he has solved this issue with some pieces of paper.

Just FYI though I am not sure it is the build sheet. As i started troubleshooting this issue i decided to shim the left side of the heated bed. I used some paper as my shims and added 3 sheets or about .3mm to the 3 screws down the left side and retired the print. Right off it appears that this helped. The first layer went down fairly good.

So… Here we are, new firmware took out quick cal, we have a calibration with no heat on. I’m getting the highest deltas I have since the first week, with this new firmware. Has anyone figured out how to level this bed? :thinking:

What i did is turn off auto cal.
Then level the bed using the calibration with the card, but turn on the heated bed with the M140 command,after starting the calibration, while you are connected with your pc over usb. Directly after finishing the level i noticed still some problematic points.
Then i went to the points using the G42 command to go to each point and adjust the offset of every point with the "M421 InJn Q-.05 " and the calibration card, until i got decent results. For activating the the new offset do a homing with G28.(If you want it quicker, then drive to the point of calibration fast using a G0 command, before using the G28)
Dont forget to save the values after the optimization with a M500.
This took me 45 minutes to get a perfectly leveled bed on my A250.

Using the dial indicator is in my opinion not necessary at all. We are not chasing hundreds of a milimeter here.
And additionally if the x-carriage really does some rotating movement, which i could “feel” using the calibration card, the position of the dial indicator is not spot on the point of the nozzle which introduces additional error. So my advice would be, just use the calibration card, because in the end you want the nozzle to have a constant distance from the buildplate, not that dial indicator.

Another thing is encountered: try to fix the heated bed starting from the middle, so first tighten the four screws in the middle and then over cross all screws outside. That minimizes any tension you could bring in by tightening the bed in a wrong way.

I agree with @Razor1990 about using the card vs the dial indicator. It’s too tricky to account for the Z difference between the nozzle and where the indicator is. But I do find the indicator (the one with a needle and dial, not the digital one) useful for scanning the surface to see just how bumpy it is or at least how off the matrix is from the the bed. You can also see the effects of the head torquing when changing directions. Backlash too.
I did a couple of additional tests on how the bed moved with heat at the four corners. It’s pretty close to what it is at the center so I’ve consequently switched to doing my level cold and then using the temperature vs Z movement curve (earlier in this thread) to shift the whole matrix up or down based on the bed temperature.
One other thing I discovered (like @Razor1990 mentioned) is that the M421 tweaks to grid points don’t take effect until some kind of movement has taken place. I assumed it required a “home” command so I would use my “Check Level.gcode” file to survey the grid points and write down on a piece of paper how much to tweak each point and make the changes in a macro.

1 Like

Just so I do this right :

  1. Turn off auto calibration. (3d print settings -> Auto-leveling off)

  2. Run manual calibration.

  3. a) While the machine is waiting for me to calibrate the first point, input from a connected computer, “M140 S60” - where 60 is my wanted bed temp.
    b) Wait 15 minutes for heat transfer into the aluminum frame.

  4. Use "M421 InJn Q-.05 ", and run “G28” in between each use of the “M421” code, and check the level. Note: “-. 05” is the change, and different values can be used but stick to small values.

  5. After leveling all the points, run “M500” to save the values.

Did I miss any steps there?

I find that it takes about 15 minutes of bed heating before the bed stops moving. Maybe even 30 minutes.

1 Like

It is “M140 S60”

In M421 the thing after the Q is how much you want to change the grid point.

Do it with small increments and do a homing in between each change. With this behavior you don’t run into your bed that easy.

2 Likes

You can do a number of M421’s and then a G28 to have them take effect.

@Tone

I have a thread created and this was what I did:

I checked everything and all was mounted correctly and properly. Anyway, I unassembled and reassembled the printer, making double sure everything was assembled and tighten properly again.

Prior to the above-mentioned, I did five level testS.

The results:

Print #1) everything was in a mess. Noticed that the filament is quite flatten. Some difficulty in removing the print out from the bed.

Print #2) recalibrate manually. Very slightly better print. However the filament was not properly adhered at points 4, 6, 7, 8. Very difficult and almost impossible to remove the filament at and around point 9. It was as if the filament has gone into the bed.

Print #3) recalibrate in auto mode. Set Z-offset up by 0.05mm. Print was slightly better and easier to remove from the bed this time. However the filament was not adhered to the bed at point 6 and 8.

Print #4) recalibrate in auto mode, with the slightest resistance when adjusting the Z-offset. This print is the best albeit the filament did not adhere properly at point 8. There were some gaps of the first layer at point 8, tiny gaps near points 1 and 2. Otherwise it is almost perfect.

Print #5) I tried to do a manual recalibration again. I notice every time during manual calibration, when come to the right side of the bed, especially bottom right at point 3, the nozzle will press against the bed hard after pressing 0.5mm down five times. However at the left side of the bed, ie around points 1, 2, 8, the nozzle will only press hard against the bed after pressing 0.5mm down seven times. I immediately paused and stopped the print when the nozzle reached points 9, 4, 3, as the filament was pressed into the bed again.

Overall, I noticed that the nozzle tends to be closer at the bottom right side of the bed, while slightly further from the bed at the bottom left.

The above-mentioned was what I had tried doing earlier. Todate, I am still trying to calibrate the bed and it is still the same issue. I had to “compromise” by lowering the nozzle during calibration so that the furthest points will be close enough for the filament to adhere to the bed later on during print. However the already closer points caused the filaments to be pressing against the bed much more than it should be. I proceeded with my print in this settings such that my first two layers always have the “elephant legs” effect. Is that the term where the first two layers are slightly flatten out than they are supposed to be?

Did you update to the last firmware version ?
Look like too me you still stuck with the inversion problem on the manual calibration point

Yes, I have updated to latest firmware for both the touchscreen and the printer. Have also stopped manual calibration long ago and have been using the auto calibration.

i was looking at my calibration-points-values and noticed that after printing that after a print where i did a Z-change during print, every point was lowered by the value i changed the Z with.

Have you tried using my “Check Level Hot.gcode” file?

@Tone @Razor1990
My intention was to be able to use these instructions to level the bed, but there are clear gaps between what I have listed, and how to actually do any leveling at all. For example, I ran G28 - must be a “home” function, Well, now how do I get back to the point I’m trying to calibrate?

G42 I0J0 F6000
G1 Z5 F6000
Then you are 5mm above the Grid point I0J0.

You don’t need to do the G28 in between each grid point. Just once at the begining.

What I ended up doing in using manual calibration, hooked up via USB. Then, I would calibrate manually, then jog the x/y/z using the computer, then check the point. Sometimes there would be slack, sometimes it was good. I started a multiprint over almost the whole surface, so… fingers crossed?
It really feels like some of us heavy users have gone to great lengths trying to level this thing… it’s kind of discouraging. :sob:

1 Like

My main problem was the underrated extrusion. Check this!
I had to get my extrusion steps/mm from E212.21 to E238.44
now its way better!

1 Like

Add that to your list of things to do after opening the box!

1 Like