To see this working, head to your live site.

Edited: Oct 27, 2020

After calibrating Z offset- the first layer still prints in the air

Im sure most people have stopped looking at this forum now but its worth a shot still.

I did have my printer running ok ish. i made some upgrades to the linear rails to make it run smoother. after putting the printer back together and recalibrating everything as per the set up guides. the printer still trys to print the first layer too high. about 0.5mm too high. i have managed one or two prints but putting an extra manual offset before printing but this isnt actaully fixing the issue.

i have tried recalibrating the z offset about 10 different times and still the same issue. has anyone had this problem and managed to correct it?

i dont think it is a mechanical issue as i replaced the stepper motor with the one i used for the 2nd extruder. So it must either be with th slicer or marlin. however the G.code looks fine to me. and i havent made any changes to marlin sinse the original set up.

3 answers4 replies

7 Comments

Johannes L

Nov 01, 2020

There is a Z_PROBE_LOW_POINT setting in the firmware (configuration.h) that was also changed by makertech3d a while ago. I'm not completely sure about this, but if I understand it correctly, it limits the maximum range of the Z-probing from the printer's original z=0.0 point. If that's the case, it might limit probing into the bed when auto-leveling. It's currently at 3mm which should be sufficient if you follow the procedure from the build manual when mounting the z-probe sensor (resulting in about 1mm distance between sensor and hotend). It might fail when the sensor is mounted too high.

Jay M

Nov 01, 2020

Replying to

I had one corner that would not probe after I increased the number of probe points, so I had to change the point from 3mm to 4mm if bed leveling with a hot bed, and 5mm if leveling with a cold bed.

Torsten Unger

Oct 27, 2020

Hi,

I had the problem of 'printing the first layer in the air' several times after calibrating and fixing the printer. As I assume meanwhile, it has to do with the distance between probe and nozzle. I guess when it's more than a millimeter, the Z-Offset does not compensate it.

Jay M

Oct 27, 2020

For me, increasing the number of measurement points for the bed leveling and then turning on babystepping and really dialing in the Z-Offsite through the Tune menu while printing, I've been able to get the first layer working well.

Both of those things are in the configuration.h Marlin firmware.

sebastian waters

Oct 27, 2020

Replying to

Thanks for getting back to me @Jay M. When i say i havent updated marlin that was a bit of a lie. when i first encountered this problem the first thing i did was increase the probing points. i currently have it at 6x6.

However i am now looking at adding in babystepping as well.

But even with these changes i dont see how that fixes the issue. if the initil z-offset is set up correctly then why when it goes to print, it doesnt go to the correct z-offset point?

Jay M

Oct 27, 2020

Replying to

@sebastian waters - I'm at 7x5 for my mesh grid. I think the magnets MIGHT impact things? So maybe 6x6 is hitting or not hitting magnets? I don't know, but that is working for me. I was going to try a denser mesh soon.

I'm also using "sub-calculation measurements" and do 3 of those per measurement.

I also want to turn on the "extend beyond measurement points" get make sure my top edge where you can't probe since the inductive sensor is off the plate follows the current calculated trajectory.

My z-height adjustment process is generally:

1) With power on and the flex plate on (or whatever you're going to print on),

2) M502 to reset all the settings to factory, M500 to save them.

3) Disable all the steppers (in OctoPi for me, Prontoface works, maybe M18 works)

4) Manually position the hotend in the center of the plate on the X and roughly 5cm "south" of the center of the bed on the Y by hand

5) Lower the gantry down by hand until either the hotend is closer than 1mm (I use a .6mm feeler gauge myself; or the inductive sensor lights up.

6) Adjust the sensor up/down until the head is <1mm, and >.5mm (roughly .6mm for me is what I've been using) and the sensor light is just turning off/on with any gantry up/down movement in that 0.5-1mm range.

7) G28 to Home all axis

8) G29 to run the 7x5 autoleveling

9) M500 to save the mesh

10) G28 to Home all axis

11) M420 S0 to turn off the level/fade

12) M114 to make sure I'm at Z0.0

13) I use OctoPi (technically OctoPod on my iPhone) to move down 0.1 at a time until a folded piece of paper is just starting to grip well but not completely STOP the paper from moving under the nozzle. This is often -0.4 or 0.5 when checked with M114 again. I will also say, that I can sometimes feel the head hit the paper but still move the paper quite easily... so I generally go 0.1 more... but if really stops the paper at that point, I go back up 0.1 and use that M114 value.

14) M206 Zx.x to store the value. As I said, I often end up with this being "M206 Z0.4" ... notice the lack of a (-) negative sign. I missed this the first couple times and that was why my Z was too high.

15) M500 to save everything.

16) I generally turn the power off and unplug the OctoPi at this point, and wait couple of seconds and then turn everything back on.

17) I start a print here. I have a 80x80x0.2mm square that I use for this.

18) Even at the skirt if I see it not sticking, I click the LCD selector wheel, go to Tune, then Z-Offset and start going negative. Past couple of times I've needed to go to -0.125 or -0.150 to get a good squish.

19) even with this, the "upper left" or "northwest" quandrant is higher for me, and it squishes too low, where almost invisible (thus creating pressure in hot end I assume), but my prints are coming out ok.

Hope this helps!

Good luck!

Jay M

Nov 01, 2020

Replying to

I've now went to 15x10 measurement points. Turned off the Bilinear Subdivision and kept the Extrapolate Beyond Grid on. In my 80x80x0.2 first layer square I still had a couple of high points, so figured this should really help the accuracy and it's a one time thing essentially, so why not.

Here is my grid now...

Recv: Bilinear Leveling Grid:
Recv:       0      1      2      3      4      5      6      7      8      9     10     11     12     13     14
Recv:  0 +1.502 +1.002 +0.970 +0.788 +0.635 +0.500 +0.375 +0.192 +0.402 +0.513 +0.592 +0.788 +1.045 +1.200 +1.580
Recv:  1 +1.407 +1.062 +0.825 +0.638 +0.470 +0.327 +0.255 +0.250 +0.242 +0.327 +0.415 +0.595 +0.855 +1.132 +1.442
Recv:  2 +1.287 +1.015 +0.730 +0.545 +0.427 +0.282 +0.207 +0.192 +0.172 +0.232 +0.332 +0.505 +0.765 +1.077 +1.360
Recv:  3 +1.170 +0.870 +0.630 +0.430 +0.297 +0.225 -0.090 -0.115 +0.073 +0.212 +0.280 +0.430 +0.678 +0.868 +1.230
Recv:  4 +0.703 +0.448 +0.558 +0.380 +0.262 +0.175 +0.100 +0.007 +0.095 +0.142 +0.225 +0.392 +0.582 +0.520 +0.952
Recv:  5 +1.097 +0.713 +0.538 +0.287 +0.147 +0.052 -0.015 -0.017 -0.032 +0.040 +0.102 +0.227 +0.442 +0.683 +0.968
Recv:  6 +1.055 +0.730 +0.445 +0.232 +0.075 -0.040 -0.085 -0.118 -0.085 -0.005 +0.067 +0.197 +0.400 +0.685 +0.988
Recv:  7 +0.940 +0.625 +0.297 +0.030 -0.130 -0.250 -0.297 -0.322 -0.330 -0.270 -0.187 -0.057 +0.145 +0.410 +0.755
Recv:  8 +0.730 +0.565 +0.260 -0.017 -0.167 -0.307 -0.568 -0.598 -0.518 -0.380 -0.273 -0.120 +0.005 +0.252 +0.680
Recv:  9 +0.553 +0.442 +0.157 -0.075 -0.295 -0.420 -0.565 -0.670 -0.648 -0.510 -0.413 -0.282 -0.100 +0.135 +0.538

Here is my mesh bed leveling config now...

#if EITHER(AUTO_BED_LEVELING_LINEAR, AUTO_BED_LEVELING_BILINEAR) // Set the number of grid points per dimension. #define GRID_MAX_POINTS_X 15 #define GRID_MAX_POINTS_Y 10 // Probe along the Y axis, advancing X after each column //#define PROBE_Y_FIRST #if ENABLED(AUTO_BED_LEVELING_BILINEAR) // Beyond the probed grid, continue the implied tilt? // Default is to maintain the height of the nearest edge. #define EXTRAPOLATE_BEYOND_GRID // // Experimental Subdivision of the grid by Catmull-Rom method. // Synthesizes intermediate points to produce a more detailed mesh. // //#define ABL_BILINEAR_SUBDIVISION #if ENABLED(ABL_BILINEAR_SUBDIVISION) // Number of subdivisions between probe points #define BILINEAR_SUBDIVISIONS 3 #endif #endif