Elegoo Mars 2 Pro Calibration XY and Z

[nighteagle]

Hello Printers,

First my Workaround and Printer/Settings

Printer Elegoo Mars 2 Pro Firmware: 4.3.13_LCDC/ 1620x2560 /F2.23
Slicer Chitubox Version 1.7.0
Calibration Settings XY:  X:82.65mm and Y:130.42mm (My Prints have in XY +0,01mm to 0,03mm depending  from the size of the models in xy)
Calibraion Setting Z (Default is 0,000625 mm/Step): Configure 0,000622mm/Steps ( over the gcode-file - expalin later)
Calibration Setting Z=0 (Default is M8084 Z0.000000; Set Z Offset ): Default is M8084 Z-0.03; Set Z Offset

Bed-Leveling Method:

1. Losen Screws of Build-Plate
2. Put paper on LCD
3. Push Button "Home"
4. Fasten Screws if Build-Plate is on the Paper
5. Press 0,1mm UP
6. Set Z=0 in the other menue
If the Paper has 0,1mm thickness then after my routine the Logical-Zero-Point is 0,2mm over LCD.


Resin: Prima Creator Value White
Layer Height: 0,05mm
Bottom Layer Count: 5
Exposure Time: 2.5s
Bottom Exposure Time: 12s
Bottom Lift Distance: 4mm
Lifting Distance: 4mm
Bottom Lift Speed: 65mm/min
LiftingSpeed: 90mm/min
Retract Speed: 120mm/min
Anti-Aliasing: OFF

The Follwing you can only see if you click on Default-Profile on GCODE and the click your Printer

GCODE Start:
G21;
G90;
M106 S0;
G28 Z0;

Interlayer:
M6054 {image};show Image
G0 Z{rise_pos} F{rise_speed};
G0 Z{fall_pos} F{fall_speed};
G4 P{light_delay};
M106 S{light_pwm};light on
G4 P{exposure_time};
M106 S0; light off

End:
M106 S0;
G1 Z{machine_height} F25;
M18;


Before i make some changes on the machine parameters i use a gcode to make a paramter-dump:

Make a txt-file with the following:

M8512 "active_config.gcode" ; Parameter Dump Gcode write to active_config.gcode

Store it yourname.gcode and print it on the printer. It read the machine-parameter and write it to the usb-stick in the file called active_config.gcode.
You can read the parameters with your text-editor.


Calibration Z-Axis Height:

GCODE
M8010 s0,000625; Set mm/Step

Here we have 0,000625mm per step and not the wrong informations all over in the forum it is 0,000625 steps per milimeter - so can anyone explain to make 0,000625 steps on a steppermotor?
Yes... it is not possible... so waste hours of life-time to check the mistake.
So if you Print a cube like 20x20x20mm and it prints to small like 20x20x19,65mm then the model is to small in Z-Height.
We calculate that 20mm need 32000 Steps by 0,000625mm/step. If we have only 19,6mm it need more steps - logical or not?
That means to become more steps you have to decrease the 0,000625mm/steps not to increase!!!
So if you use 0,000622 mm/Steps you have 32154 Steps and yes this is more and so the model prints with are higher height.

The problem is only that the Elegoo Mars 2 Pro have a lead of 2 - so the formular is
Lead = I* ((360/1.8)*16)
I means 0,000625mm/Step
Lead is 2.

Make a txt-file with the following:

M8010 S0.000622; Set mm/steps Default 0,000625mm/Step
M8500 ; SAVE MUST DO THIS IN EVER G CODE SET YOU SEND THE PRINTER

Store it yourname.gcode and print it on the printer.


So to change the M8010 value is a good idea to solve issue with z-height on models - but it is not corrrect to change a value that is physically given from the hardware.
It looks like it helps because it compensates play in the axis or retract-issues. If you measure the moevement of the z-axis only in one direction like 10mm UP and repeat it severall times like 10times also 100mm the you can see/measure it is not good to change the value from 0,000625 to 0,000622 because it move not 100mm it move only 99,5mm.
Instad if you move it 10mm up and then 10mm down ist looks nice.. because the backlash comes in front to you.

So what next?

Calibrate Logical Z=0

Often i want to print parts without supports diretly from the build-plate like cylinder or some wheels, gears.
I have often the problem the thickness is not correct - if i make another bed-leveling the thickness change..
So for this i make the same bed-leveling every time - describes above - and then i print my part - measure the thickness and make a static offset with the follwing gcode:
M8084 Z0.000000; Set Z Offset

For example my thickness is to small - i want 3,1mm and the part have only 2,8mm - the difference is 0,3mm.

Make an txt-file with the following:

M8084 Z-0.3; Set Z Offset
M8500 ; SAVE MUST DO THIS IN EVER G CODE SET YOU SEND THE PRINTER


Then store it to yourname.gcode and print it with the printer.

So this solved thickness only on the first layers!!!


And now we came to my problem...
I have a simple part like a round gear for simplest without tooths.. like a flat wheel diameter 45mm with only a thickness or z-Height of 3,1mm.
On this flat wheel i have a cylinder with diameter 16mm and z-height of 8,5mm.

And now erverbody can draw it and export it and print it - it is very simple part.
If i print it with default factory settings i have a thickness or z-height of the flatwheel of 3,08 to 3,12mm - looks not so bad but the z-height of the cylinder is not 8,5mm is only 8,07 to 8,09.

And if you make a 6mm diameter hole in the middle on the cylinder with a depth of 2,5mm it has exact 2,5mm depth.. so this is so confusing...

I have a picture of this simple part - i can also upload the stl - but the jumping point is - make your own and print it with your Elegoo Mars 2 Pro and give me your results... tell me you can print 8,5mm height cylinder on a flatwheel with 3,1mm height with a hole of 2,5mm depth or minus z-Height.

For me it looks like there are some bugs in calculations - chitubox slice the layer on 0,05mm layer-height only 0,043 and 0,083 and then 0,122mm and so on - i dont understand why this happen.

All calibrating-ideas are useless then if i start to calibrate M8010 then the hole of 2,5mm depth goes also deeper... maybe 2,6-2,7mm dependig from my changes --- thats frustating.
If i use the M8084 only the flatwheel on the bottom is affected.

So if somebody can check the problem with their Elegoo Mars 2 Pro printer - thats will be nice.

Regards,

nighteagle

---

Image of the simple parts--- in different dimensions..

[nighteagle]

Hello,


So someone have similar Z-Height accuracy problems or not?
And on Chitubox why the hell sliced it with weird values.
I i use the example cube of chitubox the model 20x20x20.stl then slice with 0,05mm Laer Height the layer are:
1. 0,000
2. 0.074
3. 0,147

that looks really weird?

Can somebody retest it on your own Chitubox? I have the 1.7.0 and Chitubox-Support write me to delete Log-Path-Files.
But this changes nothing... they told me also it is not normal to have 0,074 instead of 0,05...

Regards,

nighteagle

[nighteagle]

Hello,

i have seen some bending on the Z-Axis of the Mars2Pro even if you print objects with big surface on the bottom layers - because the fep-film bring a big force to the Z-Axis if you print large obejcts. To Prevent this bending - have measure is more than 0,4mm - i have modified the printer with two 6mm plates from aluminium and mill them in my cnc.
You have to disassembly the upper plate and be sure to protect the LCD - i have use papertape and paper to protect the LCD of both sides.
Then disassembly alle the parts on the upper plate and mill 4 holes besides the z-Axis and thats all on the plate.
An the Z-Axis itself mill 5 holes each side to mount the staibilizers and at last measure all is correct in angle and fasten all screws.

If some one need the CAD-Model of it - please mail me.

[nighteagle]

So is there anybody with a solution for Z-Height accuracy?

[DDLLC]

Nighteagle,

I don't have a Mars 2, but I've spent a lot of time trying to improve Z accuracy on a Mars. My run parameters are not directly related to yours but I can tell you that I have seen more than 1mm thickness difference on a 5 mm thick part just by varying run parameters. Light delay time is important and adding say 13 seconds may be helpful. Every single trial (maybe 50 or so on a simple open rectangle 95 x65 mm x 5 mm thickness) I've made produced thicker parts than the design by at least 0.155 mm. You are apparently seeing only thinner parts than design, so maybe the Mars 2 has different factors affecting the thickness.

Early in my testing I had weak indications that slightly increasing the lift distance from the recommended 5 mm to 5.5 mm might have made a slight improvement, but that might have just as well been caused by replacing the FEP film and having a different amount of tension. I currently run a 7mm lift but only because it was the high end of a design of experiment that I initially ran. One can definitely hear the part release from the FEP film well before it reaches the lift limit.

Whatever parameter changes you make, I'd suggest at least 3 replicates each with  a minimum 5 thickness measurements. This way you can run a statistical test to see if the parameter really might have made a change. The standard deviations have varied on a given trial from 0.068 to 0.108 mm at the low end to well over 0.200 mm at the high end for a part with 5mm design thickness.

I haven't even looked into the gcode yet, but the print precision I'm seeing suggests you may be chasing a ghost (I mean that maybe the problem is more parameter settings and lack of print precision due to machine design than the code itself).

You also seem to have an issue in Chitubox as those numbers are odd. I'm running a lot of trials with mostly 0.020 mm layer thicknesses and in Chitubox my layers are stepping that amount (in the Chitubox run window). I've switched to the latest version 1.8.1 but I don't recall seeing any odd numbers in the previous version  ( I think it was 1.7.3). It might be worthwhile testing the latest release.

For what it's worth my lengths and widths also vary with parameter settings but generally are less than 0.5%.

Hopefully somebody with a Mars 2 can try to replicate what you are seeing.

Greg

[nighteagle]

@DDLLC

yes i thik also is a problem of th stability - special in z-axis.
Here a picture for my mod to became stronger-z-axis not so a extremly bending like before.

[DDLLC]

The side supports should prevent any forward movement. Where you able to measure any differences in dimension variances?

[cliffyk]

Prior to replacing the OEM mast on my plain ol' Mars with the Jackson Products dual track mast I made a number of observations using a dial indicator with 0.0005" resolution. I saw no fore/aft sway of the mast even at it's very top while printing. I do not believe the load vectors imposed by the normal printing process are sufficient to flex that extruded aluminum component or the printer's top deck. The mounting screws at it's base must of course be properly torqued.

[nighteagle]

The side supports should prevent any forward movement. Where you able to measure any differences in dimension variances?

Hello,

yes it is more better  - but it is also a problem of backlash i think... elegoo told me all is okay they have printed and all dimension correct - but they have made a trick - the rotate by 45 degree the object and measure then the same points. But now the can't find the Z-Axis- failure because it is now printed in diagonal xy direction. Often i think the will kidding me... if i say z-Axis not correct in height - they change the environemnt and say all is okay... it is not a good working with these bad support. I have wait weeks for answers and they make all wrong and want not to see the problem.

i print large object with a big surface and the forces are very high i mean - from the fep to the build-plate - i also use silikon-grease on the fep before i fill in the resin.

But Z-Height not correct - so i print two times - first i measure the wrong Z-Height then correct it with an offset in the CAD-Model and print again. Silly Work-Around...

[nighteagle]

Her is a nice post about the z-axis problem:

https://blog.honzamrazek.cz/2019/08/test...s-problem/

I have now also a Saturn and they have exact the same problem.