New Moai 200 calibration in Asura 2.2.3

[peopoly]

hi, some of you noticed there is a new calibration profile in Asura 2.2.3 for Moai 200
If you open the file up in text editor,

You will see the data points are now in matrix

This post is to help the users understand the working and perhaps customize in the future (tools are coming)

So Moai 200 has a x/y plane of 200x200mm
In this calibration set up, we will have 21 data points (number_cols / rows) on each axis
This is means each data point is supposed to be 10mm apart from the next data point.

here is how the matrix data is defined

there are 21 rows of data, each row has 42 data points or rather 21 pair of data points.

What do I mean by row?

And what do we mean by pair of data?

so each pair is (dx, dy) where dx / dy are defined as difference in distance to center as

with 21 pairs or data per row for 21 rows, we can define how the entire x/y plane distorts

@ad_fontec @Anton23

[Whazaa]

@peopoly Is something like this coming for the little brother Moai 130?

[ad_fontec]

Thanks Mark for the information and the nice and easy to understand pictures.

I have one question above all:

Have you tested the correction on a Moai 200 and can you indicate how high the resulting absolute accuracy in the problem areas is?

This would be of particular interest to me to what extent the measurement data from my machine can be transferred to other Moai 200s.

(Currently I do not have access to the M200 for testing)

If the elimination of the distortion in this way succeeds, then I would agree to carry out the same measurement procedure on a Moai 130, which could be loaned to me.

[mauswerkz]

This is very exciting! Very much looking forward to the calibration procedure and getting my printer dialed in. If a “default” correction is transferable between printers to improve typical distortion in the meantime (assuming that’s the values in the included fiel) that would be awesome.

[Vectorfire]

Show me a printed, full grid over the entire work x-y work volume along with an .xml of the the physical x-y intersection locations post calibration, and then we’ll talk.

[johnchen]

Can anyone confirm the new Moai 200 calibration file works and is actually better then not having it?

[johnchen]

Anyone?? surly someone has printed with the new Moai 200 calibration

[peopoly]

the moai 200 on Asura 2.2.2 works fine but is basic. Asura 2.2.3 has much more refined calibration but the current profile has a bug. We have found a solution for it and will get a new update out soon

[Anton23]

Print before calibration

this is a square net test print with “plus shaped” pins at the cross of lines. Step is 10mm.
Tops of pins were painted by marker. Using “PlotDigitizer” software I extracted coordinates of these pins from photo. As a result of the processing of these coordinates I obtained the calibration file.
This method is not too accurate but it was the simplest thing that I could do.

print after calibration

On this print the shape of pins was distorted as a result of using the calibration file.

[johnchen]

if you would willing to share your calibration file that would be amazing.
Peopoly has said the current calibration file has a bug, so its not useful right now.

[Anton23]

Okay. I need to explain something first.

I started just from a very simple test print - square frame (at the top of the photo).

I wanted to test opportunities of the new calibration file using this print. For this purpose I prepared the calibration file which should shift corners of the square, like a strong distortion easy to detect.
You can see result of this at the bottom of the picture. Obviously, it was distorted but not in that way it was supposed to be. I supposed to obtain sharp corners and bent sides of the square, but I got just mess instead.

[Anton23]

After that my conclusions were - maybe this software couldn’t handle with such strong changes or maybe I did something wrong in my corrections. That’s why I decided to use this calibration just in the way it was supposed to be applied. Then I made those calibration prints which I had discribed here the day before. And again I saw that weird distortion of the shape but it was just slighter.

So my opinion is that the problem is not in the calibration file but in the algorithm of calibration which this software uses. It looks like it somehow divides the model by squares and shift them relatively to each other filling gaps by bulk material or leave them empty. This algorithm should be definitely more complicated than I’m trying to explain, but it seems that now this calibration doesn’t work at all.

I hope it’s clear what I mean.

P.S. I did all these attempts of calibration in Asura-2.23beta 20190713 two weeks ago

[johnchen]

Hey Anton, Its not just you, I tried to print the pillars with the 2.2.3 calibration and they got all crazy warped like your second photo.

As Peopoly stated, the calibration file in 2.2.3 was broken.

They have just released 2.2.4 with fix to the calibration file. If you are feeling brave to test again and share with us your findings

[Anton23]

You can just convert your stl to gcode in asura (applying some calibration file, of course) and upload it to gcode viewer.
e.g. here: http://gcode.ws/ you can check gcode file slice by slice.
This is easier than to print and you don’t need waste time and resin.

[Anton23]

i checked gcode after Asura-2.24 20190909. It looks like it should work