How to calibrate geometry of printed parts?  

Re: How to calibrate geometry of printed parts?

Possibly is the question somewhat inaccurate.

I have several cubes with an edge length of 20mm distributed on the printing surface and printed. After printing I have measured the parts and averaged the results.
The size of the cubes also differs among each other somewhat, but this can be neglected.

The average value of the width (x axis) is 19.39mm. The average value of the length (y axis) is 19.21mm.
Since the distribution on the surface apparently does not matter, I assume that the number of motor steps per distance is too small.

Can I set it and where?

Re: How to calibrate geometry of printed parts?

Jens

Yes, you can change the steps per mm using GCode M92: http://reprap.org/wiki/G-code#M92:_Set_axis_steps_per_unit

That Wiki can come in very useful...

However, I wonder if your issue is actually the printer; there are other things which can affect dimensions:

Some slicers will also include an option like "Inset Surface" which you can use to adjust final dimensions.

In your case, I would suspect belt tightness first, especially because your X and Y measures are different.

Peter

Re: How to calibrate geometry of printed parts?

I would suspect belt tightness first, especially because your X and Y measures are different.
You're most likely right, I'd just like to point out that my printer is a bit imprecise too. But not that much. There is about 0,5mm deviation on 200mm object, it's about 0,25% off in X and 0,35% in Y. I don't care. Only pointing out that X and Y deviations may differ.
But Jens got almost 10 times more, thats too much.

Re: How to calibrate geometry of printed parts?

You have to be careful. Is the distance error caused by the under / over extrusion or by the X/Y step length deviation?
I believe the X/Y step length deviation could be caused by too tight belts.
Vojtech

Re: How to calibrate geometry of printed parts?

Hi

To ensure if it's steps per distance issue, you should try a bigger test print. If the deviation is proportionnal, then it could be the issue, but if the deviation remain the same on absolute value, it's probably an extrusion/infill or slicing issue.

Re: How to calibrate geometry of printed parts?

I'm a little busy at my job at the moment, sorry.
Only a few quick things ...
I have in the surfaces very little vibration and no layer shifting. The size deviation occurs systematically throughout the whole print bed why I am not entirely clear how the belt may cause this.
Nevertheless, I have to be sure the tension slightly varies. This had no effect.

I have checked the gcode files. I see no problems and the positioning is considering the extrusion width exactly.

I have the absolute position and the relative position of each cube with each other checked and could find that this is correct for each of the reference point at bottom left.
Therefore a deviation of the motor-steps is no longer plausible and I do not actually know how to go forward.

Voitech, thx, I know. I will only change such settings if I absolutely sure what I do.

Re: How to calibrate geometry of printed parts?

I have puzzled over what can cause the behavior.

Since the change of size looked for some kind of transformation, I have tried to minimize the correction which is determined in the xy calibration.
Therefore I have disassembled the printer partially and reassembled.

After the initial assembly, the calibration result was "quite skew" was (I do not remember exactly the wording).
Now the result is "a bit skew" and the geometry is much better. The differences are now in a range below 1% which is sufficient for now.

I suppose I was with the tendency of the x axis towards the y axis in a limit region, so it was adjustable by the software but might have caused the deviations.

I will sometimes try to get it perpendicular. So far I could not done it.