The marlin firmware comes with the following steps/mm configuration:
- X: 80 steps/mm
- Y: 80 steps/mm
- Z: 400 steps/mm
- E: 400 steps/mm
But these numbers be a bit different per each particular copy of the printer. Hence, I decided to spend time and calibrate all the steps/mm values to achieve better dimensional accuracy.
The approach is pretty simple - we need to compare actual measurements with what was expected and calculate new steps/mm value using the following formula:
new steps = (current steps * expected dimention value) / actual dimention value
First of all, we need to calibrate the extruder to be sure the amount of the pushed filament is accurate.
There are a lot of videos over the Youtube describing the calibration process. I'll just present numbers and calculations.
Initial steps/mm = 400
Control distance = 100 mm
Expected distance = 90 mm
Actual offset = 13.5mm
Actual distance = 86.5mm
new steps = (400 * 90) / 86.5 = 416.19 steps/mm
I set the given value and decided to make another round.
Initial steps/mm = 416.19
Control distance = 100 mm
Expected distance = 90 mm
Actual offset = 10.78mm
Actual distance = 89.22mm
new steps = (416.19 * 90) / 89.22 = 419.83 steps/mm
I set the new steps/mm value for the extruder motor and decided it should be good enough to just stop calibrating and simply go with that value.
For the axes calibration we need to print some part and use measured dimentions as actual values. I used tips from this video to print the 100x100x100 calibration print instead of 20mm cube. That's what I've got:
I was a little surprised once I measured all the dimentions.
Initial steps = 80 steps/mm
Expected size = 100mm
Actual size = 100.19mm
new steps = (80 * 100) / 100.19 = 79.85 steps/mm
Initial steps = 80 steps/mm
Expected size = 100mm
Actual size = 100mm
new steps = (80 * 100) / 100 = 80 steps/mm
Initial steps = 400 steps/mm
Expected size = 100mm
Actual size = 100.03mm
new steps = (400 * 100) / 100.03 = 399.88 steps/mm
It turned out all the dimensions are pretty well aligned with the expected values, so it might be even not worth making some adjustments. Still, anyway, I adjusted only X-axis steps/mm - other both values are good enough to leave them as is.