Extrusion tests, steps per mm. My viewpoint
 
Notifications
Clear all

Extrusion tests, steps per mm. My viewpoint  

  RSS
tom.b6
(@tom-b6)
Trusted Member
Extrusion tests, steps per mm. My viewpoint

I’ve had my MK2S since around May last year. In that time, I’ve made a variety of changes, many of which were inspired by threads here.
Among those changes have been setting the steps per mm in the firmware so that telling the extruder to extrude 100mm actually extrudes 100mm of filament.
My results from doing this were very similar to what other’s have reported, that the factory setting of 161.3 is about 7% too low. I ended up changing it to 171 which caused my 100mm command to extrude 100mm of filament.
I recently was printing some items that require a precision fit and found that they were too large and this appears to be related to the extrusion setting. I made changes to the extrusion multiplier in Slic3r and was able to improve the parts by lowering the multiplier to ~.94. Since this was approximately the same amount I had upped the multiplier in the firmware, I decided that rather than reset the multiplier for every filament, I’d set the firmware back to 161.3.
With the firmware at default, and multiplier set to 1.0, my prints are much more accurate across multiple filaments (PLA and PETG).
So, some questions and concerns.
1) Why does the firmware have the 161,3 setting embedded when it is always (by measuring the filament extruded) wrong?
2) All the gcode on the SD supplied by Prusa prints great with the 161.3 setting. If it needs something else, why do the supplied .GCODE files print so well?
3) Have others had this same experience? That changing steps/mm hurts the accuracy?

Based upon my experience, I think it’s a mistake to change the steps/mm in firmware. Everything is more accurate, at least when slicing with Slic3r, with steps/mm set to the factory defaults. If there are changes needed, they should be made for the specific filament in the slicer.
For testing extrusion, I was using an XYZ calibration cube and an object named Extrusion_Test.stl that I downloaded from somewhere. I prefer that latter as it provides and easy to evaluate fit test. By printing two, you can see how they fit together which gives you an accurate evaluation of the effect of extrusion changes. I've attached the stl.

Publié : 12/02/2018 8:10 pm
digibluh
(@digibluh)
Reputable Member
Re: Extrusion tests, steps per mm. My viewpoint

It's always wrong because:

1. Filament itself can affect the Steps per mm - material type... ABS,PLA,PETG...
2. The speed you extrude can affect steps per mm (though you would have to check how much, depends)
3. The pressure/how tight the idler presses against the filament changes esteps (by changing the effective diameter of the toothed gear). i don't know how this affects dual bond-tech types though.

See: https://github.com/alexrj/Slic3r/issues/955

My other machines ended up at 0.94-0.95 naturally (titan extruders' with calculated, not calibrated steps per mm) after dimension and consistency checks. some filaments may vary which i calibrate if i need a good accurate part but there are too many variables to make this number fixed. The defaults are about correct and best for prints over all, so i have no clue why people are calibrating and looking at dated calibration guides... they'll just end up at less than a multiplier of 1 anyway....

The MK3 firmware previously had calibrated steps per mm closer to what they would be, and produced terrible prints. then it was changed (or slic3er profile was changed... i forget). there are threads there about it some place.

Publié : 13/02/2018 1:08 am
Partager :