Inconsistant E-Step Calibration

[Roxy]

Well... Do what ever you are comfortable with and works for you... But this is kind of like setting the Z-Probe-Offset-From-Extruder for Auto Bed Leveling. You can measure and theorize. But that will only get you so close. Once you are close, you need to just manually tweak things until the right thing happens.

Good Luck! And have fun with it!

[Chadd]

Well... Do what ever you are comfortable with and works for you... But this is kind of like setting the Z-Probe-Offset-From-Extruder for Auto Bed Leveling. You can measure and theorize. But that will only get you so close. Once you are close, you need to just manually tweak things until the right thing happens.

Good Luck! And have fun with it!

Exactly, you have to get close then tune it.

[BLKKROW]

I found it best to calibrate the extruder with the hot end off. Using the following formula:

(expected length) x steps = (actual length) x new steps

[clough42]

I struggled and struggled with the formula, and you're right. It never quite comes out right. I just marked the filament at 110mm and then extruded 100mm and checked, changing the value until it was within 2mm.

Then I tuned by hand as I printed. If solid infill squeezes out the sides and builds up on top so the hot end is bouncing over the rough surface, tune it down. If solid infill on top becomes an open mesh, tune it up.

Keep in mind that filling is extremely sensitive to Z height. Don't trust anything you see in the first 5mm of a print for extrusion tuning. A very small error in Z height can make the first few layers too dense or too sparse.

[Alibert]

Did you try measuring with the nozzle taken off, that is without any backpressure on the filament? And how does that value compare to what you found manually?

[clough42]

Did you try measuring with the nozzle taken off, that is without any backpressure on the filament? And how does that value compare to what you found manually?

I did not try with the nozzle off. I was using a brand new, very sharp index-milled bolt, but I suppose there could be some compression process going on at the engagement point. Conditions were consistent from run-to-run, but who knows? There could be something going on there.