Ultimaker 2+ Estep calibration with data, variance ok?

[X-Type 0]

Hi

I've spent the evening working with estep calibration on my UM2+
Below are the results. I hope you understand the data 🙂
The time the extruder ran was as perfect as I could measure each time. (And it seems my response time is always very close to 0.15 seconds 😉 )

First of all a notice on what I assume is some kind of bug. The first extrusion run after every M92 change of the estep value, gives a way off result. 
So, I have tested three extrusion lengths 100mm, 200mm and 300mm. It seems the longer extrusion run, the less filament is extruded. So, at the third and last setting in the data below, at E100, I tend to get 101-102mm, at E200, it's more or less spot on, and at E300, it is 2-4mm too short. So, where talking about +- 2%. Is that a normal variation? Though the variation seems to be related to the extruded length, so it is not random.

Starting off with default estep = 369

E369
E200	F500	216mm
E200	F1000	206mm
E100	F100	203mm
E100	F100	102mm
E200	F1000	204mm	12s
E200	F1000	204mm	12s
E200	F500	204mm	24s
E100	F500	103mm	12s
E100	F100	103mm	60s
E200	F750	205mm
E200	F750	205mm	16s
E200	F750	205mm	16s
M92 360
E200	F750	137mm	11s  !!!!!
E200	F1000	200mm	12s
E200	F1000	201mm	12s
E200	F1000	200mm	12s
E100	F100	101mm	60s
E100	F100	102mm	60s
E100	F500	102mm	12s
E100	F1000	101mm	6s
E100	F1000	103mm	6s
E100	F1000	102mm	6s
E200	F1000	201mm	12s
M92 355.76
E200	F1000	150mm	9s  !!!!!
E200	F1000	200mm	12s
E200	F1000	200mm	12s
E100	F100	101mm	60s
E100	F1000	101mm	6s
E100	F1000	101mm	6s
E100	F1000	102mm	6s
E200	F1000	200mm	12s
E100	F1000	102mm	6S
E100	F500	102mm	12s
E100	F100	102mm	60s
E300	F1000	296mm	18s
E300	F1000	298mm	18s

[tinkergnome 764]

9 hours ago, X-Type said:

First of all a notice on what I assume is some kind of bug. The first extrusion run after every M92 change of the estep value, gives a way off result.

 

For older versions of Marlin (like the one that is used in UM2) the current axis position is lost after M92 (or better: is not longer accurate).

Don't use M92 on its own, always send a subsequent G92 Exxx.xx to reset the position.

 

9 hours ago, X-Type said:

So, I have tested three extrusion lengths 100mm, 200mm and 300mm. It seems the longer extrusion run, the less filament is extruded.

 

I don't know the specific reason why you took all these measurements, but i think the used speeds are way to fast and the filament probably just slips more and more the longer the move takes.

I would use an equal (and lower) speed for all tests. Let's say F180 or even slower (somewhat closer to 'normal' extrusion rates). The variation should be much smaller then.

 

[X-Type 0]

Thanks for input, I will give it a try and run some different extrusion lengths using a slow speed, and see what results that gets me 🙂