This possible bug is really weird

[Tinchus]

Hello. I have a puylse encoder used as filament extrusion sensor. The wheel of the encoder ahs 100 holes, is really precise and accurate (the error in measurement is like 1/1000 of mms)
It works really well.
At the beginning I posted some problems related to this measurement: the encoder seemed to reported a calibration measured valuerd different insome ocassions. It was very stabble saying a value of 0.2 mm / pulse. But sometimes reported 0.25 mms/pulse. And sometimes reported 0.1 mm/pulse.
My guess was: something is wrong with electronics, may be the oncoder is too muy precise for this aplication? Well, at some point was like it was always reporting 0.2 so... it works and works really well, let it be.

But now I had a couple of prints stopped because of the enocedor triggering filament error, all false triggers. The value reported by the enoceder calibration data was again0.1, and even on a print was 0.07 mms/pulse

This time I think I found the reason: a student by mistake changed the value of pressure advanced normally used here, wich is 0.05, and this student used a 0.1 value in a profile, and then changed to something even bigger: 0.13

I called my atention this time that this 0.1 and 0.13 vlues were too much related to the decreasing of the calibration data reported.

So I did a test: using a value of CERO, I got a calibration data of 0.26 mm/pulse.
Using a pressure advance value of normals 0.05 mms, I was getting again the value of 0.22 mm/pulse
using a pressure advance value of 0.1 gave me a predicted 0.1 mm/pulse

SO I guess it is conclusive? Somehow pressure advance values are affecting thje calibration data for pulse encoders?

This is a duet3, MBC mode, with 3.5.2 version and encoder config is:

M591 D0 P7 C"io4.in" S1 L0.2 R20:350 E3

[dc42]

@Tinchus the problem with most pulse encoder filament monitors is that they are unable to distinguish between forward and reverse movement. When large values of PA are used, extrusion sometimes actually reverses during deceleration at the end of a move. So it's not unusual for the measured extrusion exceed the actual amount of extrusion, because this reverse motion will be seen as forward motion.

Duet3D filament monitors do distinguish between forward and reverse motion, so they do not suffer from this issue. For other types of pulse filament monitors it may be best to calibrate with PA set to zero, then set a very large upper limit on the allowed tolerance to avoid getting "too much motion" errors reported.

[Tinchus]

@dc42 Makes a lot of sense. But are you sure this could be the reason? I mean: the retractions created by the slicers dont create this effect on the calibration. And also: the calibration data remains super stable along the hole printing, if this negative movements created by pressure advance has this effect, shouldnt the value vary along the print or between different prints? because for example: i ran a 3 hs print (a cube) using the "continues perimeter "option in superslicer. This created a continues path with barely any stop in the same layer. And the calibration data, having a pressure advance value of 0.1, had the described efffect. But another print, of 28 hs with lots of stps, paths on the same layer and where I can clearly see the pressure advance making movements on negative direction, in that print the calibration data is exactly the same as in the cube and following your answer, shouldnt be different?

[dc42]

@Tinchus by default RRF disables filament monitoring during retraction and reprimes.

[Tinchus]

@dc42 OH! well that is good to know. Probably would be good to include this clarification into docs. Post solved, thanks a lot for the answers