Rotating Magnetic Sensor - Keeps Pausing Printer

T3P3Tony

@wouldstain said in Rotating Magnetic Sensor - Keeps Pausing Printer:

zero issues after running with the the sensor set wide open:

do you have a report of the window that was detected by the end of the print?

dc42

So far my test files have failed to identify the problem. I can see that @alankilian has an abnormally high AGC value which I would normally consider out of the workable range, but despite that his results are good.

I printed the drywall anchor from @wouldstain yesterday. It showed an expansion of the measured range towards the end. After 13 minutes the range was 81% to 127%, but in the final 3 minutes it increased to 53% to 130%. So I will try to work out the movement pattern that caused that - or maybe it was just slower printing time towards the end.

A Former User

@dc42 said in Rotating Magnetic Sensor - Keeps Pausing Printer:

So far my test files have failed to identify the problem. I can see that @alankilian has an abnormally high AGC value which I would normally consider out of the workable range, but despite that his results are good.

I printed the drywall anchor from @wouldstain yesterday. It showed an expansion of the measured range towards the end. After 13 minutes the range was 81% to 127%, but in the final 3 minutes it increased to 53% to 130%. So I will try to work out the movement pattern that caused that - or maybe it was just slower printing time towards the end.

If it could help I've noticed too that in my prints (when they don't stop immediately) there's ALMOST always a range expansion during the print in terms of max and min limit....at a certain point I thought that the two values could be "incremental" values....

dc42

@thedragonlord @wouldstain @alankilian @charlesmc @CNCModeller I have fixed a couple of potential issues in RRF that might have caused spurious under or over extrusion reports. My initial testing suggests that these have fixed at least part of the problem - the window at the end of the drywall-anchor print went from about 53-130% before to 93-112% after. However I have not done enough runs (2 before the changes, 1 after) to prove statistically that the changes have made an improvement.

If you wish to try out RRF with these changes, you can find the binaries at https://www.dropbox.com/sh/xeu30go894p9xbs/AAB5U66SH4wbjKIcX0snUbKxa?dl=0. Caution! These are pre-beta binaries so they have had limited testing. If you are not already running firmware 3.4beta5 then you must read the upgrade notes before installing them.

A Former User

@dc42 said in Rotating Magnetic Sensor - Keeps Pausing Printer:

@thedragonlord @wouldstain @alankilian @charlesmc @CNCModeller I have fixed a couple of potential issues in RRF that might have caused spurious under or over extrusion reports. My initial testing suggests that these have fixed at least part of the problem - the window at the end of the drywall-anchor print went from about 53-130% before to 93-112% after. However I have not done enough runs (2 before the changes, 1 after) to prove statistically that the changes have made an improvement.

If you wish to try out RRF with these changes, you can find the binaries at https://www.dropbox.com/sh/xeu30go894p9xbs/AAB5U66SH4wbjKIcX0snUbKxa?dl=0. Caution! These are pre-beta binaries so they have had limited testing. If you are not already running firmware 3.4beta5 then you must read the upgrade notes before installing them.

great! tomorrow I'll give it a try! of course I'll let you know

CNCModeller

@dc42 thanks for the update. I'll do some testing with your gcode at the weekend and let you know the results with my 600ppr rotary encoder based sensor.

wouldstain

@t3p3tony I cant say for sure that it always happens when running slower. Only that it did happen this time, with multiple occurrences. I just happened to be printing something that came to a very fine point that I felt would turn out better if printed slower.

gcode attached just in case it would be helpful...

Cone.stl

wouldstain

If by PA you're referring to Pressure Advance, I looked at that as a likely culprit as I was getting much more frequent stops before I had the sensor in place and seemed to be having a problem with the extruder outrunning the hotend. It did not seem to make much difference.

Should it be revisited at slower speeds?

wouldstain

@dc42

Tests 2-4 resulted in an error message.

Console output:

1635171097795-mfm-test-1:
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 131 agc 75, measured sensitivity 25.89mm/rev, min 99% max 103% over 48.1mm

1635171097802-mfm-test-2:
Error: in GCode file line 219 column 14: meta command: variable 'extruder' already exists
Cancelled printing file 0:/gcodes/1635171097802-mfm-test-2.gcode, print time was 0h 0m

1635171097805-mfm-test-3:
Error: in GCode file line 219 column 14: meta command: variable 'extruder' already exists
Cancelled printing file 0:/gcodes/1635171097805-mfm-test-3.gcode, print time was 0h 0m

1635173702760-mfm-test-4:
Error: in GCode file line 415 column 14: meta command: variable 'extruder' already exists
Cancelled printing file 0:/gcodes/1635173702760-mfm-test-4.gcode, print time was 0h 0m

T3P3Tony

@wouldstain

At the end of the test file is the following:

var extruder=tools[state.currentTool].extruders[0]
M400
M591 D{var.extruder}

Comment out the line starting with var and change the D variable to the appropriate one for your machine (normally 0)

;var extruder=tools[state.currentTool].extruders[0]
M400
M591 D0

wouldstain

@t3p3tony
Console output:

1635171097795-mfm-test-1:
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 131 agc 75, measured sensitivity 25.89mm/rev, min 99% max 103% over 48.1mm

1635171097802-mfm-test-2:
Duet3D rotating magnet filament monitor v3 on pin e0stop, enabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 132 agc 77, measured sensitivity 26.00mm/rev, min 95% max 102% over 99.6mm

1635171097805-mfm-test-3:
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 132 agc 75, measured sensitivity 25.87mm/rev, min 98% max 102% over 49.0mm

1635173702760-mfm-test-4:
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 132 agc 75, measured sensitivity 26.09mm/rev, min 97% max 102% over 39.1mm

1635178672431-mfm-test-5
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 131 agc 75, measured sensitivity 25.88mm/rev, min 98% max 102% over 99.3mm

1635178672433-mfm-test-6
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 131 agc 75, measured sensitivity 25.99mm/rev, min 97% max 104% over 99.4mm

1635178672436-mfm-test-7
Duet3D rotating magnet filament monitor v3 on pin e0stop, disabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 131 agc 75, measured sensitivity 25.79mm/rev, min 98% max 101% over 45.4mm

A Former User

@dc42 said in Rotating Magnetic Sensor - Keeps Pausing Printer:

@thedragonlord @wouldstain @alankilian @charlesmc @CNCModeller I have fixed a couple of potential issues in RRF that might have caused spurious under or over extrusion reports. My initial testing suggests that these have fixed at least part of the problem - the window at the end of the drywall-anchor print went from about 53-130% before to 93-112% after. However I have not done enough runs (2 before the changes, 1 after) to prove statistically that the changes have made an improvement.

If you wish to try out RRF with these changes, you can find the binaries at https://www.dropbox.com/sh/xeu30go894p9xbs/AAB5U66SH4wbjKIcX0snUbKxa?dl=0. Caution! These are pre-beta binaries so they have had limited testing. If you are not already running firmware 3.4beta5 then you must read the upgrade notes before installing them.

No way...Installed the pre beta but got this after few minutes of printing:

Duet3D rotating magnet filament monitor v3 on pin e0stop, enabled, sensitivity 24.78mm/rev, allow 60% to 130%, check every 3.0mm, version 3, mag 131 agc 66, measured sensitivity 24.49mm/rev, min 41% max 118% over 771.9mm

I've kept an eye to the range for about two minutes and the minimum and maximun have been about stable at around 94% and 104%. After two minutes I went away from the printer so I can't say if before the stop the range has started to enlarge or if the stop itself has been a single event

What I can say for sure is that once more the stop occured in the surrounding of a sharp corner where the nozzle slows down...(I can say because I've found a small blob of filament where the nozzle has stopped)

dc42

@thedragonlord is this with the file Very Fast Test Print__PETG Amazon Basic_0.2000mm.gcode that you posted earlier?

A Former User

@dc42 said in Rotating Magnetic Sensor - Keeps Pausing Printer:

@thedragonlord is this with the file Very Fast Test Print__PETG Amazon Basic_0.2000mm.gcode that you posted earlier?

no, with this one
GTS3D PLA Calibration Cube T0__PLA GTS3D_0.2000mm.gcode

wouldstain

@T3P3Tony @dc42

Just received this from the attached print:

Duet3D rotating magnet filament monitor v3 on pin e0stop, enabled, sensitivity 26.50mm/rev, allow 1% to 200%, check every 3.0mm, version 3, mag 129 agc 75, measured sensitivity 27.29mm/rev, min -1% max 109% over 406.5mm

I'm guessing the negative percentage indicates a retraction?

Cone_threaded.stl

T3P3Tony

@wouldstain thanks. I have found a section of a short print file that gives repeatable bad results.

It would be helpful to try the cone print but better if you attached the sliced gcode than the stl file and my slicing settings will be different.

wouldstain

@t3p3tony
Oops! I had passed it directly from Prusa to the printer w/o saving as a file...

here you go!

Cone_threaded.gcode

wouldstain

@t3p3tony

It occurs to me that this could be part of a broader issue I had encountered prior to installing the MFM.

In short, I was experiencing what appeared to be over-advancement on the extruder. This would often result in filament being ground in the extruder to the point where it would no longer feed leading to an eventual clog in the hotend and even in the throat.

I worked with @3lemental to tune the M92 E-value (went from 837 to 814) which yielded some improvement but did not completely resolve the issue.

Installing the MFM lead to a drastic improvement (no more clogs or ground filament!) but these pauses may be related. Since this was no longer ruining print jobs, I sort of placed the issue on the back burner.

I don't know if it's appropriate to this thread but I have more detail on what steps I had / have taken so far on this.

Just let me know if you would like me to post it here or elsewhere.

T3P3Tony

@wouldstain Probably better to start a new thread on that, thanks.

dc42

I have put new 3.4beta5+1 binaries at https://www.dropbox.com/sh/xeu30go894p9xbs/AAB5U66SH4wbjKIcX0snUbKxa?dl=0. These fix a bug that was giving me consistently poor results from the filament monitor on the drywall-anchor print. This particular bug was either introduced or made worse in version 3.4beta2. The particular instance that triggered it was a very short extruder move, less than one microstep long. This means that it would have occurred more often on machines with low extruder steps/mm.

Use these binaries with caution, because I have done only small test prints with the Duet 2 binary, basic movement tests with the Duet 3 binary, and I have not yet tested the Duet 3 Mini or Maestro binaries.

These binaries also include the race condition fix that may have caused occasional incorrect results from the filament monitor in firmware 3.3 and earlier versions.