Filament monitors setup problem

claustro

I finally mounted the filament monitor.
Led green and red blink as expected

In config I wrote
M591 D0 P3 S23.5 C3 R50 E3.0
M591 D0

I started the calibration process with
M591 D0 R-1

Started the cube print and during the print I sent randomly
M591 D0 command.
I received 2 answer form duet . Sometimes simply it doesn't answer , sometimes it send

[[language]]
Duet3D filament sensor on endstop 3, no microswitch, 23.5mm per rev, check every 3.0mm, tolerance -1.0%, current angle 221.1
Duet3D filament sensor on endstop 3, no microswitch, 23.5mm per rev, check every 3.0mm, tolerance -1.0%, current angle 36.2
Duet3D filament sensor on endstop 3, no microswitch, 23.5mm per rev, check every 3.0mm, tolerance -1.0%, current angle 309.4

If I send M122 I receive

[[language]]
Extruder 0 sensor: angle 180.7, ok, 28.15mm/rev, tolerance +35% -146%
Extruder 0 sensor: angle 340.7, ok, 28.12mm/rev, tolerance +35% -146%
Extruder 0 sensor: angle 111.1, ok, 28.09mm/rev, tolerance +35% -146%
Extruder 0 sensor: angle 68.9, ok, 28.05mm/rev, tolerance +34% -146%
Extruder 0 sensor: angle 254.2, ok, 27.97mm/rev, tolerance +34% -146%
Extruder 0 sensor: angle 254.2, ok, 27.97mm/rev, tolerance +34% -146%
Extruder 0 sensor: angle 174.0, ok, 27.50mm/rev, tolerance +32% -145%

```My questions are:
1) why sometimes I don't receive answer following M591 command even if the led blink regularly?
2) is the calibration process achieved simply following a print after a M591 D0 R-1 command?
3) How can I check if the calibration is done correctly?
4) What is the meaning of the mm/rev I can read after M122 command and why the value decrees constantly during the print?
.5) For test I stopped mannuallt the extruder and the printer stopped as expected but on resume it extruded a lot of filament as soon as the noozle reached the old position on the bed then continued regularly the print. How can I get rid of this extruded bloob?

Thank you
Andrea

dc42

#1 Sorry, I don't know. I'll test it on my printer.

#2 At present you need to take the values reported by M122 and put the mm/rev in your M591 command along with a suitable tolerance that encompasses the tolerance reported by M122. However, as you are getting a tolerance that goes as low as -146% there is something wrong with your setup. Have you constrained the sensor so that when you retract filament, the filament is forced back through the filament monitor?

#3 It's up to you to put suitable values in the M591 command. But first you need to get the tolerance reported by M122 to within something like +/-30%.

#4 It is the mm per revolution of the filament monitor magnet calculated from the extrusion commanded and the rotation measured. If it's gradually getting lower, that could mean you are getting increased back pressure from the nozzle as the print progresses.

#5 That sounds like a relative/absolute extrusion mixup. What do you have in your pause.g and resume.g files?

claustro

I thought that the filament have to run trought the sensor in both ways , extrusions and retractions. Am I wrong .
this is a photo of my last mount.

5)```
; Pause macro file
M83 ; relative extruder moves
G1 E-4 F2500 ; retract 4mm
G91 ; relative moves
G1 Z5 F5000 ; raise nozzle 2mm
G90 ; absolute moves
G1 X0 Y200 F5000 ; move head out of the way of the print

; Resume macro file
G1 R1 Z2 F5000	; move to 2mm above resume point
G1 R1			; lower nozzle to resume point
M83				; relative extruder moves
G1 E4 F2500		; undo the retraction

```

when I send the M591 D0 R-1 I don't receive an answer is it normal?

also after the end of the print if I sen M122 I read always Extruder 0 sensor: angle 358.9, ok, no calibration data.
How the calibration data is saved?

thank you

dc42

Your mechanical setup looks reasonable to me, although it would be better if the PTFE tube from the inlet of the Titan ran all the way up to the filament monitor, to make sure that when you retract filament it can't buckle but has to reverse through the filament monitor.

Your pause and resume files look OK to me too. Which firmware version are you using? I'll re-test filament error pause/resume in absolute extrusion mode.

It's normal that M591 D0 R-1 won't receive a response, except for "ok" if it is sent from USB. M591 D0 without and further parameters should report the current filament monitor configuration for extruder drive 0.

In firmware 1.19, M122 should report the calibration data if you had the tolerance (R parameter) set to -1 at the start of the print. In the latest 1.20 beta firmware, calibration data is accumulated even if is wasn't.

The calibration data isn't saved. It's up to you to use the values reported in the M122 command to set the S and R parameters in the M591 command in config.g.

claustro

Thank you for your kind responses.
My setup can accomodate 2 ptfe tubing on both side , so it is not a problem add another ptfe tube , I didn' t added it before because I was worried regarding the filament management. I 'll give a try.
Could be this the problem of my extreme tolerance range?Any idea how can I improve it?

My firmware is 1.19.2 (2017-09-01) i am gonna upgrade to 1.20 asap.

Now I am starting to undestand the calibration process. Sending M122 I always see no calibration data available unless I start a print with R-1, is it right?

dc42

@claustro:

Now I am starting to undestand the calibration process. Sending M122 I always see no calibration data available unless I start a print with R-1, is it right?

With firmware 1.19, that's correct (or you could pause the print and then send M591 D0 R-1). In firmware 1.20 the data is accumulated even if you didn't use R-1.

claustro

Thank you
Still testing.
I have some false alarm from the sensor.

My intervention for now was to raise a little the print head form the bed but the problem remain.
How can I investigate and solve this problem?

Strangely the resume now work flawlessly , no more wrong extrusion on resume
Thank you
A.

dc42

You won't be able to use the sensor to monitor filament during a print nuti you have resolved the large tolerance that is reported in calibration mode. You are getting about +34% -146%. You should be getting results within about +20% -20%. Possible causes of the problem are:

1. Filament is not being forced back through the filament monitor during retractions, but instead it is buckling in the space between the filament monitor and the extruder instead.

2. Your extruder acceleration, maximum speed, or jerk is set too high, or your E motor current set too low, so that steps are being skipped (probably during retraction or un-retraction).

claustro

Ok at first glance none of this problems will suit to my situation . I'll test more and I 'll report.
thx

dc42

I suggest you try different sorts of movement, e.g. steady extrusion, and repeated extrusion and retraction, to see what type of movements are causing the problem.

claustro

Extruder 0 sensor: angle 238.4, ok, 24.27mm/rev, tolerance +12% -9%

this is what I obtained after an evening of tests.
At the end I completely reassembled the filaments and added a second ptfe tube.

It is quite difficult for me to understand how much hobbed gear and the front ball bearing should be tight .
The R variable in the config will be quite safe with 20 or will be better some more?

Thank you

dc42

That's a good result. Try R20.

claustro

I have inconsitent behavior from the sensor.
I start the print with 4% -4% and it goes smoothly , after a while I reprint the same things and the filament sensor give me false alarm and value og 20% -100%.

I am thinking about a problem with my spring loaded extruder

Also I have sudden stop of print without tehe usual popup telling me the filament is stopped .

01:52:18Printing paused
01:52:14Resume-after-power-fail state saved
01:51:51M24
Printing resumed

dc42

There is a known issue in the firmware, which is that you only get a popup message the first time. This is on my list to be fixed before the 1.20 release.

Your -100% suggests that the filament stopped moving due to a nozzle jam or an obstruction below the nozzle, or that the extruder drive has lost its grip on the filament.

claustro

Thank you for your explanation .
During on of my test the filament jammed at the nozzle but the filament monitor didn't stopped the print. It happened during a phase of the print with very small parts and lot of retractions, I could see the filament moving up and down through the filament monitor even if there wasn't extrusion, maybe this is was the cause of no alarm

If I understand correctly the positive or negative value of reported tolerance in M122 can give an hint of where is the jamming problem, can you explain me a little more?

dc42

The figure of exactly -100% as the lower end of the tolerance suggests that over one of the sample periods (which by default are every 3mm of movement), extrusion was commanded but no movement at all was observed by the filament monitor. I guess that an intermittent connection between the filament monitor and the Duet might have the same effect.

claustro

thank you for your clarifications, I made new connection of all sensor cable