Page MenuHomeAleph Objects Inc

Fillament run out sensor not detecting
Closed, ResolvedPublic

Description

First time I ran a print on a alpha, I noticed that the display was reading a filament run out but cura was still sending it commands and it was still printing. The print was stopped and resent, this time no run out filament was detected. I noticed this is big because the connecter with the green wire is hooked up to the filament going to the E2 extruder.

Also I noticed that the console keeps displaying this message.

[14:16:39] echo:Unknown command: "M73 P56"

Event Timeline

oliver created this task.Dec 7 2018, 2:18 PM
oliver created this object with edit policy "LulzBot Hardware Products (Project)".
west added a subscriber: west.Dec 10 2018, 7:16 AM

the connector with the green wire needs to be on E1

logan added a subscriber: logan.Dec 11 2018, 10:14 AM

@oliver The M73 you mentioned is to update the print progress percentage. http://marlinfw.org/docs/gcode/M073.html
Not sure why it is echoing unknown though

It looks like the unknown command is a problem in the Quiver FW. I will fix this in .38

west added a comment.Dec 11 2018, 11:42 AM

Is this still an issue or can we close it?

@west Maybe the real issue he was reporting (because he had things connected wrong) was:

the display was reading a filament run out but cura was still sending it commands and it was still printing

Is there a ticket open for that?

west added a subscriber: samantha.Dec 11 2018, 11:52 AM

I think there was one made @samantha does this sound familiar?

west added a comment.Dec 11 2018, 1:42 PM

@logan i think T4446 is the ticket I was thinking about

@west Sounds about right but the main issue I saw was it did not matter what extruder each filament sensor was connected too, it would report a fillament run out message the first time it was ran and they it wouldn't after the print was restarted. It did it on both extruders

west added a comment.Dec 11 2018, 2:06 PM

Because E1 has to go to E1 and E2 has to ho to E2 if they are switched it doesnt matter if filament is running, marlin wont detect that filament is running through the correct tube

@oliver I was told the run out detection needs an active print timer to function properly. I was not able to detect a run out on the CB test stand until I put an M75 in. Abort/restart of prints is functionality known to have problems currently, so it may not have reset the print timer upon restarting the print which could be responsible for:

it would report a fillament run out message the first time it was ran and they it wouldn't after the print was restarted

But @west is correct that the reason the filament run out was triggered was because of the improper connections.

@west I think this ticket can be closed as invalid as the actual issues are being tracked through other tickets
See T4883 for OHAI concerns regarding improper connection of the filament sensor
T4446 for filament run out pause/resume issues

@west @logan As I said, the result was the same after the tubes were switched.

west added a comment.Dec 12 2018, 7:50 AM

what is your runout detection distance set to?

Perhaps I'm not understanding something about the filament sensor or I'm not being clear enough on my problem.

If a tube is connected to the wrong extruder it will detect a filament run out because as soon as it detects the sensor is not moving while its printing because its connected to the non-active extruder it will detect a false filament run out. My problem was that no matter what extruder a tube was connected to it would detect a run out on the first print job and then continue with the print. This happened after switching the tubes, if it was detecting a filament run out than at one point the print jobs should have always failed unless my understanding is not correct.

Something that dani told me, if we increase the run out length from 14mm to 25mm that could help.

Last time I tried to replicate a filament run out related to T3694 I was able to see motion detected in the console, but when it stops detecting motion it did not start counting down.

I was able to see motion detected in the console, but when it stops detecting motion it did not start counting down.

How long did you wait? The countdown only happens at the end of a segment, that is, at the end of a linear motion. If your head is moving across the bed extruding a long line, it will not count down until the end of that move.

logan added a comment.Dec 13 2018, 7:17 AM

I had generated a circle gcode in CuraLE and increased all of the extrusion moves (in an attempt to trigger the filament run out sooner) so that each extrusion move was about 1mm. I then inserted that section of gcode to the rest of my testing gcode and ran the print. Arriving at that section of code, filament run out data started coming through the console. It was reporting motion detected so I stopped the encoder wheel on the test unit, it then stopped reporting motion detected, but did not count down. I held this state for at least 20 of these moves. Run out distance at the time was set to 14mm.

west added a comment.Dec 13 2018, 7:19 AM

I'm not sure if im understanding correctly but if the sensor detects runout, it shoudnt countdown until resume is hit

logan added a comment.Dec 13 2018, 7:33 AM

It didn't trigger a run out, only stopped reporting motion detected in the console.

west added a comment.Dec 13 2018, 7:54 AM

Oh gotcha

Send me the GCODE and I can test.

logan added a comment.Dec 13 2018, 8:05 AM

Here is the gcode with the over extruding circle

@marcio were you able to test the above gcode?

@logan: Could you try with .54 or later? I think what you were seeing is that in earlier FW versions, once a runout was detected, the runout condition once triggered, would inhibit any further motion from being detected until the condition was cleared. Since there was no way to clear that condition from the serial console, the runout sensor would work only once. In .54 I changed it so the run-out condition is immediately reset after a message is printed on the serial console.

logan added a comment.Dec 26 2018, 9:11 AM

@marcio Yes, I can try again once we make a determination on a path forward for T3694 but I'd like to not spend time on that function of the test stand if we can avoid using it. In summary I think it adds too much time and complexity to the test.

tutley closed this task as Resolved.Jan 16 2019, 1:24 PM
tutley claimed this task.

With the latest firmwares, to my knowledge there hasnt been any issue reading a working sensor board. If anything false positives are more of an issue at this point in time, so i am closing this ticket. Also for the test stand, the pin states addition to the firmware ( in fw .73) will be a useful tool for determining the status of the filament sensor.

logan added a comment.Jan 16 2019, 2:00 PM

Confirmed, things are all good here! Thank you!