Page MenuHomeAleph Objects Inc

False positive reading from filament sensor
Open, NormalPublic

Description

When I started printing with flexy materials I started getting filament runouts. I had previously been printing with nGen and PLA with no problem. With the terminal open I got this error

[12:11:53] Remaining mm: 1.52, 12.09
< [12:11:54] Remaining mm: 0.75, 12.09
< [12:11:55] Remaining mm: 0.36, 12.09
< [12:11:56] Remaining mm: 0.11, 12.09
< [12:11:57] Remaining mm: 0.01, 12.09
< [12:11:57] Pausing print
< [12:11:57] //action:filament_runout
< [12:11:59] Remaining mm: 13.81, 14.00
< [12:11:59] Remaining mm: 13.36, 14.00
< [12:12:01] Remaining mm: 13.12, 14.00
< [12:12:01] Remaining mm: 12.77, 14.00
< [12:12:03] Remaining mm: 12.40, 14.00
< [12:12:03] Remaining mm: 12.00, 14.00
< [12:12:05] Remaining mm: 12.00, 14.00
< [12:12:05] Remaining mm: 12.00, 14.00
< [12:12:07] Remaining mm: 12.00, 14.00
< [12:12:09] Remaining mm: 12.00, 14.00
< [12:12:09] Remaining mm: 12.00, 14.00
< [12:12:11] Remaining mm: 12.00, 14.00
< [12:12:11] Remaining mm: 12.00, 14.00
< [12:12:13] Remaining mm: 12.00, 14.00
< [12:12:13] Remaining mm: 12.00, 14.00

It looks like it ran out then started reading the filament again.

Event Timeline

MikeR created this task.Jan 3 2019, 2:30 PM
MikeR triaged this task as Normal priority.

@MikeR are you seeing this issue still?

MikeR added a comment.Jan 11 2019, 8:39 AM

@DaniAO I have not seen this issue since I switched back to using nGen. I was having too much trouble with the flexy that I had to change. I don't know if there is anyone else who has tried printing with flexy material other than josh to see if they have the same problem. It looked like the flexy filament was getting jammed somehow and would stop extruding after a few layers.

I had difficulties printing flexy as well. Gave up on it.

MikeR added a comment.Jan 11 2019, 8:57 AM

HEre is the gcode I was using. Nozzle one was ninjaflex and nozzle 2 was polyflex.

@MikeR what was the runout detection setting used in this print? Can you try turning the runout detection higher and test again? I wonder if we shouldn't make this setting defined in profiles as flexible materials may require a higher threshold.

MikeR added a comment.Jan 11 2019, 9:55 AM

@Steven I first had the runout set to 14mm then I increased it to 34mm. Both ended up running out.

oliver added a subscriber: oliver.Jan 11 2019, 10:07 AM

Trying to run flexy with just one extruder right now and I had this happen on the first layer for me. Attempting a second try and I can get the run out threshold when it fails or if it happens to complete.

The job I was currently printing finished the entire print for the most part. Slice settings needs some adjustment but it did finish. This was also only using extruder 1 to print and I believe @MikeR was seeing this when his started printing with extruder 2.

I am running a flexy/PVA print with the filament sensor turned on (set to 14mm). PVA on extruder 1, Flexy on extruder 2. I have had no issues with the sensors and false positives. @west is printing another dual flexy cone, and i dont believe he has seen any false positives with flexy either.

@MikeR have you tried to run flexy again lately?

MikeR added a comment.Jan 16 2019, 1:22 PM

@tutley I have not printed with any flexy recently. I may have had an issue with my tensions with the flexy. Once I get my printer put back together I will try it again.

matth added a subscriber: matth.Feb 21 2019, 10:01 AM

Firmware .91, Cura 3.6.4

Been trying to do a large print, but twice now it stopped after an hour or so with a false positive. Jeff believes it's because of the fact I'm using 2 3kg reels and since there's not too much wiggle room with them, it's easy for filament or the reels themselves to get caught on something and prevent filament intake. I make a ticket about that here: T5813

matth added a comment.Mar 6 2019, 9:10 AM

Currently on .93

I don't know if it's the sensor or the tool head, but I've been getting a lot of false positives lately. Sometimes I can reproduce them, sometimes I can't, and it'll occur mid print. Twice now I've had a long (12+ hour) print stop about an hour into printing and I have no idea why.

MikeR added a comment.Mar 6 2019, 10:40 AM

@matth is there any chance you could jabber me a vid of it? I may be able to help.

jebba added a subscriber: jebba.Mar 6 2019, 2:03 PM

Btw, I was just speculating it could be the weight of the reel, since it started happening when we put a new black 3kg Polymaker PLA reel on the machine. Could be 100x other things.

matth added a comment.EditedApr 16 2019, 10:23 AM

Printer: KT 7, beta
Filament: noz 1: polydissolve, clear
noz 2: polylite pla, black
Firmware: 2.0.0.108
Runout detection distance: 14mm
Profile: values for standard from http://devel.lulzbot.com/filament/Testing_Docs_Master/Quiver-Setting-Guide/

Starting getting false positives again on my beta. Currently am trying to print a 30 hour model with polydissolve supports, but would get the failure an hour into the print. It would then resume for a few minutes, then fail again. After 3 failures in the span of an hour I scrapped the print.

Every time it's from filament 2, a 3kg reel of polylite PLA. It's positioned on the spool arm as optimally as I can manage. I even unspooled a bit of filament to make it looser/easier to feed, but didn't seem to help. I've done earlier, long prints with this reel in the same position with no issues.

@matth can you verify that the encoder wheel spins with the filament sensor wheel when filament is ran through it? if the screw holding the assembly is too loose the filament wheel (with o rings) can spin but the encoder will not.
Also if your idler tension on the tool head is too loose and you're getting partial strip outs or the filament is slipping on the hob this will look like a false positive

I have been doing A LOT of Ninjaflex printing to figure this out. From what I can tell the weight of the reel is VERY significant to this. When feeding ninjaflex through your fil tube, notice how tough it is to pull through, there is so much more resistance on the path to the nozzle than we ever have had before. While this is incredibly significant it did not solve the issue mentioned earlier about ninjaflex not extruding 2 layers in. For this I discovered it was isolated to the opaque ninjaflex colors and not the translucent ones. So you are going to have a harder time with Black, Purple, Red, etc. All the colors that look opaque when extruded. Other colors like Ninjaflex's Blue, looks translucent when being extruded and does not have the same extrusion resistance.

My theory: The Opaque colors are absorbing heat and expanding inside the chamber while retracted, this causes the prime to under extrude because it is swollen and now trying to squeeze through a tighter chamber. This issue can begin at any point in the print and appear to just be a section that under extruded, once it finally primes past the swollen section it then begins to print normally again until another swelling occurs while retracted. Ninjaflex is not rigid and does not get eaten away by a strip out, it just keeps pushing and trying to extrude, which eventually it does. This would explain the severe inconsistencies we have been having with failures being with different models and times in a print.

My solution: I usually keep my filament runnout threshold at 20mm as a default for ninja, I double check my filament sensor tension, and I always try and unspool slack from each reel as often as I can. I made improved profiles (posting those soon or you can find my current testing settings listed in this document here: http://devel.alephobjects.com/lulzbot/filament/Testing_Docs_Master/Quiver-Setting-Guide/)

While I do think printing with Ninjaflex is going to be more difficult and take a little more attention, it is completely possible with the Taz Pro.

@matth I can confirm what @anolen is saying with my own experiances. Anything you can do to reduce the drag on the filament and reduce the chance of the filament stretching will help a lot.

kent added a comment.Apr 16 2019, 11:12 AM

It is possible to disable the filament runout sensor by turning the threshold all the way down. If it is reading lots of false positives for a specific material or environment, then this is an option for the user. Beware the false false positive -- where you think you have a false positive but actually there is a legit filament feed issue that can't be detected just by looking at it or it's not really showing up in the print as artifact. I think this is likely the case for large spools with lots of weight and friction.

@tutley I confirmed that the wheels in the sensor spin and the pin states change when I move the filament in the sensor. On the tool head the idlers are centered for both nozzles as well.

@anolen I've mentioned reel weight on a few tickets, as it seemed like I was encountering issues that no one else was running into. My theory was that two large reels on the side are rubbing into each other and stop filament from being pulled. I tried to remedy this by printing an overhead spool holder from here: https://www.thingiverse.com/thing:1608830. Currently I have polydissolve on top feeding into nozzle 1 and PLA on the side feeding into nozzle 2. It wasn't until I started printing with polydissolve that I started having false positives, always for nozzle 2, the PLA.

anolen added a comment.EditedApr 16 2019, 11:23 AM

@kent, yeah, the only way I can verify an actual strip out for Ninja is by removing the filament from the toolhead and looking at it. If there is a dramatic kink in the filament it indicates the feed gear had been working on that area for too long and it is indeed a strip out.

@matth Yeah Polydissolve profiles need to be updated, are you printing at 235?

@anolen The filament failing for @matth is 3kg Black Polymaker PLA mounted on the side of the printer. The 1kg Polydissolve mounted on top is printing fine.

@anolen Yes, at 235. Sorry for the confusion but the false positives have been with the 3kg reel of black polymaker PLA.

To add on to @kent 's comment, each failure I'm able to successfully resume a print and it will run for a couple minutes. I can confirm that it is still extruding and not stripping out in the toolhead.

@jebba oh! I misunderstood. I recently ordered some Black PLA, I am unsure why he is getting errors with a PLA, I haven't heard of that from anyone else that I can think of but Ill keep and ear/eye out and let you know when my Black Polylite comes in and I have a chance to check updates on it.

@anolen We're trying some 1kg lulzbot green PLA with the polydissolve to see if the weight of the reel or color is somehow related. Also, @matth is using your Polydissolve profile from the spreadsheet. It looks like it is printing ok. We are adding "brim" when we enable support too.

west added a comment.Apr 16 2019, 11:46 AM

@matth can you post a picture of your setup?

im surprised you're not having issues with the top feed reel

per T6465 you can try cutting 200mm off your feed tubes, this will reduce the amount of drag on your filament and prevent the PTFE tube from getting caught in the frame.
might help if the filament is slipping on the hobbing

jebba added a comment.Apr 16 2019, 1:47 PM

@tutley when @matth brings in the printer, set the filament spool holders to the "official" position. The top spool was used when we saw some issues with doing 2x 3kg reels right next to each other. But even moving one above we're still seeing issues with the one large reel, so moving up top didn't fully solve any issues.

sounds good. Is he going to bring it to Aleph Rockies?

matth added a comment.Apr 16 2019, 2:36 PM

I will come down tomorrow. I replicated a false positive with the pictured setup, and now I've replaced the 3kg black PLA with a 1kg green PLA. I'd like to see if it will be replicated again. (I haven't shortened the tubes yet either.)

jebba added a comment.Apr 16 2019, 3:35 PM

@tutley Ya, it's going to Aleph Rockies. He was able to have it fail with 3kg black PLA and 1kg green PLA.

So @anolen witnessed the issue and was able to identify the problem and apply a fix. She'll be able to describe it better than I can, but from what I understand my errors came from a faulty wire for nozzle 2 in the tool head. I will test to confirm the fix, and then move on to testing big reels.

So we got it to fail due to the filament sensor.
I then used the LCD in "Change Filament" to extrude filament from T1 and I notices that it did begin to extrude filament right away but every couple cm the extrusion would vibrate like it was being extruded at too low of a temp. I had this issue with 2 of my 3 Alphas. When I took the toolhead off to check the wiring I noticed this:


After taking it apart, here is a better look.

I rewired that section and it seems to be extruding consistently now. @matth Let me know if you have issues with it in the future.

On another note, @kent is this an issue we are still seeing with Betas or was this a fluke?

This was orginally opened for flexy material on the runout sensor- @MikeR or @anolen any updates? Is this ticket okay to close?

I have not had a chance to print with flexy recently to check.

I suggest turning your sensor off if you are printing flexible material, I do not have it on when i am testing it, I eats up too much time.

jebba added a comment.May 22 2019, 2:12 PM

If it is generating false positives with TPUs is it possible to disable it in the Cura profile? If not, we should document this for users that TPUs should have runout disabled.

Agreed, doing it in the the profile would be really nice and convenient but if not, some form of communication to the customer is needed.

@anolen @jebba you can add M412 S0 to turn off the sensor in the start gcode, and then M412 S1 to turn it back on in the end gcode

i know that the start/end gcode is not profile specific, but those are the gcodes to turn on/off the sensor

maybe we can add a toggle option in cura that turns it off/on

@tutley That would be ideal. @karrad, @alexei and @victor_larchenko, this possible?

@alexie @anolen @karrad @victor_larchenko

if we implement this into cura, it may be useful to include a run out length section as well as the ability to turn on/off

M412 S ; S0 turns off filament sensor, S1 turns on filament sensor

M412 D ; D15 sets the run out length to 15, D25 sets the run out length to 25, etc.

@tutley ^^ This would be very useful.

@anolen This should be doable, we have added variables and new parameters in the past.

@tutley Is this the official R&D recommendation to disable filament runout when printing NinjaFlex?

@karrad according to @anolen's testing yes but i think we could try increasing the runout distance to like 25mm or something and that may work as well

@karrad and @tutley, I have working Profiles for ninja and I feel this is worth testing now. Shall I wait for the Cura update to test the new setting or just go ahead and set it manually and test it that way for approval?

@anolen Go ahead and post the exported profiles for testing, but we will still want to get them into Cura LE if they are ready. Have you already logged them somewhere?

@karrad The profiles need to have a final model test run, I was wondering if you'd like me to wait for the filament runnout setting in cura to run that test and then post or if I should run the test with a manual change of the fil runnout setting and then post.
All settings are saved and accessible in my Quiver settings spreadsheet.

@anolen Go ahead and post, if it is something we need to adjust for all ninjaflex profiles we will take care of updates at that point.

matth added a comment.May 28 2019, 1:12 PM

So I'm still getting runout errors. I'm now using a production model that Jeff picked up.

First big print was perfect. No issues, no runouts, no peeling, great results.

However, my last print ran into the same issue again. In the first 1-2 hours of printing, it kept having a false positive filament runnout error on extruder 2. After resuming the print successfully after the third time, we turned off the sensor so it'd just keep printing. I left for the weekend and sure enough it had a true positive shortly after. It was so stripped out I'm still trying to remove the filament.

All the false positives leading up to a true positive leads me to believe something is some correlation, but I don't know what.

Serial number: KT-PR0050NA-1001
Firmware: 2.0.0.124
Filament: 2x Polymaker Polylite PLA, lulzbot green (E1) true black (E2)

@matth have you confirmed proper idler tension on this machine?

matth added a comment.May 29 2019, 9:17 AM

@tutley Yes. They looks like they're in the center. They were set by the factory and I haven't touched them since we got the machine.

@matth are these with the 1 kg or 3kg reels?

@DaniAO I discussed it further with @matth and I think in this case it wasn't a false positive. When the sensor was disabled, it jammed not long after, so I don't think it was a sensor issue.

anolen removed a subscriber: anolen.Jul 1 2019, 12:29 PM