ER34 At End of Spool - More $$$ Lost

[CTotten 18]

So here I am again with another issue that popped up at the end of spool on a long print.  When I saw the machine was sitting idle, I assumed it was the material (UM CPE) stuck in the printhead again.  That has occurred multiple times since a recent firmware (5.7.2 and 5.7.3) update, but can be recovered without losing the print.

But to my surprise, it was a new error (ER34).

This seems like an odd error to have on a when the machine is used frequently.  The printcores have not been moved in over a month.  The machine hasn't been moved.  I checked the connection at the top of the printhead, which looked fine.  Anyway, I seem to be stuck.  I am at the end of a spool in the  material station (filament is not visible at the front).  The printhead is positioned like it was in the process of unloading.  I am wondering if the unloading somehow unseated the printcore connection.

 

And the troubleshooting guide for an ER34 doesn't really tell you how to recover mid-print.  Seems like this isn't an error you should get while printing...

I guess I have to shut the machine down (and lose a partial print that has already used a full roll of CPE filament - $$$).  Then I have to manually remove the filament, and hope all recovers well after this.

 

I am still wondering about the "cold pull" routine done by the Pro Bundle.  While I was not here for this occurrence, since I updated to 5.7.2 the material change process has not worked well at all.  I am running about 50% on successful transitions between spools (with UM CPE).  On the successful ones, there is a pretty significant "pop" when it starts the unload.  On the rest, the filament is stuck in the printcore, requiring me to manually free it before the unloading will continue.  While this works most of the time, it is really annoying, especially on big print jobs.  And I have lost significant amounts of material due to these errors.

 

While I understand a "cold pull" is required to keep the material station running smoothly, can someone please look into the dynamics with CPE material specifically?  If I recall correctly, the process is cooling down to 75C before pulling the material out.  Is it necessary to cool down that far?  For all materials?  In the instruction manual, it says to cool down to 75C when using PLA to do a cold pull for cleaning the nozzle, but CPE is not PLA.  I have to think higher grade materials, like CPE or PC, probably don't need to cool down that far to get the desired result.  In my brain, cooling down too far is likely requiring too much force to do the required cold pull by the machine, resulting in the "stuck filaments" and possibly even this phantom ER34 error.  And I also have to think the temp it cools down to for the cold pull should be material specific.

Edited August 19, 2020 by CTotten

[CTotten 18]

I did a little digging on the internet.  Most sites recommend cooling to between 10-25C above the glass transition temp for the material being used for a cold pull.

Based on the TDS for UM PLA, and the UM written directions to do a cold pull, UM is recommending about 15C above the glass transition temp (PLA transition temp is` 60C, cold pull directions for PLA say 75C).

 

With that in mind, the UMS5 Pro Bundle is currently trying to "unload" CPE material at 75C, which is BELOW the glass transition temp (~82C by the UM CPE TDS).  So this seems to, atleast in my mind, explain all the issues I am having with swapping at the end of a CPE spool.  Cooling down too much, resulting in much higher force required to execute the "cold pull", resulting in either a "stuck filament" error or this recent ER34 error.  I would be interested to see if this is also the case for other higher temp materials. I print almost exclusive with CPE, so I have never seen an end of spool with anything other than CPE (and Breakaway, which is similar to PLA in terms of temps).  Also, I don't recall if I ever checked the temperature plots at the end of print, but I do not seem to have any issues at the end of print, so I should see what temp it cools down to in that case.

 

So does the UMS5 Pro Bundle use 75C for all materials?  If not, is it a "bug" that CPE is using this temperature on the end of spool process?  Seems like it should be cooling down to 95-100C for the "cold pull" with CPE, based on the stated glass transition temp.

 

Beyond the errors (lost prints = wasted $$ and time), I am concerned about mechanical stresses on the machine over time.  When it does a cold pull with CPE (when it actually works), it does make a significant "pop" sound, and you can see the rails flex and the machine shudder).

[UlrichC-DE 114]

One thing is for sure, if the automatic cold pull is so strong that an ER34 is triggered, it is not good for the head and the core. To do this manually you need pliers and lots of spinach in the morning.

I suspect this will cause some wear on the head and the ER34 will be easier to release.

On Core 1 it may still work, but Core 2 is more sensitive because of its switching mechanism. If I hadn't already done a repair I wouldn't write this.

On ER34, the printer loses contact with the print head.This happens exactly when the spring pins of the print head board have no contact to the pads on the core. So if an ox pulls the filament.

[SandervG 1,521]

Hi @CTotten Thanks again for your amazingly detailed post and research. I do believe that the break temp for PLA is 60 and CPE is 75 because it is suppose to 'break/snap'. It is being pulled back a little bit so a small tip forms, then it cools down, and then it pulls back quickly to generate the snapping sound you hear. Regardless, I will pass this feedback along to the material team to collect their thoughts on CPE specifically since you have had more issues with it and report back with you. 
They do test all profiles and materials in real life, we don't just deduct 15ºC from glass temperature, but feedback is always crucial.

 

About the ER34, I doubt this is caused by a cold pull. It is peculiar the print head is sitting in that position when ER34 occurred tho. Can you check the pcb of both print cores on the back, if they are still intact? Perhaps also in the print head? Sharing a photo could help. 

 

Could you also share the log of your print when it failed with the ER34?  

 

Thank you! 

[CTotten 18]

@SandervG Any advice on how to recover from this?  My assumption is that I should power off both the S5 and Material Station, then manually remove the material that is still in the tube (basically using the ER61 procedure)?  But since the filament is not visible in on the Material Station bay side, I have nothing to "push" the filament forward through the system with.  Should I be trying to cut the filament into pieces (like cut it before the decoupler and pull that piece from the back of the machine, then pull the rest out from the print head end)?  The ER61 procedure seems to assume that filament is NOT running all the way from the loading port in the Material Station to the Printhead... (never gives an option of the filament being visible above the decoupler).

 

Once I get the material unloaded, I can check the contacts on the PrintCores and the PrintHead.  It will be a few days as I am working remotely this week.  I will try to post the logs as well.

 

I should have taken a picture last night, but the machine looks just like it did when I was getting the Stuck Material errors.  Filament spool is completely empty, with no filament visible at the entry port. Print bed is lowered, and printhead is in the forward, left position.  Filament is still in the bowden tube at the top of the printhead.  I even have the slight stringing that I usually see from the item being printed to the printhead nozzle.  So it definitely seems like it was definitely in the material change over process when this occurred.  So in my mind, something had to dynamically "move" the printcore, since the likelihood of the contacts suddenly failing without moving around is astronomical.

 

And as a general question, does the material change routine use material specific temperature profiles, or does it use a generic "one-size-fits-all"?  It just seems odd to me that I have had SO MANY problems with mid-print material changes since the 5.7.2 firmware update, but I have had NO issues with the end of print routine.  Both of those should have be using the same process, if the goal is the same (to get the nice tip on the filament being unloaded).

Edited August 20, 2020 by CTotten

[SandervG 1,521]

34 minutes ago, CTotten said:

Any advice on how to recover from this? 

Recovering from an ER34? This is I believe only possible via a reboot. 

While I would happily continue to try and help you, I would also like to ask if you would want to create a ticket in our support system via this link, that way there are no blind spots in our data and monitoring the frequencies of certain error messages. ER34 is a good one to keep an eye on. 

 

Can you still unload it from the material station? While it may not be visible anymore, the feeder could/should still have a grip on it. 

39 minutes ago, CTotten said:

does the material change routine use material specific temperature profiles, or does it use a generic "one-size-fits-all"?

It uses material specific temperatures. 

 

40 minutes ago, CTotten said:

It just seems odd to me that I have had SO MANY problems with mid-print material changes since the 5.7.2 firmware update, but I have had NO issues with the end of print routine.

That is odd indeed, and hopefully we'll get to the bottom of it together. As I mentioned before, please also write a support ticket if you haven't done so already. 

[CTotten 18]

I finally had the time (and energy) to deal with this issue.  I have a couple pictures below of the state the printer was in, with it appearing to have been in the middle of the material change process at the end of a spool.  The blue lights on the material station are hard to see, but bays E (light gray CPE) and F (breakaway) were in use, with bay A having the spare roll of light gray CPE that should have been next up.

 

Recovery: I have to do several reboots.  The material was stuck in the printcore so hard that I could not remove it without heating.  After the first reboot, I got an ER61 (sort of expected) before I could heat the core up.  After the second reboot, I was able to manually heat the printhead to remove the material.  I ended up cutting the material off as far above the printhead as I could get (left me a few inches to pull with), then fully had to remove the bowden tube between the feeder and the printhead.  I pulled the rest of the filament out through the top of the feeder.  I have a picture of the tip that I pulled out, just in case, but I don't think it really means much based on the manual heating and interrupted unloading.

 

I did not see any more ER34 errors, so it seems the "communication error" was a one-time thing.  I also pulled both printcores and inspected and cleaned both the contacts on the printhead and on the pcb on the back of the printcores.  Both looked very clean to begin with.  The spring contacts all appear to be in the correct position (atleast all of them are in the same position). On my cleaning swab, there so no "color" after cleaning, so I am confident they were clean to begin with.  I have pictures below before cleaning.

 

I attached the log files below.  Atleast, I think they are the log files.  They were the only files on the thumb drive that I did not recognize.  But these files were after the second reboot.  I never got a chance to get the the maintenance window to download the logs until after I got to that point.

ultimakersystem-ccbdd3007914.5.7.3.config.tar.zip

[CTotten 18]

On 8/20/2020 at 9:50 AM, SandervG said:

It uses material specific temperatures.

 

Is this information that can be posted?  I would be interested to see what the temperatures are supposed to be, atleast for the UM materials.

 

And I understand it is not a set 15C above glass transition, but my concern is that the CPE temp my S5 is trying to use is BELOW the glass transition temp.  Unfortunately, for me to do more testing on this requires a lot of time to set up and a lot of material to waste, to force this into the end of spool behavior.

[CTotten 18]

I thought I would play around a bit this morning.  I ran a small, 7 minute print on the S5 using the same Light Gray CPE material.  I was trying to confirm that the end of print routine was using the same 75C temp for the cold pull.  That is confirmed. The material is cooled down to approximately 75C when the machine starts pulling the filament back on the unload cycle.

 

But funny enough, I got into the wizard for the material being stuck in the printhead.  I am pretty sure this is the first time that has happened at the end of a print.

 

The reheating at the tail end of the graph was during the wizard process to correct the material stuck in the printhead (heats it up, has you manually pull out of top of printcore, then cut off the tip, before it unloads it).

Edited August 22, 2020 by CTotten

[tinkergnome 927]

8 hours ago, CTotten said:

 I would be interested to see what the temperatures are supposed to be, atleast for the UM materials.

 

Material files are in the Cura program folder and on GitHub. That's the easiest way to confirm the parameters. If one is a bit "nerdy" one can even check the history.

 

https://github.com/Ultimaker/fdm_materials

Search for "break temperature" in the appropriate xml-file.

PLA uses 60C, Nylon 145.
For CPE it was recently increased from 70C to 75C (in Nov.  2019).

 

[SandervG 1,521]

Hi @CTotten, can you share your Curaproject file so we can reprint your file in the office and see if we experience something similar using CPE?

[CTotten 18]

On 8/24/2020 at 4:16 AM, SandervG said:

Hi @CTotten, can you share your Curaproject file so we can reprint your file in the office and see if we experience something similar using CPE?

Sander:

Unfortunately I cannot share the file.  It is proprietary.  The reason this is the file that is the bane of my existence is that it is a large print, requiring about 900g of one material (UM CPE) and another 150g of support material.  So I have no way to avoid a material change during this print.  But, this problem has occurred on other, smaller print jobs as well, when material changes occur.  Those are less frequent in occurrence because I don't have to print those as much, and the material changes are just the normal course so that I used all the material on the spool.

 

Just as a reference, before I updated to 5.7.2 (and then 5.7.3), I printed 10 of these without a single issue.  Since the firmware updates, I have not completed a single one without the material being stuck.  Several have been saved at the "Material Stuck Wizard" stage, but several have been lost by various errors that I am assuming are associated with whatever the main issue is (ER61, ER34, etc).

 

And this morning, I was trying to run the XY calibration to account for a new AA 0.25 core that we were trying to use for the first time, and the Breakaway material failed to unload from core 2.  This is a first.  I got the wizard prompt, started it, then the machine basically left the wizard on its own, sat there for 20 minutes, then went to a "Ready to Print" state.  I got no prompt to acknowledge the print was removed.  Both materials are still loaded in the print cores.

 

I am fed up.  Can you provide directions on how to rollback the firmware?  I have a copy of 5.5.12 stored on my network.  I am pretty sure that is the version I was using before I upgraded to 5.7.2.  I am losing too much production time trying to figure out these issues.

Edited August 28, 2020 by CTotten

[UlrichC-DE 114]

@CTotten

I would check the bowden at the back of the unit. There is a lot of Bowden in the back of the flow sensor.

If only half of it is plugged in, the DEPRIME will not work properly. These effects occur.

The DEPRIME has several stages. It is a loop of cooling, waiting and retracting. The configuration of the PRIME and DEPRIME is based on a standard printer with fixed Bowden length.

If the Bowden is too long, the first step is not really retracting, but only relaxing the material in the Bowden. This causes the filament to stick to the core.

The following steps, which involve further cooling and further retraction, will not work properly.

This is only an observation. This may sound bizarre but it solved my problems.

Please correct me if I am completely wrong.

 

@SandervG

To test this you could print any smaller object. It doesn't matter if it is a DEPRIME at print abort or after 750 gram CPE.

 

I wish you all a pleasant weekend

[CTotten 18]

2 hours ago, UlrichC-DE said:

@CTotten

I would check the bowden at the back of the unit.

Thanks.  I did, but I will check it again more thoroughly.  The weird part is that "most" of the time, the end of print depriming seems to work fine.  Most of my issues have been in the material changing process mid print, where the CPE in core 1 "sticks" almost every time.  I can recover most thru the stuck filament wizard, but a few have been lost, most recently because I am pretty sure the stuck filament caused an ER34 (comms error to printhead) trying to pull the filament out of the printcore.  Failing a print after a whole spool of material is painful.  So if this was a bowden tube problem, it should impact all depriming actions the same, not just the mid-print material change.

All of the bowden tubes unique to the material station are the ones that came with the unit new.  The bowden tubes on the S5 itself were replaced with parts from FBRC8, the US repair parts supplier.  I hope they are the right length, but those changes were done months and months ago (December or January), and this issue just started in the last few months, right after I updated to 5.7.2 (which did tweak the material change process, from my understanding).

Edited August 28, 2020 by CTotten

[UlrichC-DE 114]

True
If the Bowden is not right, the Deprime always goes bad.
The PRIME in Core 1 is then also not good and in Core 2 the filament breaks.

 

I can't interpret the LOGs completely. It has several NFC read errors. Otherwise I only read coffee grounds.
Unfortunately this is also the case in my LOGs. So I would consider this as normal.

 

I print large nylon parts and I have roll ends in it. It works fine so far.
Since the material station has errors in recognizing filament types, I only admit the most necessary. All other filaments I leave outside.
With the firmware 5.7.3 I have to put the material in the material station to the left wall each time, so it recognizes it at first go. With 5.7.2 he often detected over two shafts at the same time (that was not nice either).
So if the material station wanted to detect the filament during the Deprime I would only bet 50/50 that it actually works.
The suspicion would then be. That he loaded CPE but the Deprime of another filament makes the filament loaded.
Of course I don't know if the material station works like this.
Again just an idea

[CTotten 18]

On 8/30/2020 at 5:35 PM, UlrichC-DE said:

The suspicion would then be. That he loaded CPE but the Deprime of another filament makes the filament loaded.

Of course I don't know if the material station works like this.
Again just an idea

Possible, but I generally have all CPE loaded in the machine.  Well, CPE in bays 1-4 (sometimes 5), and Breakaway in 6.  I have noticed the issue with it detecting an adjacent spool.  It is a pain...

[UlrichC-DE 114]

The new version 5.8.0 shows a different behaviour when changing the spools. Somehow everything is different. Maybe this is coinced for your problem. However, the version is not completely error-free and has new surprises.

 

If I wanted to test it, I would cut the filament during printing in the MS or I would loosen the bowden under the feeder. Better to lose two meters of filament instead of 20 hours of printing time. If you then have the bug, the people from Ultimaker can recreate it. (When Ultimaker lends me a roll of CPE, I test .. address is known  🙂 )

Ultimaker CPE and Generic CPE each have their own profiles. But changing that doesn't make sense for you either. Because you print in different colors.

There are people on the firmware forum who are closer to the productive firmware.

Maybe @sj3fk3 can write if Ultimaker has the problem on the screen.

 

Again just an idea

[UlrichC-DE 114]

I also had ER34 a lot today. (Otherwise I have only 998, 56 etc.)

 

I noticed an (old) problem with the print head. The cable is plugged in. The cable leading into the print head is very thick and moves the plug.
In my case it lost contact as soon as I moved the print head. The upper / rear cover acts like a lever on the cables.

 

Maybe your problem is also connected to this print head connector.

https://ultimakernasupport.zendesk.com/hc/en-us/articles/115004169003-Reseating-the-Printhead-Cable-Ultimaker-3-Series-

I only unscrewed the cover with the two screws at the back and screwed it on again. I kept the cable under pressure towards the plug. If I get the problem again, I build a strain relief over the cable. That the angle of the cable is not so close to the connector.

 

Again just an idea. This time for the sympthom and not for the cause