[BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
Hello Prusa Community,
I'm posting here after spending considerable time attempting to get this issue escalated through Customer Support without success. My hope is that others who have encountered this can corroborate it, and that it reaches the engineering team.
--------
The Bug:
--------
When the pre-print Nozzle Cleaning Routine fails, the printer correctly pauses and presents the operator with three options:
1. Retry — Retry the Nozzle Cleaning routine
2. Ignore — Ignore the failure and continue
3. Abort — Abort the print
The problem occurs when Retry is selected. The subsequent Nozzle Cleaning routine executes with the bed in an incorrect Z position — rather than lowering properly, the bed is driven further upward toward the nozzle. The nozzle then drags across the print sheet surface as it travels between cleaning touch-points, leaving physical scratches.
Additionally, the travel move from the cleaning position to the bed leveling positions also fails to perform the correct Z drop, causing a second set of scratches — this time in the printable area of the sheet.
To be clear: the primary concern is not that the nozzle cleaning routine can fail. The concern is that the Retry path has a Z-axis positioning bug that causes physical damage to the print sheet.
Occurrences: I have seen this issue occur twice since I received and setup the Prusa Core One L. Upon the first occurrence I chalked it up to the printer losing track of it's Z-axis position and reran the homing procedure prior to restarting the print; however, not that this has occurred more than once, under the same conditions of a first-round Nozzle Cleaning routine having failed, followed by a second-round Nozzle Cleaning with this Z-positioning bug, I can no longer consider this a fluke.
------------------------
Reproduction Difficulty:
------------------------
Prusa Support has asked me to reproduce and record the issue on video. I have made multiple attempts, including deliberately blocking the nozzle with ABS filament while running a PLA profile (idle temp ~170°C). In every attempt, the cleaning routine has returned a success result — which is itself a separate concern — and the print process has continued normally - continuing to the bed leveling routine.
The failure condition appears to require a very specific set of circumstances that I cannot reliably manufacture on demand. I cannot reproduce it on command, but the scratches on my print sheet are physical evidence that it has occurred.
-------------
The Evidence:
-------------
The scratches visible on the sheet correspond exactly to the nozzle travel path during the Nozzle Cleaning routine — not during the print itself.
The following images show my various attempts to create a first-round Nozzle Cleaning routine failure (for the purposes of reproducing the issue). TO BE CLEAR, THIS IS AN INTENTIONALLY MANUFACTURED FAILURE CONDITION (or attempt to do so). This is NOT how this printer functions - it is thoroughly cleaned between (after and before) each print:
Images: Attempted Manufacture of Nozzle Cleaning Failure Condition
Images: Attempted Manufacture of Nozzle Cleaning Failure Condition
Images: Attempted Manufacture of Nozzle Cleaning Failure Condition
In each of these cases, while the offending PLA filament was not successfully removed from the nozzle, the printer "passed" the first-round Nozzle Cleaning routine.
The following image shows my attempt to create a first-round Nozzle Cleaning routine failure with a nozzle-covering plug of ABS filament:
Again, while the offending ABS filament was not successfully cleared from the nozzle, the printer "passed" the first-round Nozzle Cleaning routine.
--------
The Ask:
--------
I am not asking for a workaround. I am asking that this be treated as a firmware bug and escalated to the engineering team for investigation of the Retry code path in the Nozzle Cleaning routine. If anyone else has experienced this, please reply — additional reports may help get traction.
As customer service has so far refused to assist unless I can provide video evidence, I will also be requesting a replacement Satin Print Sheet, as in my attempts to reproduce the issue, I have sacrificed this print sheet.
I am starting to pull my hair out over here as I am consistently being told they (Customer Service) will not escalate the issue until I have provided a video of the issue being reproduced; however, the issue cannot be reliably reproduced - as stated and for the reasons listed above. This is not the level of customer service I had come to expect with Prusa and this is incredibly disappointing and discouraging given one of the primary reasons I chose to continue being a Prusa customer and purchase the Core One L is because of the long-standing history of exceptional support and customer service. I'd hate to imagine that those days are over...
To help this move along, I have included a copy of my Customer Support chat transcript - redacted. My apologies to the community for the noticeably growing frustration as this chat progresses:
- Transcript 1 / 2: Chat Transcript 1 of 2
- Transcript 2 / 2: Chat Transcript 2 of 2
Thank you in advance for any help you may be able to provide!
Best,
AX
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
I understand your frustration. I have felt like Prusa support made me jump through hoops once in the past, when they asked for video evidence of a problem for which I had already provided clear measurement evidence. In your case, it seems that in the first chat transcript the support agents simply did not understand the challenge of the two-step process. (Can't force the initial failure, but if it fails, the retry reproducibly goes wrong.) And the second chat got off to a bad start since you lost patience right away -- somewhat understandable, but in the shoes of the new support guy I would have switched to "let's end this call soon" mode too.
Two ideas how you might force the initial failure:
There is an early user report from the original Core One which describes that "nozzle cleaning failed" happened often with the textured sheet, since it does not "grab" the PETG filament well. If you have a textured sheet (and are willing to risk a scratch on it), maybe that is worth a try?
https://forum.prusa3d.com/forum/prusa-core-one-hardware-firmware-and-software-help/nozzle-cleaning-consistently-fails-with-petg/
Alternatively, maybe the error can be forced by putting some thin, slightly soft material onto the print bed -- a piece of thin cardboard, or a few layers of fabric or tissue? I assume that the printer detects "failed nozzle cleaning" by a gradual increase of the force on the load cell as it approaches the bed, which signals the presence of something squishy instead of the expected hard contact with the print bed. Trying this would have the additional benefit that the cardboard will protect your print sheet from scratches.
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
...m2c based on own experince...
Bed scratching
- try to do a "Auto Home" routine (Control -> "Auto Home"); after this it should work "normally", meaning the nozzle is not scratching
When I experience this behaviour, I immediately stop the print and do a "Auto Home" routine. Don't know why it scratch the bed sometimes...
Nozzle cleaning
- I don't get the point/reason why you do "AN INTENTIONALLY MANUFACTURED FAILURE CONDITION". Keep the nozzle clean/clear of filament debris eg. after filament loading. Use a cleaning fleece, this is very imprtant when using a ObXidian nozzle. Heet the nozzle to print temperature, give it "rotating cleaning movement" and let it go.
From my understanding the "cleaning the nozzle routine" is not intend to clean what you "fabricated"...
Core One L
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
- I don't get the point/reason why you do "AN INTENTIONALLY MANUFACTURED FAILURE CONDITION". Keep the nozzle clean/ [...]
I don't think you got it. Nozzle cleaning failures sometimes happen (sporadically, rarely) -- that's not what the OP is complaining about. But if and when they happen, the firmware offers a "Retry" option, and that appears to be buggy: According to the OP, in two out of two cases where he experienced this scenario, Retry caused the nozzle to crash into the print sheet.
Because Prusa support demanded a video demonstration of the second step (Retry failing), the OP is now trying to recreate the first step (nozzle wiping fail). It's just a means to an end, namely the reproduction of the presumed software bug in the Retry step.
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
@Jürgen
I agree with you that there might be a software bug. But I still don't think that plastering the nozzle full with filament will get us the info about "Why?" this happens. Would I be support and see this nozzle I would suggest to clean it first...just to be sure everything is as it should be duricng "Normal" use...
Yes, I also got such "Nozzle cleaning failure" in the past combined with scratching the bed. But the nozzle was clean, there was no debris on the nozzle tip. I just stopped the print and will stop when it occurs again, do an "Auto Home" routine and start the print again.
Yes, I agree...this is annoying...and especially not a soulution when you start the print not being next to the printer...but it is - for me - a "workaround" I can live with. But it doesn't give us a hint where the "bug" is living whithin the code/machine...
Thinking about when it happend during prints to get a pattern, but I can not see a pattern...at least an obvious pattern...
Core One L
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
Well, if the firmware offers a "Retry" option, it should work. It might not be successful (again) in cleaning the nozzle, but it must not bring a risk of damaging the print sheet. This is all the OP seems to be asking for, and it seems like a reasonable ask to me.
Some way to reproduce the problem is required -- to convince Prusa support, and to enable the software developers to systematically test and debug. So what else would you do than to try and provoke a failed nozzle cleaning? I like the idea of putting some soft material on the print bed (instead of on the nozzle) to simulate a cleaning failure, as suggested above. Seems less messy and easier to control/reproduce; but I have not tried this myself yet.
RE:
I reported this to Pursa CS 4 times in January, I could even make it repeat. The new Beta firmware 6.5.3, does nothing to address it. CS reached out to me recently and attempted to send a replacement load cell due to "microfractures". Package was rejected by customs, see if they ever get it to me. Don't believe for a moment that is the problem, as it only occurs during a specific squence of commands. The lack of problem solving ability is beyond comprehension. Or the best one yet, factory reset. They seems to suggest this every single one of their firmware issue. Even if factory reset "fixed" the issue for the moment, it is not a solution. Factory reset should never be required to fix a firmware problem. Imagine if you had to wipe your phone for every problem it had.
Here are all the known posts on this topic.
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
Very, very interesting:
https://www.reddit.com/r/prusa3d/comments/1rqib31/does_prusa_monitor_their_github_repo_anymore/
My workaround - making a "Auto Home" routine - fits in the issue description "[...] This change in behavior results in the printer losing track of z height slightly (its an accumulated error between prints without a power cycle) and aggressively drags the nozzle into the bed during nozzle cleaning [...] of the Reddit poster.
Core One L
RE: [BUG] Core One L: "Retry" Nozzle Cleaning Routine Does Not Lower Bed — Causes Nozzle to Scratch Print Sheet
@Ruebarb: Yes, you already postetd the link to Github where the issue was described and a soulution presented in January (!). I didn't read this Github post, have to admit never done it...just stumbled over the Reddit post...
Core One L