RE: Nozzle cleaning consistently fails with PETG
After feeding the filament into the tube, if you keep your finger on it. You feel how it keeps "moving in" more filament while it's flushing the nozzle. Once the screen says "Is the color correct?" (or something like that), and you press "Yes" while keeping your finger on the filament entry – you feel how it pushes back some filament out again. I think this the new behavior that a firmware update made
I'm just speculating, but I think the nozzle cleaning failed was due to filament sticking in the nozzle. And I distinctly remember it starting to push filament back out after loading it.
RE: Nozzle cleaning consistently fails with PETG
I still do not understand, are you saying that the FW is now setup to continue extruding filament after confirming the color? I am not sure what yyu mean by pushing back. Did you mean pulling back (retracting) or pushing out (extruding)?
If the behaviour is to extrude more filament after confirmation, then the filament will stuck at the nozzle. I ask because I experience this with Nylon. After a couple of minutes of oozing then this is stopped. Is this the behaviour you are talking about?
RE:
I still do not understand, are you saying that the FW is now setup to continue extruding filament after confirming the color? I am not sure what yyu mean by pushing back. Did you mean pulling back (retracting) or pushing out (extruding)?
If the behaviour is to extrude more filament after confirmation, then the filament will stuck at the nozzle. I ask because I experience this with Nylon. After a couple of minutes of oozing then this is stopped. Is this the behaviour you are talking about?
That was happening to me with Prusa PETG and color changes mid-print. After confirming the color, the filament continues to ooze out of the nozzle for around half a minute and then drags this over the print. I basically have to pause and manually remove it. Was thinking about adding filament retraction in the color change g-code section.
RE: Nozzle cleaning consistently fails with PETG
Retract. It retracts after it confirms the color.
RE: Nozzle cleaning consistently fails with PETG
Also, I don't really want to sit and baby sit my prints for the first 5-10 mins each time.
Tough.
It was usual with early RepRap machines to watch the first two or three layers in case of early fails. Prusa Mk2 reduced it to watching just the first layer and Mk3 just to watching all the first layer perimeters and the beginning of infill. With my XL I glance a few times at only the first two perimeters and then walk away. The Core One is of the same generation, I would expect to monitor the the first two perimeters.
A few minutes at the start when the whole print may take hours or even a few days is trivial.
Cheerio,
I don't understand the 'it used to be worse so why try to keep improving on it now since its better' mentality. Nothing would ever improve if everyone lived by that. It's clearly still an issue, even if its not as bad as it was. And its also a largely fixed issue.
There are multiple things to unpack here:
- There is something inherently wrong with the nozzle cleaning step
If it isn't happy with the readings during cleaning after 5 probes it will do all 20 probes and fail. I've never had a nozzle cleaning stage conclude without error after the 5th probe. This is an error
It also doesn't really do any cleaning, especially on certain steel sheets. Nothing comes off the nozzle during the probing unless using the smooth sheet
These were my main concerns when making this post in the first place, the nozzle cleaning stage just seems to not work properly- There ARE solutions to this issue outside of poking the build plate over and over
Other manufactures and also user mods add a silicone/similar brush and a scrubbing phase. Which just automates the cleaning with a brush before a print without the worry of the Core One crashing when the door is opened, which happens all the time and will be fixed eventually, but isn't yet.
I have never had a clean fail or a print fail with my MK4 with MMU3. This is because it doesn't load the filament until it starts printing. This eliminates the possibility for oozing. I would start a print remotely and it would just print without supervision. I had a camera to make sure it was working, but didn't need to baby sit it with a wire brush.- There are also likely other solutions i.e.
Retract the filament more after a print when the nozzle is cooled slightly to try and pull most of the plastic out of the heat zone
Wipe the nozzle over the build plate instead of just poking it (I'm pretty sure there is custom GCode in the Core One test prints on the USB to do this
I can't imaging its fun if you have a print farm of 5, 10 or more printers to have to sit scrubbing the nozzle on each printer at the start of every print so that your print doesn't fail during a stage which is meant to clean the nozzle, that just seems like a big problem to me. But hey, I guess we are just meant to baby sit our print farms because it used to be way worse.
Mate, some people are just stuck in the past.
If you really want to be concerned, monitor the serial output via usb, the amount of 'NOK' messages during the startup process is deeply concerning. It's amazing any print actually starts.
I fully expect these to be hidden in later firmwares.
RE: Nozzle cleaning consistently fails with PETG
when you say hidden, do you mean swept under the rug or fixed?
Also what are these sequences you mention? Sorry, I am not a complete newbie but I am certainly not connecting my printer to USB any time soon. I will do some GPT and see if I can figure out what you meant, but I believe you are insinuating that the firmware is somewhat rough?
As of late, I have been thinking of trying the new H2d and see how it prints in comparison. I might just put the whole prusa project behind... I am yet on the fence but these things worry me to some degree.
RE:
when you say hidden, do you mean swept under the rug or fixed?
Also what are these sequences you mention? Sorry, I am not a complete newbie but I am certainly not connecting my printer to USB any time soon. I will do some GPT and see if I can figure out what you meant, but I believe you are insinuating that the firmware is somewhat rough?
As of late, I have been thinking of trying the new H2d and see how it prints in comparison. I might just put the whole prusa project behind... I am yet on the fence but these things worry me to some degree.
My guess its they'll probably be reformatted so that Prusa can tell the errors are there, but your average bloke looking at a serial monitor probably won't. I've worked for companies that have done this in the past.
Nearly all related to the nozzle probing the bed and I presume readings coming back from the load cell.
Right now it's spitting out almost as many NOK messages as OK messages...and this is when a print successfully starts.
I could well be wrong but I find it hard to believe this is all normal operation of a printer....I don't think the firmware's 'rough', I guess the firmware is doing whats its been programmed to do, report bad/weird/out of range readings when they occur.
It's simple enough to monitor, just connect a usb-c lead into the bottom usb-c on the printer, other end into your computer, and then you can watch all the action with any serial monitor. I use the Arduino one that comes with the IDE as I have that installed and ready to go. You can also view the messages in octoprint if you have that set up.
I should mention I really like this printer, it's a little beast and I don't regret my purchase at all, but still kinda feel we're in the teething process of this machine. I have faith Prusa will address and correct these issues.
RE: Nozzle cleaning consistently fails with PETG
My issue still persists and has only gone worse.
The only way I can print anything is with custom G-code in the slicer where I do this:
- Auto home before heating *anything*. Homing will fail miserably if a) nozzle heating is on and/or b) bed heating is on.
- Turn off heating before nozzle cleaning and mesh bed leveling.
- If there is even a tiny amount of actual filament residue in the nozzle, it will still fail (as it should, I guess...). In that case retry will fail because bed heating is on. I have to reboot the printer and start over again.
Whenever there is a good day and it agrees to print something, the quality is good. But the amount of hassle and waiting before printing begins is just ridiculous, also when everything works normally. Small prints are done much faster and reliably with my old Mk3S, print quality being essentially the same for most of my purposes. Sadly also that printer just broke down with some thermal issue anomaly (not debugged it yet).
I think maybe in 6 months with better firmware and proper fixes to the heating / nozzle clean fail issues Core One can become a good printer, but even then I expect to need to customize the g-code to disable most of the silly startup procedures, which is basically just heating and cooling the nozzle and bed back and forth.
There are also other minor issues such as bugs in the firmware, door handle getting loose, annoying maintenance due to rivets instead of screws, etc. But I can live with those as long as it prints reliably.
RE: Nozzle cleaning consistently fails with PETG
Looks like I have finally solved the nozzle cleaning issue in my Core One.
I noticed these things:
- Printer was more likely to suffer from nozzle cleaning issues with certain materials. But not because of the material itself - because of the temperature setting for the heatbed.
- Again, printer was more likely to have this issue depending on whether the heatbed was already warm or not from previous prints.
- Turning heatbed heating off from slicer custom g-code seemed to help most of the time.
My theory: the heatbed is large and heating it up to 100 degrees or so requires fairly strong current and moderate voltage (24V?) via heatbed power cables. However, the loadcell sensor is low-voltage, sensitive analog sensor (I think). If their wires are located close to each other and not protected, it is possible that the motherboard will pick false readings from the loadcell due current changes in the heatbed power cables, especially if heatbed target temperature is reached and it is controlled with frequent ON-OFF style PWM approach.
I examined the printer from this point of view. It seems that the loadcell sensor cables are not protected in any way from such problems. They are not twisted pair, not 24V (just a guess), and not inside a protective aluminium foil. When I removed the metal cable cover from the back of the printer, I saw that the cables were all tightened together with zip ties. Of course, my printer being a Kit, I had built it this way myself, following the instructions.
So, I simply removed the zip ties and reorganized the cables in this back panel area so that all high-power cables are separated from the rest: heatbed cables, motor cables, LED cables. Then I used zip ties to create two separate groups: low-power and high-power cables and made sure they are as far away as possible inside the area that the back panel covers.
Before making these changes, I tried to print and got the errors, as usual. After the change, I have been printing the whole weekend and got it only once - when there actually was some filament residue stuck to the print head.
Overall, this change only took a few minutes and seemed to remove all the weird issues with the load cell sensor: nozzle cleaning failures and homing failures. To my understanding, it also makes a lot of sense why this worked. I had similar phantom reading issues with a CNC style machine's limit switches a few months ago because of strong currents in stepper motor cables, which reminded me that this could be a similar issue.
This makes me wonder, why Prusa has not done anything to protect the load cell cables from interference from stepper motor, heatbed and extruder power cables. They are fairly long and by building instructions packed together with zip ties. In hindsight, that seems like looking for trouble.