The "Filament 3 Error" - Successful loads turn into unload and MMU Load Failure
Hello everyone! As a background, I own and operate 7 MK3's, with 3 of them having MMU's installed with a 4th on the way. They all run around the clock and therefore have accumulated 10's of thousands of hours amongst them. I've dealt with many issues, but there's one that I haven't seemed to be able to resolve: The Filament 3 Error.
"Filament 3 Error" isn't something that shows up on my screen or anywhere else for that matter. But for some reason, all three of my printers have experienced this issue over time, but ONLY when switching to Filament 3. Here's how it goes:
- Printer is doing fine on filament 1, 2, 4, or 5.
- Gcode calls for a swap to filament 3. Unload goes perfectly.
- Selector moves to Filament 3 without issue
- Filament 3 begins loading
- Filament 3 is successfully sensed by the MMU FINDA, and is then sent down the Bowden quickly. The hot end FINDA successfully senses Filament 3 and stops the rapid descent of the MMU. Then the magic happens...
- The printer acts as though it is going to print with Filament three, but never begins the wipe on the tower.it
- The MMU retracts the filament.
- The above cycle will repeat steps 4-7 four to five times before it gives up, the print is paused, and then the hot end retreats to the left and the bed comes forward.
Has anyone else experienced this? Everything is going correctly MECHANICALLY multiple times consecutively which is a feat in and of itself. But there's clearly something wrong in the gcode or the printer software causing this to happen. All three printers have had this happen from time to time, so I'm beginning to think it's the slicer or gcode causing the failures to happen.
Temporary fixes (~50% success rate at recovering prints, different tactics work for different prints)
- Reset button on MMU and resume print. This sometimes is all it needs.
- If 1 fails, pause print, reset button on MMU, load filament 3 to nozzle, resume print. This will often work, but it will cause for the MMU to throw the "light up like a Christmas tree crash error" when it goes to unload filament 3. So it's manually pull out Filament 3, then reset MMU, resume print.
- If 1 and 2 fail, then you convince it to feed the filament to the hot end while manually forcing the filament into the nozzle by pushing the filament on the spool side of the MMU. This SOMETIMES works, and is a last ditch effort before the printer gets flipped off and the wasted parts get hurled across the room.
TL;DR - filament 3 frequently goes through a cycle of successful loads without ever restarting the print, then throws an error and pauses prints.
Any tips or guidance is welcome! Printer and MMU running most recent firmware iterations.
Does this happen with different filaments? I have seen that behavior on one filament I have. Sometimes it looks like it loads fine, but gets reloaded anyway.
I have just added an indicator LED on the filament sensor and have seen it once do that and I think it was because the sensor wasn't triggers all the time with the filament loaded. Very irregular filament dimensions? I have adjusted the sensor a little and at the moment it looks fine.
I tested this hypothesis and it appears to be the correct path to take. I just can't understand why this would be problematic. The FINDA at the MMU is triggered by a ball bearing, as is the FINDA at the hot end. There should be some allowance for minute diameter differences between filaments. The physical characteristics, aside from diameter (hardness, color, opacity, awesomeness) shouldn't matter in my case because the FINDA's are both successfully recognizing "yup, there's filament up in here."
Come to find out my Priline Orange PLA and Mika3D gold silk filament are the problem children, but with this caveat...ONLY the oldest spools. I tested several filaments on the Filament 3, 4, and 5 channels (1 and 2 are exclusively dedicated to white and black) and the only 2 failures were older spools of the Orange and gold. Newer spools of the same filament performed fine. Interesting, and fortuitous that I had multiples so that I can complete some orders that are building up.
Time to pull a CR-10 from storage and print up the extension cylinder for my food dehydrator. My filament appears to be wet.