Newly installed MMU3, prints look good but crashes on wipe tower  

RE: Newly installed MMU3, prints look good but crashes on wipe tower

FWIW, the MK4S is on firmware version 6.2.6+8948 (to fix an issue where it would stall out and never load in the filament, as suggested on another post) and the MMU3 is on version 3.0.3.

RE:

I consider it a bug, the nozzle spends too much time too close to the wipe tower surface while the MMU swaps filaments:

- the nozzle melts into the wipe tower

- any dripping filament builds blobs that eventually catch the head

My solution is a custom tool change G code e.g. G1 X10 Y10 F5000 that moves the head to an empty spot on the print bed while the MMU is busy. For small prints also G1 Z10 F5000 works but this will crash once the wipe tower exceeds 10 mm in height (adjust number as needed but it gets slow).

RE: Newly installed MMU3, prints look good but crashes on wipe tower

Well that sucks but I appreciate the answer. I don't understand how, if it's an inherent problem, I see all these photos of super clean wipe towers. I just figured that meant I had something wrong. Okay so do you know if there is any good instructions on adding custom GCODE? I vaguely understand it but I also know you can screw a lot of stuff up if you mess up GCODE. And is that something that can be locked in? Or would I have to edit the GCODE for every print I want to do? Sorry for the ignorant question, I'm a mechanic not a programmer haha!

RE:

Yes... the MMU isn't really mainstream. There are too many situations where the printer's "reset" button serves as "cancel" button e.g. with the firmware in a loop you can't get out of. Also, error messages are not good e.g. stuck in "G-code execution" often means the MMU can't complete something because the hardware is in an unexpected state. All that said, I don't want to bash it too badly because it works quite nicely. The snags are on the learning curve, not so much in operation once it's "pipecleaned".

I copied this from my last successful print tonight.

G1 X10 Y10 F4000

Paste it into "tool change G-code" (Prusa Slicer: Printers/Tool Change G code) and watch the printer for the first few color changes. It also shows in the toolpath preview. The head shouldn't travel over printed objects, so place the wipe tower accordingly or edit X and Y using preview. I usually paste it in without much thought´, even if the nozzle scrapes against the print it's not critical. But a solid blob of several millimeters on the wipe tower is...

RE: Newly installed MMU3, prints look good but crashes on wipe tower

Awesome man, I appreciate it! I work in a different state than where I live so I'll have to give it a ride when I get home in 2 weeks. Thanks!

RE:

Wow... there's a lot of trial-and-error involved... you didn't pick an easy challenge. 

Some of the "snags" I can think of:

- it's probably better to set the springs in the extruder idler not too tight. Same goes for the MMU, as the pressure can wear out the filament, especially in small prints (little forward motion, mostly back-and-forth). Don't forget to calibrate the extruder if you change the screws.

- have a look at filament friction, how hard the extruder needs to work to pull it through the whole system. Don't place the spools for good looks, but try to make it move as easily as possible. I actually use a tiny little bit of oil on the filament but that's a topic in itself, simple placement of the PTFE tubes for bend radius can make a big difference.

- check that the FINDA sensor switch point is reasonably "half-way" between "filament in" and "filament out". What you want to avoid that it's marginal and randomly mis-triggers (which will stop the print). As you can rotate the sensor to change the distance, it's an easy adjustment. Determine the two switch points, then set it halfway and tighten it.

- there are printed mods that straighten the five exiting PTFE tubes. Convenient (insertion gets easier) but not necessary.

- problems from one print can carry over to the next one. If in doubt and a hurry, sacrifice one meter of each spool. Or slide between fingers to check for damage (the pulley pattern itself is OK unless it's flattened)

- if there are any open "construction sites" (bed adhesion, filament drying, belt tension), set the MMU aside and solve them first. It only gets harder.

- If you remove the 90-degree PTFE tube bend on top of the extruder (depends on version), make sure the umbilical cannot droop and jam between X motor and print head. If this happens, it's nasty (layer shift and no clue, why). The trick here is that the PTFE tube provides some lift to the umbilical from the print head side.

- If the MMU plays dead and refuses to calibrate on power-up, check the FINDA light (filament fragments stuck? Push some filament through from either side or disassemble)

- Do yourself a favor and make sure the filament is dry, and stays dry. Don't assume it comes dried out of the sealed bag (it's  often barely printable with some controlled water content from fabrication as plasticiser). However, extremely dry filament may become brittle and difficult to handle in the MMU - it just snaps. Leaving it out one night can be a good compromise, some brands stay brittle though - not everything works. Where I live (50% RH), leaving filament loaded is not an option for more than a night or two. Arguably it may be still good enough but with the MMU I tend towards better margins in everything - there is just too many things that can go wrong. 

RE: Newly installed MMU3, prints look good but crashes on wipe tower

Whew, that's a lot to go through haha! I'll check it all out when I get home. The strange thing is that it prints the actual part totally cleanly, no problems, regardless of which or how many spools are being used. It's only on the wipe tower where issues arise.

RE: Newly installed MMU3, prints look good but crashes on wipe tower

edit time expired... one more random thought, especially if the printer isn't at hand:

Spend time with the slicer. Understand the different "Preview" settings e.g. "actual volumetric flow" or just "speed" on some test object. Experiment with wipe tower settings (for example "sparse wipe tower" can be very useful when the object is designed for color changes in narrow Z intervals - this strips hours off the print. But you need to manage fan shroud collisions yourself). These are just tools to tackle the MMU problems when they come. And they will come 🙂