On filament tolerances, printing with 0.25mm nozzle and filament change.
For a project I switched to a 0.25mm nozzle to get better definition in small structures. The project also requires printing layers in different colours. As my printer is the plain i3MK3S, this means manually inserting filament change commands in PrusaSlicer. I did switch to the 0.25mm print profile before generating gcode.
The manual filament change never gave problems printing with 0.4mm filament, but now it does with 0.25mm. Typically one colour tends to get stuck while unloading, sometimes requiring a partial disassembly of the extruder. The problem appears to be a bigger than normal blob created at the hotend side getting stuck in the lower end of the PTFE tube. Other colours didn't give problems. This colour is of a different brand than teh others, and its dimensions appear to be somewhat smaller compared to the other brands. I did some measurements with a micrometer, put those into a spreadsheet to calculate an average. My problematic black filament clearly deviated from the others. But first a description of my semi-scientific method.
Each spool was unwound for 4 meters, and measurements were taken each meter, horizontally and vertically. To keep it simple and fast, I didn't try to use the complete resolution of the micrometer, but rounded to the closest five micrometer. All the values were averaged, divided by the standard value (1.75mm) and multiplied by 100 to get a percentage.
The problematic spool came to 97.83%, the other used spools were within 1.5% (98.6%, 99.89%, 100.31%) the specified 1.75mm.
Why this happened only with a .025mm nozzle is not very obvious. It might be that due to higher pressure in the nozzle, the extruder forces are larger, and these start to slip. If the filament remains stuck, the extruder slips, starts removing material until it looses grip completely. A lesser diameter of the filament makes this worse.
The condition of the PFTE tube is also a possible cause; The lower part is warmer than the upper end, so is more flexible and can accomodate the blob. The upper end is cold, meaning somewhere the blob gets stuck. The blob size might be influenced by nozzle size / filament pressure.
The procedure for filament change by an M300 command is slightly different from the normal filament unload. Most annoyingly it is not very helpful in this situation, and tells you to fix it yourself.
Maybe someone can shed some light on situations like this. My conclusion for now is to get a better quality black filament and mark the current one as "0.40mm"-only.
Heat break will cause jams like this: normally filament pulls out okay, though sometimes you need to open the idler door and just pull the filament up.
There is a known issue with the Prusa supplied heat break. It has a step that causes jams, and can also cause unload problems after printing a while. The filament usually looks like this.
I've had the same thing happen to me. Had to take out the PTFE tube completely to get the filament out. It looked just like the image pictured above. I was multi-layer color printing as well. Two different colors, both prusament filament: lipstick red and white. First the Red, then the white. When it stopped to change filaments, to the white, the red wouldn't come out. Had to go as far as removing the tube to get it out. While it was in the headend I had to put a hair dryer to it to soften it up enough to pull it out.
So since this is a known issue, what is being done to fix it by Prusa?
I'm not sure the problem can be completely fixed in the printer construction or firmware. As I see it, when the filament diameter is below the standard, the BondTech gears have less grip on it and cannot pull as hard on it. This means slipping and the gears lose grip altogether by scraping the filament even thinner. Adding a slip detector might help? It would complicate the extruder construction for sure.
Since I had the problem, I had to use the problematic filament again combined with the 0.25mm nozzle. By being alert and pull very hard when requested, I prevented the problem from happening again.
The simplest solution (for the problem I had) is to demand filament that is closer to spec. Maybe this is the reason Prusament exists :-).
Soon I hope to get a MMU installed, so I have a reason to avoid filament that gives this problem. There is no option to assist in pulling there.
At last I managed to make a half decent image of the stuck filament. Note the point where the BondTech gears lost grip on the right.