It started randomly underextruding during a print. I paused the print, manually extruded some filament, it seemed fine, so I started the print again. Then the extruder started clicking and it stopped extruding. Unfortunately, the filament broke off trying to do a cold pull, so I had to take the extruder apart. I unscrewed the nozzle from the heated block and found a lot of black ash. That or dust was probably the culprit. I pushed the remaining filament out of the PTFE tube, tried to clean up as much of the ash as I could without completely disassembling the hot end. I switched to a new nozzle. Since I don't have a blow torch, I'm going to try to bake the old filament out of the nozzle. The cheap nozzle/PTFE tube combo I bought came with a tube that was exactly 4mm OD, while the nozzle holes are slightly larger than that. What I think happened was that the pressure caused a film of melted filament to flow around the end of the PTFE tube, between the tube and the nozzle wall, and then up into the threads, where it carbonized. Some of that ash might have made its way back into the nozzle. I went back to the old PTFE tube I had that was ~4.1mm OD and fits more snugly. Even if that wasn't the cause of the clog, the ash is bad. PTFE tubes should be a tight fit in the nozzle. The tube's ID should be just a little large than the filament so that it can slide freely, but if it's too large, the filament can buckle.
Lesson learned: buy tight tolerance PTFE tubes or use an all-metal hot end.
I got it all back together and extruded about 50mm of filament. All good. I then tried to restart the print. Before I killed the print job, I noted what the Z axis was at. I modified the gcode following this guide to try to restart the print at the last layer. The gcode modification appeared to work fine, but two things were wrong: 1. The new nozzle was about 0.1mm less screwed in (the ash must have gummed up the threads), so the nozzle was a bit too low. 2. It was printing about 0.5mm too far to the left. The X-axis shift could have been due to a slightly different home position or the build plate shifting slightly (more likely). This goes back to what I've been saying about 3D printers NOT being precision machines. UGH. I killed the print and started it over. The first few layers seemed fine, but then it started under-extruding pretty badly again. Killed it, extruded ~100mm manually no problem. Didn't seem clogged. Started a test print cube. The first ~10mm had lines on the sides of the print and one missed layer; these typically indicate inconsistent extrusion. Inconsistent extrusion has many causes, but the main ones are partially clogged nozzle (pretty sure it's not), inconsistent temperature, too low temperature, and poor filament quality. I already know the filament is poor quality/cheap, but I can't help that. I tried turning the temperature up from 195 C to 200 C, which seemed to help a lot. It's possible that the ash in the threads is acting as an insulator. Also, because the nozzle didn't screw in as far, slightly less of it is covered in insulation now. These things might be keeping the nozzle from getting as hot as it was before I replaced it, so I may have to run at a slightly higher temperature now. The layer adhesion at 195 C was nowhere near as good as at 200C, so I think that's my new minimum temperature.
I decided to also check the filament diameter again. Near the beginning of the spool, it was 1.7 +/- 0.02mm. Now it's 1.68 +/- 0.02 mm. That translates to about -2.4% volume, which is pretty significant, so that might also be part of the problem. Cheap filament is cheap...
The print failed again due to underextrusion. Not sure what's going on. I tried increasing the extrusion multiplier (120%), but it didn't seem to do anything. It's like it can't push enough plastic. I did another cube test print at 200 C for a sanity check. Looked pretty good. I really don't think the extruder is clogged. For the big part, speed was 40 mm/s at 0.25mm layer heights, while the test pieces were 40 mm/s at 0.2mm layer height. New hypothesis: the ash in the threads is so insulative that I need to run at a much higher extruder temperature in order to get the correct volume flow rate. Trying again at 215 C. Seems to help, cranked it up to 230 C. Looks much better, so that's probably the problem. I'll have to run at ~220-230 C for high flow rates. Not sure what else that will affect. I should probably run another bridging test.
Another lesson learned: Need good metal-metal contact in heat block. Ash/carbon and corrosion are bad.
Update: I experimented some more with speed and temperature. At 220 C, 40 mm/s, 0.25 layer height, the outer surface had a slight glossy appearance. I found out from my previous temperature tests that that was ~200 C (40 mm/s, 0.2 mm layer height) before the nozzle replacement. Upping the speed to 50 mm/s using the speed multiplier under "quick settings" seem to work fine, but the glossy surface finish went away, which corresponded to ~195 C on the previous temperature tests. I started getting under extrusion around 185 C. So the 15-20 C hotter I need to run now is somewhat consistent, i.e. 200 C = under extrusion, 220C = slight glossy-ness. Bumping the speed up to 60 mm/s and temperature to 225 C isn't glossy, but it isn't under extruding either, which is good.
Lesson learned: Extruder temperature and setting temperature are functions of flow rate and thermal conductivity. Higher flow requires a higher temperature set point. Lower thermal conductivity (ex: ash in threads) requires a higher temperature set point. But this only works to a limit.
The print failed again due to underextrusion. Not sure what's going on. I tried increasing the extrusion multiplier (120%), but it didn't seem to do anything. It's like it can't push enough plastic. I did another cube test print at 200 C for a sanity check. Looked pretty good. I really don't think the extruder is clogged. For the big part, speed was 40 mm/s at 0.25mm layer heights, while the test pieces were 40 mm/s at 0.2mm layer height. New hypothesis: the ash in the threads is so insulative that I need to run at a much higher extruder temperature in order to get the correct volume flow rate. Trying again at 215 C. Seems to help, cranked it up to 230 C. Looks much better, so that's probably the problem. I'll have to run at ~220-230 C for high flow rates. Not sure what else that will affect. I should probably run another bridging test.
Another lesson learned: Need good metal-metal contact in heat block. Ash/carbon and corrosion are bad.
Update: I experimented some more with speed and temperature. At 220 C, 40 mm/s, 0.25 layer height, the outer surface had a slight glossy appearance. I found out from my previous temperature tests that that was ~200 C (40 mm/s, 0.2 mm layer height) before the nozzle replacement. Upping the speed to 50 mm/s using the speed multiplier under "quick settings" seem to work fine, but the glossy surface finish went away, which corresponded to ~195 C on the previous temperature tests. I started getting under extrusion around 185 C. So the 15-20 C hotter I need to run now is somewhat consistent, i.e. 200 C = under extrusion, 220C = slight glossy-ness. Bumping the speed up to 60 mm/s and temperature to 225 C isn't glossy, but it isn't under extruding either, which is good.
Lesson learned: Extruder temperature and setting temperature are functions of flow rate and thermal conductivity. Higher flow requires a higher temperature set point. Lower thermal conductivity (ex: ash in threads) requires a higher temperature set point. But this only works to a limit.
No comments:
Post a Comment