I have a two cylinders that are stacked on top of each other with no space in-between. The bottom one prints from the left nozzle (external spool) and the top one prints from the right nozzle (AMS) on the H2D. On this transition between them, it’s slightly indented like the top one has been chamfered or filleted slightly. I was doing a scarf seam on this model which I thought might be causing the glitch but it still occurred after turning scarf seams off. Could this be some sort of z offset issue between the two nozzles? or something else? I’ve also include a screen capture of the layers before and after the transition in the slicer and nothing looks out of the ordinary.
Oh… I just realized something. I had turned off the prime tower because this print has zero filament changes because it’s using just two colors. Do you think that might be causing it?
Ok, I’ve learned something. That problem was cause by not having a prime tower. Here’s the transition from one color to another with the prime tower. It was perfect. Disregard glitch before the color change. That was some sort of clog.
It triples the amount of filament needed to print the model for that tiny defect. I think I’ll probably leave the prime tower turned off in this case. It sure seems like there must be a way to do at close to the same thing using the purge chute. Maybe some sort of compromise that does close to what a prime tower does.
And here is the FAQ for why do you need to turn on the prime tower when dual color printing on the H2D.
Is the thing solid or hollow? If it’s hollow inside, add some infill and flush to the infill to maybe get the new nozzle “reprimed” in an area that wouldn’t be visible. And just do infill first before the walls.
That’s a good idea! Unfortunately it’s hollow (a whistle) so there’s no infill. But it does have inner walls. I’ll try changing the order to do inside walls first and see if that moves the problem to inside the walls.
You might also be able to put a “sacrificial” object on the build plate. Like a prime tower only smaller so it doesn’t use as much filament. Like, a 4mm square post that’s got solid infill. Though I’m not sure how to guarantee the slicer picks the “right” order to print that layer…
I adjusted the number of walls from 3 to 2 and that ended up changing the order it prints the walls. So now it prints enough of the inside walls to “prime” the filament change inside the model before it starts printing the outside wall. The defect is gone and I’m also not using a prime tower.
The issue has changed slightly. Here’s some more investigation into printing my tin whistle on the H2D in one piece. This video shows it near the end of the print at about 275 mm off the print bed where it switches from printing with the external (left) nozzle to the AMS (right) nozzle. It’s recorded at 240 frames a second so all of this happens quickly.
It heats up the nozzle, gives it a wipe and then hesitates before it starts printing. While it sits there, a tiny bit oozes out and solidifies somewhat. It drags that piece of ooze over and deposits it on the top of the whistle. It then goes on to print the inside walls and eventually comes back around and hits the deposited ooze and the whole thing breaks off from the print bed. It’s not definitive that hitting the ooze is what causes it but I think that is the reason. This is the third one I’ve printed and they all fail at this same spot. It doesn’t seem to have a problem when changing from the right nozzle to the left one.
I also paused the print on the layer before this when it was printing with the right nozzle and there was no hesitation after it did the wipe before it started printing again.
I’m still not using a purge tower which maybe would fix this problem. Getting rid of the hesitation before it starts printing might help things but I don’t think I have control over that.
I understand printing something so tall might not be realistic. I also tried printing it in “Silent” mode and it does print the Silk filament nicely with a smooth surface but it still breaks off of the print bed.
Ok, the next thing I’m trying is creating a 2 wall, 1 layer thick ring that is just hanging out in space at 275 mm above the print bed at the level of the transition. It’ll get printed first and maybe that’ll get rid of any ooze and get things primed.
That orphan circle may hang from the nozzle and get dragged into the main print. Or not. Sometimes strands will drop away under their own weight but might put you back into the missing filament scenario like from the ooze?
As predicted, it looks like the circle was dragged onto the main print. It made a different kind-of blemish but at least it didn’t get knocked over this time. I was watching the print on the camera and when I saw it was switching nozzles, I ran down stairs. Unfortunately it was too quick and had already placed the dangling filament so I don’t know exactly what happened. It doesn’t look like enough filament for how big the loop was. I’ll probably experiment a little more to see if it’s repeatable.
Molten filament is kind of wild in how it behaves. As the molten plastic comes out of the extruder, the outside of the extrusion cools faster than the inside. In effect it makes a straw that can also act a little like a balloon.
The molten filament expands the straw and the diameter gets larger so it doesn’t get quite as long as the extrusion would be if laid down normally. How long vs how thick just depends on how the viscosity changes as the filament cools.
I see this a lot with silk PLA. When the extruder is purging into the poop chute, instead of a rat’s nest of fine filament like with regular PLA, silk leaves short thick almost teardrops.
And it’s also how extruders can fill up and encase themselves in extruded plastic especially with filaments like PETG HF. That gets curly and sticky if it has too much moisture in it and a straw can curl and twist up until it hits the extruder. Then the straw just lets the molten filament pump up into it.
Might I suggest you give thought to adding a small blip onto the couple of layers before the join. This blip would be a small slim fillet at 30 degrees to avoid support and stick out maybe 3mm at uppermost. that way you should get a clean change. The fillet could then be trimmed off, admittingly leaving a small scar.
This may not be noticeable if small yet practicable. Likewise the same could be done at the start of the upper tube (upturned)
For those who didn’t watch the interview with Dr Tao, the need for the tower may disappear, as they believe this is a fixable software issue.
Not a bug, but rather a complicated set of algorithms that cannot then have knock-ons for the existing ones.
The same is true for the current limitation of not being able to use two different nozzles.
The solution will require complex algorithms, as it can easily fail given that everything previously has been based on single nozzles. This is not exactly true, but sufficiently true to explain the level of complexity involved.
I’m looking forward to those fixes. Maybe I shouldn’t publish any print profiles that are challenging for the H2D right now until they have a chance to refine it a bit.
