Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: 3D print Z reinforcement via injected loops

mgunlogson.github.io

Post reply on HN
Print view View original

31–40 of 60 posts

Re: Show HN: 3D print Z reinforcement via injected loops

#31

Post by slabity »

joshvm wrote: post #25
slabity wrote: post #7

I've seen this technique a lot, but mostly as a post-processing technique where resin, fiber, or some other type of plastic is injected into the channels after printing is completed. It would be interesting to see this done during the normal printing process. I am a little skeptical on the technique though. FDM printed walls are known to not handle pressure well, especially during printing when its past its glass-tra…

The simplest option is to print the part raised at an angle so the layer lines aren't parallel to faces. Clough42 has some good videos on support/rib design in Fusion: https://www.youtube.com/watch?v=XXaLxSmtnbQ , based on https://www.youtube.com/watch?v=8NKVNwVaZU0 . But you can definitely get printers to dump a blob of filament out without worrying about cooling problems, if the extruder speed is high enough. I was…

> But you can definitely get printers to dump a blob of filament out without worrying about cooling problems

Yes, but we're not talking about dumping a blob of filament. We're talking about injecting filament into a well-insulated channel where it's physically impossible for it to receive any active cooling whatsoever.

That's not a situation where you can just ignore cooling.

Re: Show HN: 3D print Z reinforcement via injected loops

#32

Post by LordHeini »

I doubt that filling a long thin channel will ever work. The heat + pressure will always collapse the thin infill walls.

In general infills does not provide much strength to a part, it is way better to have stronger walls.

And z-direction does not need to be more stable than the other directions so there is no need for long continuous strands anyway.

Maybe it would work better with smaller, less tall, slots at the inside of the walls.

Lets say 2-3 layer heights tall, continuously filled slots, which are then interleaved with each other. More like bricks less like columns. The outer wall layers would provide stability to prevent collapse. And over spill or bulging would occur towards the inside of the part.

Re: Show HN: 3D print Z reinforcement via injected loops

#34

Post by Toutouxc »

LordHeini wrote: post #32

I doubt that filling a long thin channel will ever work. The heat + pressure will always collapse the thin infill walls. In general infills does not provide much strength to a part, it is way better to have stronger walls. And z-direction does not need to be more stable than the other directions so there is no need for long continuous strands anyway. Maybe it would work better with smaller, less tall, slots at the in…

> z-direction does not need to be more stable than the other directions so there is no need for long continuous strands anyway

I’m not sure I understand. In FDM printing, Z is the only direction where you currently CANNOT have long continuous strands, even if you need them. You always need to sacrifice one direction in which the part is going to suck.

For example, you can easily print an airplane wing with a beautiful, perfectly smooth and continuous airfoil, but you have to include a channel for something like a carbon fiber rod. Without it, the slightest bending force would instantly split the layers apart. Any other orientation will give you a rough surface with steps and a disgusting amount of supports. Being able to add a few strategically placed “columns” (i.e. members in the spanwise direction) could really help this particular usecase.

Re: Show HN: 3D print Z reinforcement via injected loops

#35

Post by Toutouxc »

Aurornis wrote: post #10

Interlocking layers is an interesting idea, but I don't see how this is supposed to work. You can't use the nozzle to inject that much filament into a large cavity because it will cool and solidify right out of the nozzle. Anyone who has ever cleaned blobs of filament off of a nozzle after a print failure can tell you what happens when you try to pump hot filament into empty space. Filament cools below the melt tempe…

It doesn’t have to be a large cavity in order to be useful. Imagine being able to reliably fill a hole that’s 5mm deep. Not amazing, but that could mean 25 layers. That’s 24 layers more than what we can fuse together now.

Re: Show HN: 3D print Z reinforcement via injected loops

#36

Post by LordHeini »

Toutouxc wrote: post #34
LordHeini wrote: post #32

I doubt that filling a long thin channel will ever work. The heat + pressure will always collapse the thin infill walls. In general infills does not provide much strength to a part, it is way better to have stronger walls. And z-direction does not need to be more stable than the other directions so there is no need for long continuous strands anyway. Maybe it would work better with smaller, less tall, slots at the in…

> z-direction does not need to be more stable than the other directions so there is no need for long continuous strands anyway I’m not sure I understand. In FDM printing, Z is the only direction where you currently CANNOT have long continuous strands, even if you need them. You always need to sacrifice one direction in which the part is going to suck. For example, you can easily print an airplane wing with a beautifu…

Adding a long strand of Filament in z direction is what the Author of this article tried. By injecting molten filament into a long channel in the infill.

What i meant that the z-direction does not need to be MORE stable than the other directions.

Adding a long strand of filament in the z-direction in the infill (close to the geometric center of the print) might make the print more resistant to stretching but not necessarily bending.

Carbon rods don't bend but filament does. The wing would break apart at the seams event if it had channels of filament along side its z-axis. It would still be more stable, but not as much as one might think.

Walls give a print most of its strength by a huge margin. And interlocking the walls in z-direction would have a proprietorially larger impact.

Re: Show HN: 3D print Z reinforcement via injected loops

#37

Post by Aurornis »

slabity wrote: post #30
Aurornis wrote: post #10

Interlocking layers is an interesting idea, but I don't see how this is supposed to work. You can't use the nozzle to inject that much filament into a large cavity because it will cool and solidify right out of the nozzle. Anyone who has ever cleaned blobs of filament off of a nozzle after a print failure can tell you what happens when you try to pump hot filament into empty space. Filament cools below the melt tempe…

So I'm pretty skeptical about this as well (see my other comment on this), but the particular failure mode you're discussing is not what's happening in reality. > Anyone who has ever cleaned blobs of filament off of a nozzle after a print failure can tell you what happens when you try to pump hot filament into empty space. Filament cools below the melt temperature quickly, especially when it comes into contact with y…

> That's completely irrelevant because this isn't printing into empty space at all. This is injecting molten plastic into confined channels, with no active cooling, made from material that doesn't conduct heat well. You're saying that the plastic will cool too quickly, but I believe the opposite will be true.

The filament will cool on contact with the part. Think about it: How are you expecting this filament to stay as hot as the inside of your nozzle while it’s flowing through the part and touching all of those walls? How are you expecting the walls to not melt, but the melted filament to flow through them?

Even filament extruded into free space cools enough to become a solid string barely inches out of the extruder. It will cool even faster if it’s touching something.

The only way this works is by heating the entire part up to a temperature where the filament stays hot enough to flow. So you’d need to heat the chamber and the part to 210C to get PLA to flow through it.

Re: Show HN: 3D print Z reinforcement via injected loops

#38

Post by Aurornis »

Toutouxc wrote: post #35
Aurornis wrote: post #10

Interlocking layers is an interesting idea, but I don't see how this is supposed to work. You can't use the nozzle to inject that much filament into a large cavity because it will cool and solidify right out of the nozzle. Anyone who has ever cleaned blobs of filament off of a nozzle after a print failure can tell you what happens when you try to pump hot filament into empty space. Filament cools below the melt tempe…

It doesn’t have to be a large cavity in order to be useful. Imagine being able to reliably fill a hole that’s 5mm deep. Not amazing, but that could mean 25 layers. That’s 24 layers more than what we can fuse together now.

> Imagine being able to reliably fill a hole that’s 5mm deep

You can’t inject filament into a hole without it cooling and solidifying at the top. That’s the problem with this whole idea.

Re: Show HN: 3D print Z reinforcement via injected loops

#39

Post by Aurornis »

slabity wrote: post #31
joshvm wrote: post #25

Earlier quoted context omitted.

The simplest option is to print the part raised at an angle so the layer lines aren't parallel to faces. Clough42 has some good videos on support/rib design in Fusion: https://www.youtube.com/watch?v=XXaLxSmtnbQ , based on https://www.youtube.com/watch?v=8NKVNwVaZU0 . But you can definitely get printers to dump a blob of filament out without worrying about cooling problems, if the extruder speed is high enough. I was…

> But you can definitely get printers to dump a blob of filament out without worrying about cooling problems Yes, but we're not talking about dumping a blob of filament. We're talking about injecting filament into a well-insulated channel where it's physically impossible for it to receive any active cooling whatsoever. That's not a situation where you can just ignore cooling.

> We're talking about injecting filament into a well-insulated channel where it's physically impossible for it to receive any active cooling whatsoever.

Look up the thermal conductivity of air.

Then look up the thermal conductivity of 3D printing filaments which form those channels that are being injected into.

The filament will be cooling faster in the channel than in free air.

This cannot work unless the part is heated to a temperature where the filament flows.

Re: Show HN: 3D print Z reinforcement via injected loops

#40

Post by amluto »

Aurornis wrote: post #39
slabity wrote: post #31

Earlier quoted context omitted.

> But you can definitely get printers to dump a blob of filament out without worrying about cooling problems Yes, but we're not talking about dumping a blob of filament. We're talking about injecting filament into a well-insulated channel where it's physically impossible for it to receive any active cooling whatsoever. That's not a situation where you can just ignore cooling.

> We're talking about injecting filament into a well-insulated channel where it's physically impossible for it to receive any active cooling whatsoever. Look up the thermal conductivity of air. Then look up the thermal conductivity of 3D printing filaments which form those channels that are being injected into. The filament will be cooling faster in the channel than in free air. This cannot work unless the part is he…

I can imagine it working with a needle shaped nozzle that inserts into a hole and extrudes filament as it withdraws back out. This is probably much more than a software change, though.
Post reply on HN