New type of 3D printing uses sound waves to build up objects | Hacker News Reader

New type of 3D printing uses sound waves to build up objects

newatlas.com

Neat. It's an old process being used in a new way. Ultrasonic soundwaves are already used to melt plastic, and is used for ultrasonic welding of plastics. It works really well for certain use cases.

I had expected them to use something like a phased array to focus and "scan" the soundwaves electronically, but that doesn't seem to be what they're after.

Overall this method can work with different materials from the usual UV-curable resin stuff.

Pet peeve:

> the temperature inside them rises to about 15,000 degrees Kelvin (14,727 °C or 26,540 °F) and the pressure within them climbs to over 1,000 bar (14,504 psi)

Those conversions are way too precise. The initial values have 1-2 digits of precision; the converted numbers have 4-5. Should be:

> the temperature inside them rises to about 15,000 degrees Kelvin (14,700 ºC or 26,000 ºF) and the pressure within them climbs to over 1,000 bar (14,500 psi).

and even those are probably too precise.

I had a flash of insight a few years ago that the real way to think about 3d printing. We can separate "data transfer" about desired solid regions in a volume, from what I call "locking".

Once I made this transition in thinking, I saw hundreds of new ways to do 3D printing. Ultrasound was only one of a huge multitude of other "data transfer" techniques that just kind of felt obvious all of a sudden.

Have a look at this: https://opg.optica.org/ol/abstract.cfm?uri=ol-47-5-1279

Matches 100% what you're saying: use low intensity uv to write the spatial area you want to solidify, and when you're done flood cure it all together.

So it sounds like the novel feature is that it can print something inside of something else, but the article doesn’t say what’s required to make that happen: if I’m going to repair a plane fuselage or some surgical implant: how do I get the material inside to print? Does the fuselage require the material inside already? How do we inject the material inside of a patient for an implant?

Maybe you don't see the graphic but in relation to your last question, they're literally using a syringe to do it.

I see. So you can inject some material through a very small opening and print the full size piece in place.

Multiphoton is the only true form of printing inside of a substance

The mechanism reminds me of sonoluminescence

https://en.wikipedia.org/wiki/Sonoluminescence

bit off topic but could it possible to create "virtual objects" that can be moved around, provide/detect tactility? I know I've been obsessing with this since I read someones comment on here but you would essentially use sound waves to "conjure up" objects in a room that has ultrasound devices on walls, ceilings, floor etc.

https://www.youtube.com/watch?v=jOnxr9Ez_Kc

When interacting with a real (solid) object, you only ever interact with its surface. When you touch a ball, you feel its surface. If you press into the surface, you feel that resistance as increased pressure (near-instantaneously) and you gain a lot of information about he object by that sensation (consider the difference between poking a balloon vs. an anvil) and by the reaction of the object. With sonic "virtual objects", there would be no resistance, meaning you couldn't really interact with the object in a normal sense (your hand would just pass into/through it). Even if the "user" were to understand this, it seems like the range of experience would be very limited. Without any sensations associated with resistance, could you make a balloon feel any different than an anvil?

I've spent some time thinking about this in the past, so please forgive the random tangent: IMO, the creation of virtual objects that can be interacted with and perceived as real objects will require a combination of haptic garments and exoskeletons. The haptic garments would manage any minor sensations (the fuzzy texture of a kiwi), while the exoskeleton would apply force / prevent motion. That way, if you try to poke a virtual balloon, you feel a minute jolt of resistance at the moment of contact. If you poke an anvil, you feel your finger come to a dead stop, with all of that force on your fingertip conveying the sensation of poking a heavy, stationary object.

It seems like a suit could simulate a lot of object interactions. Imagine hefting a virtual stone in your hand. You feel its texture on the palm of your hand via the haptic glove, while the exo-suit conveys the weight as downward force, applied at the wrist, elbow, and shoulder joints.

The virtual object could be relative to the person interacting with it rather than a stationary zone in space. With sound waves, for example, it could be something like a tightly focused beam combined with motion/finger/hand tracking... As your hand moves along the "object", the beam is constantly re-aimed (at the hand) and re-adjusted in real time to give you the sensation of feeling around the profile. No suit required.

Creating tactile focal points in a specific 3D coordinate seems possible if we use the transducers described in this video. You can see that they are able to "place" tactility in specific 3D location.

https://www.youtube.com/watch?v=2QkbVr4J7CM

All that would be needed here would be an AR solution that projects an opaque hologram of a 3D object in the same location where tactility is being placed in mid air.

What stops us from increasing the power of the transducers as to magnify the torque, pressure? Like parent's comment: anvil vs basketball.

What is a safe and realistic range of pressure for anvil? Could such virtual object injure the user if dropped?

I agree a suit is the more immediately realizable tech but something that isn't so clunky like what Mr. Zuckerburg uses here.

https://www.youtube.com/shorts/oQig6L2BtOo

There must be a way to provide torque and tension electronically in a suit. Millions of small points all over.

Localized linear pcb motors that push/pull on threads built into the suit

I wonder if an alternate, not-as-good-but-acceptable set of conventions would arise. The best weighted MIDI keyboard doesn’t have the responsiveness of a good grand piano, but great music can still be made with them. Same with electric vs acoustic violins.

It definitely can be - I had a demo of Ultrahaptics' (now ultraleap) ultrasonic tactile system[0] 3 years back and you could definitely "touch" objects that weren't there. They only had the development kit out at the time - I now want to check out their new module to see how it's progressed!

[0] - https://www.ultraleap.com/product/stratos-inspire/

Pleasantly surprised they are still around since the whole MagicLeap fiasco

Which fiasco are you thinking of? Or could you be confusing Magic Leap[0] (the Augmented Reality headset) and Leap Motion[1] (the 3D motion controller) (admittedly I swap the names around all the time!)

[0] - https://www.magicleap.com/en-us/

[1] - https://www.ultraleap.com/product/leap-motion-controller/

Ultrasound transducer arrays can be used to create tactile sensations (haptics) using a fields projected in 3d. A recent Kickstarter for this type of device is Emerge[1], and a more mature company is Ultrahaptics[2]. Have not tried either though, so do not know how well it works in practice. Ultrasound can also be used for gesture sensing in 3d. Elliptic Labs is one of the companies in that space. mmwave radar is also an option here, along with optical methods.

1. https://www.kickstarter.com/projects/emergehome/emerge-home-...

2. https://www.ultraleap.com/haptics/

Once the resin solidifies, won't that impact the sound waves' ability to pass through the solidified region?

It probably goes furthest to nearest so that scenario never becomes an issue.

Is this the same principle as ultrasound scans? And microwave ovens?

What if there’s a print error, how do you extract the material?

Surgery

I imagine a solvent could be injected in that liquifies the 3d print so it could be sucked out. Maybe they've already thought of this.

Doesn't seem like there's going to be many solvents that are safe to inject. Water, but if it's water soluble, you have other issues.

Reply on news.ycombinator.com