Hacker News Reader

Comment by minihat | Hacker News Reader

Not a secret to the research community... Let me balance the hype with some of the remaining challenges.

Depth is a problem; the deeper you go, the less effective it is. Heat is another issue; you can inadvertently damage nearby tissue. Targeting accuracy is vital, especially near critical structures (think nerves), and we're not at sub-millimeter precision yet. Also, real-time monitoring like MRI-guided FUS is expensive and complicated, and without it you have to guess that you're affecting the right tissue. Great promise, but multiple engineering hurdles to clear before FUS lives up to the hype.

View on HN

> and we're not at sub-millimeter precision yet

How do you get sub-millimeter precision with ultrasound? At 300KHz, the sound has a wavelength of 1mm. My understanding is that sound with a frequency greater than about 150Khz dissipates after passing through 5cms of air; it must dissipate faster in flesh.

Also, a wavelength of 1mm should give you a resolution of about 4mm, right?

I wish I knew more about the propagation of ultrasound; I'm getting interested in the bats that live around here.

Isn't this like saying how can we move something one millimeter if our fingers are bigger than one millimeter?

Yes, I think it is like saying that, but I also think that OP was considering targets less than 4mm in size. The focus of the FUS wave being no smaller than 4mm, very small targets would be larger than the smallest focus.

It's like saying "how can a syringe extract the cytoplasm of a cell?" if you take OP's angle.

Ah I guess in the case of ultrasound, the thing you're trying to do is not nudge or move something, you're literally trying to "ablate" or hit it and only it. That's more like pressing a tiny button with your finger and trying not to touch the outside of the button.

More like saying “how can we draw a 1mm dot with a pencil bigger than 1mm”.

Though not a perfect analogy as you can do stuff with interference patterns to get sub-single wavelength resolution IIUC, I think this is used in silicon wafer photolithography.

> My understanding is that sound with a frequency greater than about 150Khz dissipates after passing through 5cms of air; it must dissipate faster in flesh.

Flesh is basically water. Water transmits sound extremely well.

Ultrasonic imaging typically uses frequencies in the low MHz. Like 1-10 MHz.

> Ultrasonic imaging typically uses frequencies in the low MHz. Like 1-10 MHz.

Oh, thank you! So the dissipation is low (because it's water), and the theoretical resolution is 500x greater than my ignorant estimate. I now consider myself better-informed.

[Edit] Now I guess I'm off to see if I can find out what a MHz-grade ultrasound transducer looks like...

All I know is they charge a lot for them. Ultrasound imaging at home would be amazing, but the transducers are hella expensive.

You can actually get quite cheap ones now. They make integrated scanners that work with a phone app. Easily under £1k second hand. Maybe close to that new if you shop around.

I wouldn't recommend it though. The only reason we use ultrasound imaging is because it's cheap, easy and completely harmless. As an actual imaging method it's terrible. There's so much speckle you can barely see anything, except in some situations like pregnancy where you have a convenient bag of water around the thing you're looking it.

There's a reason ultrasonographers are well paid - it's really really hard to read an ultrasound.

> At 300KHz, the sound has a wavelength of 1mm

And that’s in air. In water (which is more like what most bodily tissues are made of) the speed of sound is almost five times higher so the wavelength is five times longer.

I know this is maybe completely out there, but could focused ultrasound be used to heat up food? The current issue with microwaves is that they heat up food very unevenly. I am sure the current cost is quite prohibitive and all. I'm more interested in the science part itself.

Seems like the challenge would be heating the right places. How would you know what food to heat more or less?

Also, ultrasonic transducers today, as far as I know, require essentially direct contact to impart energy. You probably don't want to dip something into your leftovers just to heat it.

Rather than heating food more or less, I care about equally. Couldn’t combination of thermal camera and ultrasonic transducer achieve the goal?

If they indeed need a contact with the object, then obviously that wouldn’t work. But I am not sure why the air molecules couldn’t carry it.

Impedance mismatch and attenuation. The impedance difference between air and the piezoelectric elements of the transducer is too great - this will cause soundwaves to reflect on contact (and thus never leave the element). If some transducer innovation changed that variable, then attenuation would be the next hurdle.

In general a form of couplant (water, oil, some form of gel, etc) is necessary to eliminate air between the transducer and a material in order for soundwaves to pass.

Thanks for the context. That seems to invalidate the idea.

it feels like (as an uneducated outsider) that a microwave could use some kind of spatial awareness (lidar/sonar/whatever) along with beam-forming and array style transmitters to evenly cook something while sampling temperatures remotely.

the size/shape/cost/performance of such a device is left as an exercise to those better suited to execute that kind of artistry than I am.

I keep thinking about this topic for a couple of years now (hence my question). I wonder if combination of new technology could be made into much more useful microwave, even if the cost was not suitable for personal use (at least initially).

The dream is a microwave with "single button" (start/stop) and a temperature scale (lukewarm, warm, hot, boiling) that consistently heats up food correctly and doing so within very short period of time (below 3 min).

There is work done one this. search for: software defined microwave (SDC)

Super cool. Thanks for the suggestion found this [0] but there appears to be more out there.

[0] https://www.hcii.cmu.edu/news/software-defined-cooking-using...

Solid state microwave ovens (that avoid standing waves/dead zones) do exist, they are just very expensive.

Dumping kWs of ultrasound into food would be a bit tricky.

Thanks for putting me on this topic. Will do much more research.

I think an air fryer or toaster oven solves 80% of your problem using off the shelf tools.

The issue is time. They absolutely work better than microwaves, but microwaves are popular precisely because they are just much faster.

Does slapping a chicken count?

https://m.youtube.com/watch?v=LHFhnnTWMgI

I've got a cyclonic inverter microwave, and maybe it's all just hype, but I'm pretty sure it cooks more evenly. I also usually choose a lower power level, which helps a lot. The power level button is highly underrated.

I am not expert, but my understanding of cyclonic inverters is that they are able to lower power output and thus achieving more even cooking. That also implies slower time. I am more interested in precision while retaining the speed. But it's something I looked into myself too.

Does anyone know how this works? They demo it cooking a steak without melting the beeswax it is wrapped in, or one half of a salmon filet:

https://www.miele.com/brand/en/revolutionary-excellence-3868...

I'll speculate this is phase-controlled microwave cooking.

Microwave power transistors are now cheap at high power levels. If you have an array of such emitters, you can synthesize beams that can deposit energy in small, local volumes. This has been done in actively-scanned military radar for several decades. No moving parts, and beam steering in microseconds.

As a bonus, you could also run a pre-cooking process where you scan the food and map its microwave reflection/absorption properties. This would let you calculate the cooking algorithm based on the particular food item..

Thanks for the commentary.

My immediate reaction was I'm sure there have to be challenges and drawbacks. But hopefully we'll see the tech progress.

> and we're not at sub-millimeter precision yet

... out of curiosity, why not?

Also, do you need MRI-guided or can one just use a much higher frequency ultrasound to see where it's targeting? Regardless, seems like a classic RL problem (presumably some sort of ~ms level cycle of low-power targeting then short burst of power).

Reply on news.ycombinator.com