The last time Hackerfall tried to access this page, it returned a not found error. A cached version of the page is below, or clickhereto continue anyway

Independent/Scientist/Entrepreneur : Ultrasound, thermodynamics, and robot overlords.

Ultrasound, thermodynamics, and robot overlords.

As some of you may know, I play with ultrasound for a living, and I study what certain modalities do to specific kinds of cells. A collaborator of mine wrote a blog post questioning the wisdom of investing $10M in uBeam, and wow, did it go beyond what either of us expected. It was blowing up twitter, tumblr, Y Combinator’s Hacker News picked it up, and it even made it onto Business Insider. It’s amazing how one small social critique can reach so far when it touches on such an apparently sensitive subject.

It’s dismaying how the resultant conversation has – for the most part – split into two camps, the uBeam is Great No Matter What And All Dissenters Are Idiots camp and the Those Rich Fuckers Have It Coming; Take Them Down camp. That’s indicative of yet another trend, the one where every issue becomes a deeply personal, almost religious, cause.

I’m not on either side of this. Why? Two main reasons:

- Meredith Perry should try whatever she wants to try.

- That said, she needs to acknowledge the parameters in which she needs to work.

Let’s dive in and pick this apart.

Meredith Perry should try whatever she wants to try.

I firmly believe that if someone has a hunch that something might be possible, they should try it. Done. I will not change my mind on that. Even if others know it isn’t possible, even if people have done and redone whatever it is, the person should absolutely be empowered to find a way to get their hands dirty and dive in.


Because school is currently crap for developing the explorer/scientist/artist/entrepreneur/boundary pushing tendencies we all have to varying degrees in assorted fields. We’re doing things like teaching chemistry the exact same way it was taught in like 1950 and that tech class is somehow less valuable than math class, and that is complete bullshit. We’re exclusively taught in school to do other people’s work, to be a cog in a larger system, that there are right answers, that we should study to take the test (not learn the material), and that mistakes and being wrong are bad. This is not to say that people who enjoy working for larger organizations are somehow less than those of us who strike out on our own; I’m saying that for those of us who harbor strong out-of-the-box tendencies and the desire to learn with our hands, school can be torture.

Along similar lines, I resent how Perry’s educational pedigree is held up as an absolute metric of whether she was qualified for uBeam in the first place. Regardless of what’s on her resume, none of us have any idea about what she’s done with the time she was given in her education and accessory programs unless we’ve sat down and picked her brains. We all have at least an inkling that this world is not a meritocracy. You can study for the test, or you can learn, and to be honest, learning seems to happen more outside of the classroom. I don’t know what she’s done. I don’t know how she copes with stress, how she reacts to questions she doesn’t know the answers to, or what her team looks/will look like. I don’t care about her GPA; I’m more interested in whether she’s the kind of person who understands what she knows and doesn’t know, and surrounds herself with people possessing complementary skill sets who are in varying ways smarter than she is, or is she the kind of person who needs constant validation and reacts to the insecurity by only working with people with whom she can feel superior?

None of us know the answer to that question, except – hopefully – uBeam’s investors.

That said, she needs to acknowledge the parameters in which she needs to work.

Let’s settle this once and for all; is it physically possible to charge a phone with sound waves?


This is an energy transfer; you’re depositing ultrasound onto a device that can convert the mechanical energy of moving air molecules into electrical energy. And you could do it, even through the air, without vibrating the phone to pieces. (I say “even through the air” because air is a shit conducting medium for ultrasound, which is why they smear that gunk on pregnant folks’ bellies for imaging.) So, in the stripped-down terms of energy transfer, this is physically possible. No one will disagree.

Before I launch into some serious sciencing, I have to address the demo. It’s been dissected by people on Hacker News, and as was pointed out there, it’s missing half the information. The only parameter Perry mentions is the voltage. Electrical power (P), typically measured in Watts (W) is equal to the voltage (V) times the current (I). Without a current measurement, there is no way to understand how much power is being transmitted, and Watts are the fundamental way in which electrical power is measured. Depending upon how much of your high school physics you remember, Watts are Joules per second, Joules being the unit of energy required to move a Coulomb of charge through 1 Volt of potential. We need to know how many Coulombs of charge are moving through Volts of potential per second (which is where Amperes of current come in; they’re Coulombs per second) in order to understand exactly what the power output is. Your standard Apple iPhone charger delivers 5 W. Telling us the voltage? Not that meaningful because I can’t compare it to the device I use. I wouldn’t have anything to say about this if there was a demo that showed the whole thing. But there isn’t, and so I’m going to.  Show us one that demonstrates all of these things beyond a reasonable doubt, and I will be on board with uBeam’s effort.  

What happens when you add living organisms to the equation? You know, the users and their pets? You leave the relatively clear and mathematical world of pure physics/engineering and end up in my field, exploring the effects of ultrasound on the unpredictably murky corner of biology, physics, and chemistry. Your federal health and safety agencies set the limits on the maximal safe levels of sonic exposure for humans, because bad things start to happen to biological tissue when exposed to high amounts of energy, sound included.

First of all, there is a sonic spectrum. Sound waves are pressure waves, or oscillating mechanical compressions of the matter through which they travel. For sonic waves to radiate, they need matter to travel through, which is why sound can’t travel through a vacuum (and why spaceships don’t actually make noise). We have highly evolved organs, our ears and brain, that can detect and process sound waves in a narrow portion of the full sonic spectrum. Some creatures (bats and cetaceans like whales, dolphins, etc.) can use ultrasonic frequencies for echolocation. Cats and dogs and other small mammals can hear in the ultrasonic range, too. We use certain frequencies and strengths for medical imaging and sonar. Other modalities, like high intensity focused ultrasound (HIFU), are used to burn away carefully selected portions of tissue without having to cut the body open. Still other frequencies and intensities of ultrasound are used to mechanically clean tools and jewelry, cause temporary male infertility, and speed up bone fracture healing. Essentially, from surgery sans scalpel (HIFU) to manipulating genetic expression (fracture healing), this simple mechanical stimulus of propagating pressure waves interacts with biology in intensely interesting myriad ways. At Sonify, for example, we’re using ultrasound to treat early-stage melanoma.

If we take this chunk of knowledge and relate it back to uBeam, uBeam’s existing patents mention using ultrasound in the 20-120 kHz range, at a strength up to 155 dB SIL, which is the maximum amount of ultrasound to which people are legally able to be exposed.

Problem # 1.

Cats and dogs can hear up to 64 kHz and 44 kHz, respectively, and rats and mice can hear up to 76 kHz and 91 kHz respectively. PETA won’t be thrilled if the frequencies used are in that range.

(Potential fix: Go above those frequencies or, er… manufacture pet earwear?)

Problem #2.

I know from firsthand experience that at 27 kHz and 0.3 W/cm2 or 154.7 dB SIL according to the units converter found here, human keratinocytes become damaged in a way that’s incompatible with life in a matter of minutes due to mechanical stress.

You should care about this because keratinocytes are the most plentiful type of cell in your skin, and they produce keratin, a structural protein that helps hold all your insides together. They’re strong little bastards, and if they’re hurt by ultrasound, you can bet that less structurally focused cells will suffer too. Mechanical damage… think of a scrape, just larger. Like large swaths of your skin. I don’t want to even contemplate eyeball damage.

(Potential fix: Um… one of these for everyone?)

Problem #3.

Basing this on my own lab experience, I know that for my cells to receive that much ultrasound, the transducer was so loud, I could only use it after normal business hours, with a 24 hour advance email warning to everyone on the floor, putting signs on the lab doors, making foam earplugs available to everyone in my lab area, and making sure that no one was doing small animal work at the times when it would be on (see Problem #1).

(Potential fix: MHz ultrasound transducers quiet down, but they affect cells in all kinds of incredibly interesting ways, like acting as a temporary male contraceptive by temporarily knocking out spermatogenesis.  Totally not joking here; this was a big deal.)

Problem #4.

What Danny calculated; the fact that even if uBeam’s people start at the absolute upper limits of federally allowable ultrasonic energy, outside the audible range of small mammals, AND we assume that those upper frequencies somehow magically don’t utterly destroy biological tissue (which I strongly, strongly doubt, but best case scenario and all that) charging a phone via ultrasound will take at least 100 times longer to charge than if you plugged it into a wall. That maximum? 155 dB SPL? Regardless of how you reach it, whether by focusing an array of transducers or not, that is the fundamental constraint of this problem. We’re talking about a basic energy transfer, from sonic to electrical, and the rate at which that can be done in terms of the limitations set by the human body.

(Potential fix: Break literally all of thermodynamics. I will happily accept my technological overlords and joyously look forward to perpetual motion machines and alternate explanations for quantum physics should this happen, but considering all the exceptions that exist for every other set of physical laws, thermodynamics has exactly zero exceptions, so I’m inclined to believe the absolute truth of it. What’s that? You don’t understand thermodynamics? Pick up a copy of McQuarrie, and learn some serious shit about how the universe works.  I hate/love that book.)

(This dude could probably could break the laws of thermo.)

Problem #5.

Related to Problem #3 and Problem #4, the conversation about this has been peppered with people throwing around words the meanings of which I’m not sure they understand, like “nonlinear.”

Let’s clear this up. If you’re in a room with a focused ultrasound beam, and that beam is strong enough, the medium through which it’s traveling (air in uBeam’s case) becomes nonlinear, the oscillations of which causes beams of widely ranging frequencies to be generated, and because they have different wave propagation speeds due to dispersion, they don’t focus, and instead propagate in all different directions. Air becomes a dispersive medium at frequencies greater than 28 kHz thanks to the presence of CO2. These propagating beams are hard to predict and control, and given that they’re going to be at lower frequencies than the original beam (because physics), they will probably drop into the audible human range (upper bound is around 23 kHz) and will be at the very least annoying as hell. Factors that influence this phenomenon is air temperature, pressure, altitude, CO2 concentration… controlling this is well-nigh impossible.

(Potential fix: You got me. No breathing?)

(Or spacesuits.  But then, farts.)

So, look; you can’t build a charge-anywhere phone charger without your users, and your users are people, so you have to abide by the limitations humans and their pets pose. As such, your energy transfer rate has an upper limit, and it’s really inefficient and slow. I kind of wonder what is the energy loss from an efficiency standpoint when comparing ultrasound charging to plugging the phone into the wall.

Personally, I see a rock-solid project here, a wonderful opportunity to learn, but it’s a student project. Sonify completed proof-of-concept work on a melanoma treatment that required biosafety level II (bsl-II) facilities, built previously nonexistent lab equipment, generated results that outperformed commonly used chemotherapeutics, and submitted three patents in a year with less than $80k. $10M for an undergrad or bare bones master’s level project… I don’t understand it. I’m always jealous of how in engineering, you can sit down and crunch the numbers to get a 99.9% accurate prediction of how your device will perform. As such, I’m not sure why the uBeam people don’t have a meaningful rebuttal to “a random person on tumblr,” to paraphrase Marc Andreesson, and are instead blowing the whole thing off. It makes them look like they’ve been caught with their pants down.

Thermodynamics is not subjective, it’s not magic, it exists whether you believe in it or not, there are no exceptions to it, and it’s not personal. It doesn’t matter whose hands this project is in, the same parameters exist. I find it very odd that the reaction is to reiterate that it’s not about the ways in which it won’t work but rather it’s about the ways that it will work, and yet there’s an actual calculation that counterindicates its success that no one from uBeam is directly addressing.

So what?

I don’t know. It depresses me to see what could have been an honest simple project blown into an artificially loud PR campaign that looks like it can’t answer a high school physics question, even with all its patents filed. It doesn’t make me happy to think of uBeam failing. The whole thing is sad and generally shitty, not least of all because if uBeam isn’t much beyond an elaborate PR job, it’s just going to make it harder for the public to trust science and for investors to trust other ventures. So, Meredith Perry, you better pull it out of the bag and your product better work, because we sure as hell don’t need more shit stacked against Sonify and the rest of the innovators in and out of the ultrasound community if this field becomes associated with a Silicon Valley investment boondoggle.

Continue reading on