I once lit a large fluorescent tube on the field generated by the antenna of a fairly powerful transmitter so that joke has some basis in physics. But you'd be hard pressed to run your vacuum cleaner that way :)
Nokia had a project to charge mobile phones from ambient radio waves. But they had real engineers and scientists and even though it actually worked they reported how badly it worked. That is, the charging rate was so low that only the simplest mobile could be charged with it. I think they used a modified Nokia 100 or something similar. See this 'article' on Gizmodo: http://gizmodo.com/5285565/nokia-developing-wireless-accesso...
"But even amid a tidal wave of publicity, the company has never publicly demonstrated a fully functioning prototype of its system. Nor has it ever produced an outside expert (who wasn’t an investor) who could attest to its ability"
Well, either it works or it doesn't. What's there to talk about? It's not that they're doing rocket science.
> "But even amid a tidal wave of publicity, the company has never publicly demonstrated a fully functioning prototype of its system."
This is what I don't get - how the hell did they manage to get investors? I mean, ultrasound charging isn't magic, anyone who knows their stuff around EE should be able to throw together in a month a simple prototype of the tech to show to the investors. That is, a prototype that actually charges a phone, not this[0] which a) doesn't really charge the phone, and b) seems to be geared towards confusing people clueless about electricity. The fact that they didn't, after 4 years, produce anything suggests that they don't even have anyone on board who knows this stuff, so why are they getting money?
Seriously, I think I'm gonna plug my IRC bot to TTS and go raise money for my AGI startup.
Wait, why is she using a volt meter in that video? That's not useful information without knowing how many amps the system can push. I could probably get a volt meter to read 8 volts just from my hair after walking on carpet a bit, but that's certainly not going to charge my phone...
More specifically, THIS voltmeter, which is a little bit "under-spec'ed", to say the least: Radio Shack (22-223) https://www.google.de/search?q=radio-shack+22-223&tbm=isch It's set to the 10V-AC range. I'd expect a VC backed company to invest in a little more elaborate measurement equipment than that.
On the other hand, given a a good RMS meter operated in a frequency range where it's spec'ed to work and a knowncircuitimpedance will give you a very good measurement of the power: P=Vrms²/R. And analog meters make for rather pretty demos, to be honest.
Actually, that makes it more believable, not less. Let me explain why: A voltmeter can't just measure volts, it has to make a tiny current flow through a resistor. The value of that resistor will be lower for cheaper volt meters because the cheaper volt meters need more energy extracted from the circuit under test to function. A more expensive one has a very high internal resistance, and so loads the circuit under test far less.
In this case being able to supply 8 volts into a meter with a low internal resistance is actually a plus ;)
That said, the current is still tiny, even a crappy volt meter will pull no more than a few hundred micro-amps from the circuit that it is measuring. So let's be generous and say that it pulls a whopping 10 mA (very likely on the high side (edit: it turned out to be 37 uA)), in that case the power generated would be 0.08 W. Not enough to charge your phone by a long shot, but enough for a pretty demo with a moving needle.
Finally, these cheap meters tend to output complete nonsense values when presented with high frequency AC, one would expect that they would at a minimum measure power transfered using a dummy load and some kind of dedicated power meter.
All in all this does not make things look any better for uBeam. I'd really like to know who did their technical due diligence.
What's really bad to me - but I'm a nitpicker - is that in the video the lady refers to the volt meter as a power meter, which it definitely is not. It measures potential, not power, any power that it takes from the circuit can be chalked up to imperfect design and the laws of physics. Maybe the person doing the demo was not knowledgeable about the physics behind the equipment but that's betraying a complete mis-understanding of the basic concepts involved.
The transducers used appear to be of the piezo-electric type, those have a fairly high internal resistance and the voltages produced by those devices would be relatively high, the currents correspondingly tiny.
What gets me about this whole saga is that uBeam makes it seem as if they invented wireless power transfer using ultra-sonics, but any sound is wireless power transfer and that includes ultra-sound. The whole crux is how much power can be generated and whether or not there is a practical way to overcome all the obstacles.
To transfer some power doesn't prove a thing, when I whisper in your ear I'm transferring power using soundwaves...
In that case we might hope for an accidental slip or a deliberate simplification for the audience but she certainly should know better than to call that thing a power meter.
That's the point. It's likely that at the time, she honestly didn't know the difference (or didn't understand the difference well enough to realize that it would be obvious to other people). She did a Ted Talk where she bragged about how she didn't know anything about engineering when she started uBeam.
As I had remarked, a good RMS meter on a known impedance, on a frequency it's spec'ed for. The setup shown is neither.
Also: For low-level AC measurements you invariably need active rectification. So your meter will never be directly connected (or through a series resistor) to the source.
The guts of that meter register 37uA for full scale (I found some specs), at 10V that's an internal resistance of 270K so probably it really is driving the meter directly.
For 8 V that translates into 0.25 mW of power produced, which I don't think will be charging any cell-phones.
> The transducers used appear to be of the piezo-electric type, those have a fairly high internal resistance and the voltages produced by those devices would be relatively high, the currents correspondingly tiny.
Exactly. These transducers look just like low-cost piezos used for ultrasonic ranging in air (e.g. car parking sensors, alarms, etc.). I have not seen any in that package that would be made for any significant amount of power. High power transducers look quite different.
Yes, but even cheap digital voltmeters (this one is analog) would not work well for high frequencies and does not measure power, a volt meter is simply the wrong tool for the job.
It very much depends on the guts of the meter what it will indicate. "Something" is not a very scientific measure :)
The biggest factor are the ways in which the sample is taken and what kind of algorithm (if any...) is used to determine RMS. Simple meters simply rectify the AC current, add a small capacitor and a resistor to drain it and measure the DC voltage across the resistor. More advanced meters will use an isolation stage with a pre-amp driven from the auto ranger that will present the wave-form in one piece and suitably scaled to a DA stage where the sampling will take place. This allows for pretty good determination of the actual voltage (which in the case of AC has been agreed upon as root-mean-square, whereas, if the circuitry is bunk could easily become 'peak' or even 'nonsense').
Anyway, we're all in violent agreement that this measurement is bogus and since that is all the quantitative info there is I'll let it go here.
A better "instrument" would have been LED(s). Real power, more blinky blinky.
You could even attach a ultra lower power microcontroller dev board to it[1] and show the less technical audience "look! computer powered by spooky magic!"
> This is what I don't get - how the hell did they manage to get investors?
Because they did a couple of things absolutely right which was to pick something that'd have huge, giant, insane rewards if it works and picked something that's not so crazy implausible that it would require totally new physics to work.
Sadly even if it does work I wouldn't want it because even though you can't hear ultrasound I'm sure it wears on your ears just the same. The wavelength of 40kHz is 0.339 inches and 60kHz is 0.226 inches which are definitely on the right order of magnitude to resonate inside your ear which seems...sub optimal. I seem to remember that their tech operates in this band.
Another thing that nobody seems to be bringing up: animals. I don't know what's the full range of hearing of rats, but they talk a lot over ultrasound, with little-over-22kHz band being used to signal danger, and something around 56kHz to signal friendliness/food locations[0]. If their idea worked, they would likely piss off a lot of rats. Not many people keep rats as pets, but some other animals and usual city wildlife probably hear those ranges too.
[0] - source: my ex-gf had rats as pets and I did some googling a few years ago to figure out what kind of experiments I could do on them ^_^.
I've come to the conclusion that most VCs do not care at all about technical due diligence, or more broadly speeking analyzing the idea from the inside. They look at trends, momentum, traction, etc. It doesn't help that they are investing other people's money.
There's some logic to that: they can't invest the time and effort to understand all the problems in all the industries inside out. As a result, they look for proxies. The flip side is they get bitten by Goodhart's law. They end up being adversely selected by founders who mimic the signals the VCs have learned to pick up on.
Technology X is trending? Problem Y is a big problem? They'll fund a company that applies X to solve Y without asking themselves if X can actually solve Y or isn't even suited to solve Y. It's harder to be gamed if instead of following such trends you focus - as YC seems to be doing - on founder based signals. But people learn to mimic those signals too.
Signals which are costly to fake tend to be better. So investors will pay close attention to those. For instance, working full time on a project pre-seed investment instead of being payed for a full time job is a costly proposal, and thus potentially a good signal for investors. The flip side is that it can be faked by people with mediocre skills and no real opportunity cost, or people who are otherwise financially secure.
Not the ones I work for. And usually not the ones that my competitors work for either, DD is taken pretty seriously as a rule, including the tech portion. The VCs that don't take it serious end up getting burned and it doesn't take many of such cases to really wreck your returns. That's what makes this case so interesting. Either this thing works and nobody has figured out even the basics or a fairly large number of well-known names will end up with lots of egg on their collective faces.
Because clinkle does not seem to violate some basic laws of physics and could probably be made to work if it was pared down to basics rather than adding everything and the kitchen sink before launch. Audio modems are old school.
Even if it works perfectly, from the other article they had, pointing that they can get ~ 3W at a device (and their transmitter is how much, around 300W as well) should be enough to cast a lot of doubts
Many weekends ago, I did a bunch of research on UBeam as I had an opportunity to invest.
I hope I'm wrong (because wireless power would be cool), but after ~5-10 hours of research, I think wireless charging is unlikely to happen at any scale that would make UBeam a good investment.
The amount of sound power required to charge even small electronic devices is enormous relative to the volume of the sounds we hear every day. Therefore, charging must occur at a pretty high frequency that we can't hear. Power efficiency begins to drop at these high frequencies, however, and sound at these frequencies might have other unintended consequences beyond human hearing. Power transmission efficiency is low pretty much regardless.
The only way for UBeam to work, I think, is for people to think it's so cool to wirelessly charge things that they'll pay a lot of money to try it out. "Killing the power cord" on any sort of larger scale seems very unlikely.
I've seen, with my own eyes, working prototypes deliver amperage. That's pretty easy, though--just crank resistance down to 0, and you can get as many amps as you want.
These are the concerns that I have had since I first heard about this product. It's really simple physics... the falloff here is far too great. Come on, efficiency people!
45 comments
[ 4.8 ms ] story [ 99.8 ms ] threadI once lit a large fluorescent tube on the field generated by the antenna of a fairly powerful transmitter so that joke has some basis in physics. But you'd be hard pressed to run your vacuum cleaner that way :)
I wonder how many people clicked 'order'.
Sounds a lot like Theranos
> "But even amid a tidal wave of publicity, the company has never publicly demonstrated a fully functioning prototype of its system."
This is what I don't get - how the hell did they manage to get investors? I mean, ultrasound charging isn't magic, anyone who knows their stuff around EE should be able to throw together in a month a simple prototype of the tech to show to the investors. That is, a prototype that actually charges a phone, not this[0] which a) doesn't really charge the phone, and b) seems to be geared towards confusing people clueless about electricity. The fact that they didn't, after 4 years, produce anything suggests that they don't even have anyone on board who knows this stuff, so why are they getting money?
Seriously, I think I'm gonna plug my IRC bot to TTS and go raise money for my AGI startup.
[0] - https://www.youtube.com/watch?v=RoHxyweJcZI
On the other hand, given a a good RMS meter operated in a frequency range where it's spec'ed to work and a known circuit impedance will give you a very good measurement of the power: P=Vrms²/R. And analog meters make for rather pretty demos, to be honest.
In this case being able to supply 8 volts into a meter with a low internal resistance is actually a plus ;)
That said, the current is still tiny, even a crappy volt meter will pull no more than a few hundred micro-amps from the circuit that it is measuring. So let's be generous and say that it pulls a whopping 10 mA (very likely on the high side (edit: it turned out to be 37 uA)), in that case the power generated would be 0.08 W. Not enough to charge your phone by a long shot, but enough for a pretty demo with a moving needle.
Finally, these cheap meters tend to output complete nonsense values when presented with high frequency AC, one would expect that they would at a minimum measure power transfered using a dummy load and some kind of dedicated power meter.
All in all this does not make things look any better for uBeam. I'd really like to know who did their technical due diligence.
What's really bad to me - but I'm a nitpicker - is that in the video the lady refers to the volt meter as a power meter, which it definitely is not. It measures potential, not power, any power that it takes from the circuit can be chalked up to imperfect design and the laws of physics. Maybe the person doing the demo was not knowledgeable about the physics behind the equipment but that's betraying a complete mis-understanding of the basic concepts involved.
The transducers used appear to be of the piezo-electric type, those have a fairly high internal resistance and the voltages produced by those devices would be relatively high, the currents correspondingly tiny.
What gets me about this whole saga is that uBeam makes it seem as if they invented wireless power transfer using ultra-sonics, but any sound is wireless power transfer and that includes ultra-sound. The whole crux is how much power can be generated and whether or not there is a practical way to overcome all the obstacles.
To transfer some power doesn't prove a thing, when I whisper in your ear I'm transferring power using soundwaves...
Also: For low-level AC measurements you invariably need active rectification. So your meter will never be directly connected (or through a series resistor) to the source.
The guts of that meter register 37uA for full scale (I found some specs), at 10V that's an internal resistance of 270K so probably it really is driving the meter directly.
For 8 V that translates into 0.25 mW of power produced, which I don't think will be charging any cell-phones.
Exactly. These transducers look just like low-cost piezos used for ultrasonic ranging in air (e.g. car parking sensors, alarms, etc.). I have not seen any in that package that would be made for any significant amount of power. High power transducers look quite different.
The one the parent mentioned has 10kOhm/Volt sensitivity, so for 10V range it gives 100kOhms
http://support.radioshack.com/support_meters/doc73/73794.pdf
Correct, but it is going to measure something (it most likely rectifies and sends it to the meter, should work fine in the 100khz range)
> a volt meter is simply the wrong tool for the job
I don't disagree with this
The biggest factor are the ways in which the sample is taken and what kind of algorithm (if any...) is used to determine RMS. Simple meters simply rectify the AC current, add a small capacitor and a resistor to drain it and measure the DC voltage across the resistor. More advanced meters will use an isolation stage with a pre-amp driven from the auto ranger that will present the wave-form in one piece and suitably scaled to a DA stage where the sampling will take place. This allows for pretty good determination of the actual voltage (which in the case of AC has been agreed upon as root-mean-square, whereas, if the circuitry is bunk could easily become 'peak' or even 'nonsense').
Anyway, we're all in violent agreement that this measurement is bogus and since that is all the quantitative info there is I'll let it go here.
http://www.analog.com/en/products/linear-products/rms-to-dc-...
You could even attach a ultra lower power microcontroller dev board to it[1] and show the less technical audience "look! computer powered by spooky magic!"
1: like this: https://www.silabs.com/products/mcu/lowpower/Pages/efm32gg-s... super low power microcontroller demo board that runs off of all sorts of energy harvesting rigs like tiny solar cells, thermoelectric piles, and so on.
Because they did a couple of things absolutely right which was to pick something that'd have huge, giant, insane rewards if it works and picked something that's not so crazy implausible that it would require totally new physics to work.
Sadly even if it does work I wouldn't want it because even though you can't hear ultrasound I'm sure it wears on your ears just the same. The wavelength of 40kHz is 0.339 inches and 60kHz is 0.226 inches which are definitely on the right order of magnitude to resonate inside your ear which seems...sub optimal. I seem to remember that their tech operates in this band.
[0] - source: my ex-gf had rats as pets and I did some googling a few years ago to figure out what kind of experiments I could do on them ^_^.
There's some logic to that: they can't invest the time and effort to understand all the problems in all the industries inside out. As a result, they look for proxies. The flip side is they get bitten by Goodhart's law. They end up being adversely selected by founders who mimic the signals the VCs have learned to pick up on.
Technology X is trending? Problem Y is a big problem? They'll fund a company that applies X to solve Y without asking themselves if X can actually solve Y or isn't even suited to solve Y. It's harder to be gamed if instead of following such trends you focus - as YC seems to be doing - on founder based signals. But people learn to mimic those signals too.
Signals which are costly to fake tend to be better. So investors will pay close attention to those. For instance, working full time on a project pre-seed investment instead of being payed for a full time job is a costly proposal, and thus potentially a good signal for investors. The flip side is that it can be faked by people with mediocre skills and no real opportunity cost, or people who are otherwise financially secure.
I hope I'm wrong (because wireless power would be cool), but after ~5-10 hours of research, I think wireless charging is unlikely to happen at any scale that would make UBeam a good investment.
The amount of sound power required to charge even small electronic devices is enormous relative to the volume of the sounds we hear every day. Therefore, charging must occur at a pretty high frequency that we can't hear. Power efficiency begins to drop at these high frequencies, however, and sound at these frequencies might have other unintended consequences beyond human hearing. Power transmission efficiency is low pretty much regardless.
The only way for UBeam to work, I think, is for people to think it's so cool to wirelessly charge things that they'll pay a lot of money to try it out. "Killing the power cord" on any sort of larger scale seems very unlikely.