Spin-torque oscillators for WiFi band transmission and energy harvesting | Hacker News Reader

Spin-torque oscillators for WiFi band transmission and energy harvesting

nature.com

Energy harvesting is real, but the amount of energy that can be harvested from stray WiFi signals is vanishingly small. Even with a hypothetical 100% efficient collector, stray WiFi energy wouldn't be able to power much at all in a small device.

Practically speaking, sticking a tiny solar panel on a device would collect more energy unless the device is completely in the dark.

For an idea of exactly how small the FCC limits maximum transmitter output power (as fed into the antenna) for the band to 1 Watt. I.e. the low numbers in the results are expectations not early previews for much larger numbers in the future.

In practice even high power WiFi APs tend to top out around 300mW. Multiply by free space power loss and the inefficiencies of the system and the total available power is approximately bupkis.

Those tiny solar cells that power cheap calculators are a far more practical solution.

Exactly, moreover they speak about WIFI band, but in the experiment they didn't use an Access Point but a signal generator with the test device at * 2.5cm (0.9 inch) * from the antenna, if I've understood: "For the energy harvesting, we used the 2.45 GHz resonant patch antenna with a return loss of more than −35 dB at 2.45 GHz and a gain of 7 dBi. The antenna is fed by the signal generator at 0 dBm. The antenna is placed at ~2.5 cm away from the MTJ array". Let's try to measure at 2.5 meters. That is a laboratory experiment not intended to prove that you can give power to something moving around. They speak about neuromorphic computing: I know there are devices implanted int the brain of blind people receiving power from a radio device installed very not inside the cranium but at the outside, on the skin. That kind of application is more credible.

And WIFI is total bullshit. They say "WIFI frequency" aka 2.4 GHZ because, as we all know, that is the deregulated frequency reserved for consumer application. So, IMHO the researcher shouldn't have used that term, is confusing ! From there to say that the experiment involve the problem of powering cordless devices in the sense of electronic gadget with an access point is a complete misunderstanding of the sense of that research. They intend to power something very close to the emitter in medical/niche applications with a dedicated emitter. This is an important clarification because WIFI data transmission delivered power is NOT constant and too weak, differently from the one from a signal generator near the device.

Some conspiracy theorist freak want transform a legitimate scientific paper in some kind of proof of alien technology existance, maybe ! :-)

We changed the title from "First steps to converting 2.4GHz WiFi into usable energy to power small devices", which is not a good rewrite because it's impossible to quickly determine whether it's accurate, and because it doesn't use representative language from the article itself.

I've taken a crack at shortening the article's own title to fit HN's 80 char limit. Since I have zero idea what a "spin-torque oscillator" is, I probably got this wrong. If anyone can suggest a more accurate and neutral title, we can change it again.

Yeah, exactly the point I was try to explain being bullied for.

Thanks for the correction! It’s difficult to write an short title that can explain easily a scientific paper, and I failed =]

Meanwhile I can make a loop antenna on a banana box and tune to 880 6PR local am station and collect a few volts with enough drive for a loudspeaker (via a tx), light a dim led continuously or store up in a cap to make a burst of 2.4ghz data now and then.

Ok, it's the size of a banana box, but it is 24/7 dependable and a relatively decent amount of power compared to picowatts.

Isn't that sort of like stealing power? Wouldn't the station need to broadcast with more energy to reach any listeners "behind" your antenna-siphon? And isn't the energy transfer extremely inefficient?

It kind of sounds like the apocryphal tales of farmers who would lay big loops of cabling underneath HV lines to power their electric fences.

In Denmark the government have allowed people who live along some of the largest underground powerlines to install geothermal equipment to heat their houses.

You better be really, really sure the person who comes out to mark the utilities before you dig has the correct coordinates...

That might actually save the utility companies some power usage - the resistance of a material increases with temperature and the homeowners would be reducing the temperature of the ground.

seems like that's a win for both parties? the transmission line heating is a waste product, which increases resistance, which increases the waste. so if folks are willing to pump the heat out and productively use it, that's great.

It's very interesting. Do you have a link about this?

Anyway, this is very different.

In an electric wire part of the energy is always transformed into heat and wasted, so it's nice is someone can use it.

In this device, they add something to absorb part of the energy that otherwise would have traveled to your neighbor, so the transition tower must increase (slightly) the energy used.

It's like digging a hole and replacing a part of the high voltage cable with a device that makes more heat on purpose to heat your home.

1. No, because the power has already been transmitted. There would be problems if there were billions of these devices around a transmitter, because that would affect the loading. But one isn't going to make a difference.

2. Potentially yes, but the actual shadow is tiny. And transmitter powers are fixed and limited anyway.

3. Yes, of course. So is any form of undirected radio transmission.

4. This actually works. Kind of. You need to run the line parallel or build a resonant mix of L and C at 60Hz (or 50Hz.) You can also do things like power fluorescent tubes by induction.

https://www.trendhunter.com/trends/magnetic-field-fluorescen...

The problem is the voltage/current is very hard to control, and the whole point of electric fences is that they're not lethal. You'll get something out of a resonant circuit, but if you're not a qualified electrical engineer it won't be the clean 110/60 or 220/50 needed for an electric fence.

Points 2 and 3 seem like they could contradict each other. Wouldn't a tiny amount of power for you equate to a much larger amount of power at the radiation's source?

No. The transfer is inefficient because the vast majority of the transmitted power is dissipated in the environment as heat.

The transmitter cannot tell if switch the receiver on and off.

It only works because the LC circuit has a high "Q" and it peaks up at resonant frequency, at the expense of current. It is sort of like a torque convertor for electricity.

Now in most applications a minimum voltage is required and you will counter the small current by using a capacitor to accumulate power, at a usable voltage, until you have enough to briefly do something.

I would make a YouTube video if I could be arsed, but way to busy right now.

'TheOtherHobbes offered a detailed reply, but I'd like to add a useful mental model for such questions. Like all other EM radiation, radio waves are just light. They diffuse a bit differently than visible light due to much longer wavelengths, and different materials are transparent/opaque to them, but to a first approximation, you can view the transmitting antenna as a lightbulb, and the receivers as opaque and weakly-reflective objects.

Your questions are thus similar to asking, "If you put a tiny solar cell on a glass table, wouldn't that steal light from the lightbulb? Wouldn't we need a more powerful lightbulb to illuminate the room? Isn't this energy transfer extremely inefficient?". Yes - it consumes some power; no - it casts a tiny shadow, and enough light is scattered around that you won't see the difference; yes, if the only reason you turned the lightbulb on is to power that solar cell, it's very inefficient.

That would mean listening to the radio was stealing power - they don't care, they want you to listen to their advertisements...

So they're using low to medium energy microwaves to power devices? Well if anyone doesn't buy that they can always try and sell it as a heater [1].

[1] https://www.youtube.com/embed/Di3fPj0pUbQ

Seems possible if a crystal radio is possible.

https://en.m.wikipedia.org/wiki/Crystal_radio

Not to play the old man card, but I remember building them when I was younger -- and they worked for sure! You could tune in and listen to all sorts of broadcasts, no battery required at all.

Same here. I'm now an Amateur Radio operator. I wonder if there's a connection.

This is a big deal for IoT, and ubiquitous sensing. It's not there yet, but you can see that a few years down the road we might be able to power very simple sensors via existing wireless networks (for free). That will allow us to put sensors on everything.

Really not sure what they are trying to get at here, but their "usable energy" is on the pico-watt to micro-watt scale. It looks more and more like one of those "free energy" type devices.

Edit: Yes I understand that there is energy in RF signals, I am just questioning the actual practicality and claim that they can get usable energy from this. For a 200 mW (23 dBm) WiFi transmitter the free space loss at about 6 inches is 23 dB at which point you would have the same 0 dBm input that they used in the experiment. I have not included any antenna gain as they did not include the antenna gain from their microstrip patch antenna in their experiment, just the 0 dBm power input number.

The device can receive:

> For the condition of synchronization of four oscillators at 2.4 GHz with Idc,sync = 3 mA (1.2 mA) for the parallel (series) configuration

That’s very low but I think a first step in order to power small biometric implants, for example some kind of pacemakers. It’s very frustrating have a surgical operation to change the battery every 10/15 years.

Would be an ironic way to power a pacemaker as you'd be asked to start standing near the microwave.

Ahah that’s true, because some pacemakers send datas to the receiver/hospital through 2.4ghz band, so it can have interferences. But mine was only an example because my father has one of those device into his body and sooner or later it will need a “battery swap”. Maybe not a pacemaker but to power a small temperature sensor, or something similar, it will be useful, send datas and get energy, quite nice and futuristic!

The current you refer to there is not the power that is being generated though, it seems to be an input in to the STO devices.

> The STOs are connected in parallel (Fig. 1a) or in series (Fig. 1b) configuration, stimulated by a single dc source. For the parallel connection, at the dc bias (Idc) value of 1.5 mA

Fig 2c and d shows the total output power which maxes out at 0.85 uW. Maybe this is enough to run a pacemaker or something small, I don't actually know. Fig 6 shows the rectification of the signal in to a voltage that can turn on an LED. The dc-dc converter puts it in the ~3V range. Assuming the dc-dc converter is near 100% efficiency (some are pretty damn good) this is a current of only 0.28 uA.

Its a novel way, but they talk about in their own paper that it is not a new more efficient way to convert RF energy to DC energy

> in the range of 1.65–2.8 GHz at Prf = 0 dBm and zero dc bias ac to dc conversion efficiency of ∼6% at Prf = −20 dBm. This is higher than the reported values for STOs53,55 but less than the recent reports of power conversion efficiency of 40% and 40–70% at an input rf power ~0 dBm for MoS2-based flexible rectenna and state-of-the-art Si and GaAs rectifiers56

> their "usable energy" is on the pico-watt to micro-watt scale

devices which have long-term average power consumptions of micro-watts are perfectly reasonable. it's how some BLE sensors can run for years off a single coin cell battery.

you'd need some energy harvesting circuitry and something low-leakage to build it up in, but that's doable. more expensive than just stuffing a coin cell into the circuit, though.

> It looks more and more like one of those "free energy" type devices.

it's perfectly well understood that there's energy in RF. just very very little of it, as you note.

that's qualitatively different from the claims of most free energy devices.

I wonder if these could also improve bandwidth for other devices since the absorbed waves won't bounce / interfere with actual signals.

Reading the abstract, my 'this isn't real science' detector went crazy...

It's something about the fact they explained very basic things (series Vs parallel, battery-free), while also having a lot of words I didn't understand.

Eh, it's basically a new(-ish) way to make a small radio receiver but instead of converting the output power of the antenna into an input signal for an amplifier, they try to feed a small electric device with it. This is super possible and doesn't violate known physics in any way, but the amount of power you can transmit is minuscule. Still, it might be useful for some type of sensor where it is non desirable to have to replace the battery for some reason.

The problem is getting enough power to broadcast the sensor report. It means you have to run your sensors with a thimble of energy and store up remaining energy for a burst of reports (hope someone is listening :))

There is this company [https://uwinloc.com] , I've been working for. However ,they are using 968MHz band to harvest energy, thus (lower freq = less loss due to distance). It works up to ~20m if the antenna are correctly "aligned".

Products similar to that already exist, albeit not as low power as the device in the paper. For example, this passively powered UHF RFID tag has a moisture sensor built in: https://www.dipolerfid.com/RFID-Tags/Smartrac-Dogbone-Humidi...

> a burst of reports

That assumes the device needs to be running constantly collecting data.

What if a circuit passively charges a capacitor that once full powers up the device just long enough to take a reading and transmit it? This wouldn't be viable for things like air monitors where the sensors need to warm up for minutes, but it seems like it would be ideal for temperature sensors which can be influenced by the waste heat of the device.

If you don't need much range (a couple feet, say), you can use ambient backscatter which doesn't need to store up energy for bursts.

Here's a video about it from 2013 [1] showing some test devices. Here's an article about it [2].

A bit later they extended this to make it work with off-the-shelf WiFi devices, so you could have a battery-free device that can communicate to ordinary WiFi devices [3]. Still short range, but now you could make it so the reader is a commodity WiFi device like a smartphone, instead of a specialty device.

Here are some other researchers who demonstrated a backscatter tag that diddled with Bluetooth Low Energy (BLE) instead of WiFi, and was correctly received by an iPad at over 9 m using the existing iOS Bluetooth stack with no modifications [4].

And here is a battery-fee phone that communicates with a powered base station [5]. It works up to 30 ft away. The phone uses energy from ambient RF and small photodiodes.

[1] https://www.youtube.com/watch?v=gX9cbxLSOkE

[2] https://www.washington.edu/news/2013/08/13/wireless-devices-...

[3] https://iotwifi.cs.washington.edu/files/wifiBackscatter.pdf

[4] https://www.washington.edu/news/2017/07/05/first-battery-fre...

> and small photodiodes.

I bet the small photodiodes collect far more energy than the RF harvesting...

How would you handle payment for the energy and how would you make sure the energy ends up at the proper device and not at some man-in-the-middle?

It isn't meant for energy delivery, but energy harvesting. Put a metal stick in place of the antenna, the same power is harvested, but dissipated away.

I can see this useful for very low power devices. I don't have specific example in mind, but you can picture something like sensors embedded inside smart concrete high-rise buildings, that alert when crackling is found.

The energy to generate the 2.4GHz signal is already being lost; I don't believe that placing a spintronic device attenuates the signal in any meaningful way. In certain ways it's similar to tidal energy; one small tidal generator is not going to have any measurable effect on an ocean-scale system. The power absorbed is measured in microwatts whereas the power consumption of a consumer router is measured in watts.

Or maybe just convert the ubiquitous 50Hz/60Hz power line signal?

Or a small solar cell? (like many calculators?)

Probably more efficient. But less hype worthy.

If someone reading the post here is thinking that will be possible give power to "small devices" using WIFI, that is not and will not be impossible. We are talking about few micro amperes when a coin cell battery can deliver 220.000 micro amperes, as paragon. There are some niche applications in micro ampere range, but in that cases there are some reliable and more stable solution, like cesium batteries. Dave of EEVBlog has a mission, debunk this wireless energy affirmation, explaining why it's simply impossible, for example: https://www.youtube.com/watch?v=P8s3Xjeg0sk

Moreover the power is related to the distance to the wifi source and the presence of an active WIFI obviously doing this kind of solution at least impractical even for ultra low power consumption applications. An Apple tag uses a coin battery and I think will use that solution for a veeeeery long time.

Maybe he’s right but his video is from “jan 2010”, maybe in 11 years some improvements are been made… and this research doesn’t talk about “power a smartphone” but a very very small device.

So you provide a video of Dave showing the math that you can get only microwatts out of radio waves to debunk something that ... claims to provide power in the microwatts range?

The "charge a phone" thing is obviously bunk, but that's not what's discussed here.

Show me applications in microwatt range to justify the "usable", moreover you need a wifi. Why someone should use that instead of a solar cell, like the one on calculators ?

this didn't claim to have anything to do with "common user"s. energy harvesting/storage for IoT-type nodes is of huge interest, with lots of work being done on it.

> Dave of EEVBlog has a mission

and i don't think he'd appreciate your "contribution" here to that mission.

I wrote it fast, I rewrote it. My point is that the title is misleading and suggest you can use WIFI radiation as power source to do something useful in the near future in consumer electronics. You simply will not.

you're the one who hallucinated "consumer" into all of this. nobody said that. it wasn't in the title. this isn't a consumer electronics forum.

as the abstract of the article points out, it's for energy harvesting applications. which have been demonstrated as being perfectly useful with microwatt collection rates.

The administrators changed the title of this article for a reason. Moreover, that device is intended to be installed very close to the "consumer", that is useful in medical application, for example, where is no advisable to install the battery in a device to be implanted in the body. So you can install the emitter , transmitter, whatever you want call it on the skin. And WIFI is total bullshit. The say "WIFI frequency" aka 2.4 GHZ because as we all known is the deregulated frequency reserved for consumer application. So, IMHO the researcher shouldn't have used that term, is confusing ! From there to say that the experiment involve the problem of powering cordless devices in the sense of electronic gadget is complete misunderstanding of the sense of that research. Man , science is not an opinion.

it kind of looks like you're having a rough day. i hope you feel better.

Simply I don't like fake news.

errata corrige: ... that is not and will not be possible.

The title of this article was changed, by the administrators I guess, I think it's good in times where someone think opinions are more relevant then science.

Reply on news.ycombinator.com