Slash Boxes

SoylentNews is people

posted by Fnord666 on Friday June 04, @10:24AM   Printer-friendly [Skip to comment(s)]

Scientists at Sandia National Laboratories have built the world's smallest and best acoustic amplifier. And they did it using a concept that was all but abandoned for almost 50 years. According to a paper published May 13 in Nature Communications, the device is more than 10 times more effective than the earlier versions. The design and future research directions hold promise for smaller wireless technology. Modern cell phones are packed with radios to send and receive phone calls, text messages and high-speed data.

Scientists tried making acoustic radio-frequency amplifiers decades ago, but the last major academic papers from these efforts were published in the 1970s.

The new and improved amplifier is more than 10 times as effective as the versions built in the '70s in a few ways. It can boost signal strength by a factor of 100 in 0.008 inch (0.2 millimeter) with only 36 volts of electricity and 20 milliwatts of power.

[...] So how long until these petite radio parts are inside your phone? Probably not for a while, Eichenfield said. Converting mass-produced, commercial products like cell phones to all acousto-electric technology would require a massive overhaul of the manufacturing infrastructure, he said. But for small productions of specialized devices, the technology holds more immediate promise.

The Sandia team is now exploring whether they can adapt their technology to improve all-optical signal processing, too. They are also interested in finding out if the technology can help isolate and manipulate single quanta of sound, called phonons, which would potentially make it useful for controlling and making measurements in some quantum computers.


[Source]: Sandia Labs

Journal Reference:
Lisa Hackett, Michael Miller, Felicia Brimigion, et al. Towards single-chip radiofrequency signal processing via acoustoelectric electron–phonon interactions [open], Nature Communications (DOI: 10.1038/s41467-021-22935-1)

Original Submission

This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 2) by driverless on Friday June 04, @12:04PM

    by driverless (4770) on Friday June 04, @12:04PM (#1141741)

    Maybe I'm missing something, but this sounds like a really, really inefficient, power-hungry, physically huge equivalent to a transistor. Where's the magic?

    Or, alternatively, you can see why they stopped playing with these things fifty years ago.

  • (Score: 0) by Anonymous Coward on Friday June 04, @12:10PM

    by Anonymous Coward on Friday June 04, @12:10PM (#1141742)

    I wish they would just leave those old highpotheses up on the shelf where they found them. Do you know how much dust those things collect in 50 years?

  • (Score: 5, Interesting) by VLM on Friday June 04, @12:35PM (1 child)

    by VLM (445) Subscriber Badge on Friday June 04, @12:35PM (#1141747)

    You'll probably find more stuff online from the Carter-Reagan years under the name "Acoustoelectric Devices" and similar variations.

    So say you take a piezoelectric media like quartz and couple a RF surface wave onto it (it has a skin depth just like copper). Everyone's used to coupling a bulk chunk for crystal oscillator stuff but I'm talking about beaming a wave right into it at one end and removing at the other end.

    First thing is even at whats now low frequencies like 500 mhz the sizes are microns so to keep the waves from self-interference and noise sources you need it clean to like nanometers. Which immediately provides a scaling problem where if you wanna do military X-band radar instead of UHF you'd need to manufacture to like hundredths of a nanometer. So this ain't happening even after half a century of semiconductor manufacturing size shrinking. However it HAS improved such that fractional GHZ IF devices were unmanufacturable at high enough performance in 1970 but "some nanometers" fabs are now available enough to realistically build stuff at that size even if its quartz instead of semiconductor silicon but whatever.

    Second thing is instead of running in a vacuum (which is a PITA BTW) you can slap the peizoelectric thing up against a semiconductor and even in 1970s you could make something almost exactly like a traveling wave tube but solid state and more like a UHF frequency MASER. Couple problems: micron scale structure requiring nano scale precision to keep things clean because impedance is controlled by shape and gain is strongly dependent on impedance. Tiny variations in 1D are bad enough, add tiny variations in 2D or 3D due to material imperfections and your amp is a mixer with incredibly bad IMD specs. Being what amounts to a UHF MASER in a non-resonant lump of material, it operates in multiple modes and then mixes them together for staggeringly poor performance. Most DC power turned into heat and the best gain (which was usually pretty low anyway) came at ridiculous voltage/cm fields so a hundred times the frequency where it gets "fun" would be a hundredth the gain aka maybe a loss...

    There were hints it could be interesting in the 70s. You could get crazy gain per lineal inch. In theory you could produce the things as easy as oscillator crystals or capacitors, well, in theory. It "could" be more radiation proof than junction transistors under some circumstances? Although the performance would be miserable it could be more broadband maybe for ECM EECM military type stuff, but they love their traveling wave tubes so ...

    The biggest problem long term is its a 3d bulk material operator unlike junction transistor technology thats sorta 1-D ish so its easier to work around mfgr imperfections with junctions rather than bulk. Also there seems to be no way to make bulk material less noise than junctions. Finally both for bulk material like this and junctions its possible to run right up to thermal limits and junctions simply give better performance if you're "heat sink limited" or "battery limited". So at low signal levels the high noise of this tech means transistor tech wins. At high power the higher efficiency and better specs of transistors means they win again. At frequencies lower than MHz COTS opamps crush what the tech could be capable of. At UHF frequencies this tech was borderline unbuildable half a century ago and is possible now in 2020s. Above UHF range like cell phone range the feature size and imperfections would have to shrink smaller than we can now build CPUs and memory chips so that aint happening for yet another half century or whatever at any rate its not happening now. As a tech it seems doomed, hemmed in from all directions, transistor tech seems to beat it under all circumstances from all directions.

    Its one of those obscure attractive EE things that never amounted to anything like memristors or tunnel diodes or varactor parametric amplifiers or magnetic amplifiers or fluidic valves where its interesting to think and know about and I wouldn't mind having some parts to play with on my lab bench but its never gonna be in something that ships, not in 2020s anyway.

    • (Score: 5, Interesting) by VLM on Friday June 04, @12:55PM

      by VLM (445) Subscriber Badge on Friday June 04, @12:55PM (#1141749)

      You could get crazy gain per lineal inch.

      To edit my remarks, crazy gain per inch IF you look at it like a solid state traveling wave tube. A mid UHF traveling wave tube would get like 5 dB/inch not 50 dB/inch like these things could.

      The problem is a pile of tiny fraction of an inch transistors can do into a discrete shielded circuit or MMIC chip that gives far more "gain per inch".

      Great performance compared to 1950s vacuum tubes. Not so good compared to 2020 ICs or 1970s RF transistors.

      I guess the appeal of this tech is its a perfectly valid "what if" alternative history. What if in 1940 the guys doing RADAR with traveling wave tubes said F it and went with acoustoelectronic components instead? 1950s performance in 1940 isn't so bad... How would the EE world change maybe in the 60s, 90s, and today? I think acoustoelectrics might have made old fashioned NTSC colorburst circuits microscopically simpler maybe we'd have had more earlier more reliable color TVs.

      Sorta like AFAIK nobody in the ancient computer era (like 1950s) ever tried mass scale magnetic amps for bulk computing, like in an ALU. The power requirements would have been staggering but at least the reliability is higher. Imagine something like a classic IBM 700 series built with magamps instead of vacuum tubes. Well it would probably glow about as hot as the sun and require its own nuclear power plant, but aside from that minor problem...

  • (Score: 0) by Anonymous Coward on Friday June 04, @01:21PM (4 children)

    by Anonymous Coward on Friday June 04, @01:21PM (#1141753)

    OK, I know it's Friday and I'm tired, but there are so many groups of words that I cannot make sense of... I have to wonder if it's just me:
    - "acoustic radio-frequency": Which is it? Are we talking about moving air or electrons?
    - From TFA "the wavelengths of sound at these frequencies are so small". If we're talking about sound (*low* frequencies of 20-20 kHz), then isn't the wavelength huge, even if in a solid material?
    - "Boosting a signal by a factor of 100 with the old devices required 0.4 inch [...] more than 500 milliwatts of power" and "The new and improved amplifier [...] 0.008 inch [...] 20 milliwatts of power". So, for the same amplification efficiency, the new amplifier has higher heat loss, and that's "more than 10 times as effective"?

    Maybe there is some amazing breakthrough there, but TFA doesn't make it easy to get.

    • (Score: 2) by theluggage on Friday June 04, @02:08PM (3 children)

      by theluggage (1797) on Friday June 04, @02:08PM (#1141768)

      OK, I know it's Friday and I'm tired, but there are so many groups of words that I cannot make sense of... I have to wonder if it's just me:
      - "acoustic radio-frequency": Which is it? Are we talking about moving air or electrons?

      I'm guessing, too, but I'd guess that 'acoustic' means that the signals are being processed in the form of mechanical motion (which doesn't have to be at what human ears regard as audio frequencies) rather than fluctuating electric currents or EM waves.

      Just imagine the HMV logo re-imagined as an electron microscope image of a cute tardigrade with it's antennae/cillae/whatever cocked towards a tiny wind-up gramophone with a horn and diaphragm mechanism and hearing his master's voice sped up a few million times. It's probably nothing whatsoever like that, but it's a great way to pass the time. :-)

      • (Score: 0) by Anonymous Coward on Friday June 04, @04:04PM (2 children)

        by Anonymous Coward on Friday June 04, @04:04PM (#1141812)

        Once I worked out that this wasn't more audiophile porn, then I realized the same thing as your guess. Signal processing done as mechanical vibration (in a very small mechanical oscillator).

        Anyone know if VLM's comments above make sense? He rants and raves off-topic so often, that it's hard for me to know what to think when he writes something that looks sensible(grin)?

        • (Score: 0) by Anonymous Coward on Friday June 04, @05:12PM

          by Anonymous Coward on Friday June 04, @05:12PM (#1141836)

          Almost makes sense, yeah.

          Piezoelectric acoustic components are actually pretty common - especially in cell phones. They're called SAW or BAW filters, which stands for surface or bulk acoustic wave (depending on if the wave travels along the surface of the material or through the middle of it). The incoming signal (which is moving electrons, not air molecules) gets put into the input end of the piezoelectric chip, then comes out the other end as an (again electronic) filtered signal. This allows a nice bandpass filter to be constructed, effective in radio frequencies, without the need for space-and-power-consuming networks of capacitors and resistors. By varying the properties of the chip, the filter can be adjusted. The drawback is that the filter's properties are baked in at manufacturing time, and it has an attenuating effect on the signal due to the dual conversion. Overall it works well as long as someone makes a filter in the frequency you need.

          But those are filters, not amplifiers. So, different tech.

        • (Score: 3, Informative) by epitaxial on Friday June 04, @08:14PM

          by epitaxial (3165) on Friday June 04, @08:14PM (#1141870)

          Reading VLM's comments was refreshing. Nice to see someone with knowledge of a subject instead of simply dismissing the story.