Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Audio amp for sonoluminescense

  1. Feb 4, 2016 #1
    I'm trying to replicate a simple single bubble sonoluminescence (SBSL) experiment. I have several sources that I'm using for reference and my setup is pretty similar to the one here: http://techmind.org/sl/ but our lab does not have an amplifier. I have tested the circuit w/o one (just running the signal straight from the wave form generator) but it doesn't seem to have the voltage I need. I then tried a small lab amp (from Vernier) that we have several of, but it did strange things (clipping and changing the wave form from sine to triangle, etc. - it even made a weird, uneven square wave show up on the scope when the amp was turned OFF... idk, and neither did my profs), plus I knew before testing it that it is only rate at +/- 10V and I am shooting for closer to 30V peak to peak. I have seen people saying they have used an audio amp successfully for this application, but I worry about distortion. I have only rudimentary understanding of electronics but am teaching myself as much about it (as far as is relevant to this experiment) as I can, and, if I understand correctly, introducing proper inductance between the amplifier and the driver transducers should act like a high pass filter to help clean up the amplified signal. I have calculated the theoretical inductance for my setup (iirc ~30mH) and have some home made type inductors ready to test. Although I don't have a meter to measure inductance, it shouldn't be too hard to match it through trail and error.

    So, I am looking for a suitable amp and am having trouble finding one. Price is a concern. I don't know the exact budget at this time, but cheaper is better. Building one is not completely out of the question, but time is also a concern, so buying one is preferred. I'm not 100% on what specs I should be prioritizing because none of my references give much detail on the amp they used. The only firm stats I have seen concerning the amp is that ~30V peak to peak seems to be the goal. Most of what I'm seeing for sale is for stereo system application and doesn't even list the detailed specs that I'd like to see. I have found some low distortion amp boards with voltage ratings that are too low, but that have good distortion ratings (~0.006% or less) and I have seen audio amps with appropriate voltages but distortion ratings that I'm not sure would work (~4%). Any insights or suggestions would be appreciated greatly.

    I am also wondering if my basic understanding of the inductors acting as a high pass filter here is correct. The transducers I'm using are effectively acting as plate capacitors, which is how I was able to calculate the theoretical inductance needed. If I understand correctly, (excuse my simplistic wording, this isn't my field) the inductors will introduce a low impedance magnetic resistance to voltage spikes, which will help to keep the signal clean when properly matched to the circuit. Also, as I understand it, I can set the inductors up to be variable by having a ferrite rod that I can slide into.out of the center of the coil as needed as well as having 2 inductors, wired in series, that can be moved closer/farther apart to vary the total magnetic field strength. Clearly I could use some insights here as well...

    EDIT: After some reading on audio amps it seems I should be able to get away with using one so long as it is rated higher than what I will need. Problem is I'm having a hard time finding one with those ratings. I need a minimum of 30V peak to peak output that can handle at least 30kHz and as low distortion rating as I can find (<<1%). If anyone knows a good source for an amp like this please let me know.
    Last edited: Feb 4, 2016
  2. jcsd
  3. Feb 4, 2016 #2


    User Avatar
    Homework Helper
    Gold Member

    I'd never heard of sonoluminescence before, so thanks for introducing me to something new.
    In the absence of any other responses yet, I thought I'd offer some of my thoughts on the electricals.

    If you look in their postscript (and FAQ), they point out that only about 1W of power is needed. They also indicate one way of achieving the voltage increase that you need - using a transformer. (They suggest an audio line transformer.)
    Since a 1W amp will output about 2.7V, you need to increase this by a factor of 30/2.7 = say 12 to get your 30V
    If your amp can provide say 5W as 6V into 8 Ohm, then you can get your 30V with a step up of 5.
    I don't have much experience of using audio amps in this way, but they generally have good low distortion performance, when operated within their design parameters. Often this means at well below max power, say 10% of rated power, and into a correct load, typically >4 Ohm, mainly resistive. (I'm being a bit conservative there. The allowable % power level varies a lot. Some argue that audio distortion is only audible at low volume settings, so they don't worry about distortion so much when the volume is turned way up.)
    This would represent about 25W into 8 Ohm.
    It seems a reasonable amp to go for, though your link suggests you could get away with less. As I said above, with audio amps operating them well below rating is good for keeping distortion down. I've no idea what distortion you should tolerate. The tuning to resonance will greatly reduce any other frequencies in any case. ~0.006% sounds good to me, but even ~4%, if you can run this at a low volume setting, sounds plausible to me.

    As shown in link the inductor would be low pass. A parallel inductance would be high pass. However, I think low pass is what you want. The ripples on your sinewave are generally higher fequencies. Harmonics and clipping distortion give rise to higher frequencies, not lower. Mains hum is generally the only low frequency problem and this is so far from your frequencies of interest, that I'd expect it not to matter.
    In fact the inductance is, as you say used to tune the ceramic discs to resonance. Resonance discriminates against other frequencies, both high and low.
    I'm not really clear about why you are so worried about small amounts of other frequencies. Harmonics seem to be the main concern, because they can cause other nodes in the standing wave pattern. Provided they are significantly weaker than the main frequency, which itself needs to be in a narrow amplitude band, then I'd have thought they must be outside the range of levels wich matter.
    You're quite right here. Suck it and see! There's a lot of non-ideal behaviour in this sort of thing (IMO), but unless you come across it, don't worry too much. My biggest concern about inductors would be frequency. Radio frequency (low RF) ferrites should be fine, but if
    you used a 50Hz transformer for voltage step up, that might have big losses. That might not matter if your amp has a bit of power to spare. Audio transformers I'm not so familiar with, but I'd guess they might be ok at 25 or 30kHz as they have to operate well at 20kHz plus.
    Remember P x R = V^2 so if you know the RMS constant output power and the speaker resistance it is designed for (usually 4 or 8 Ohm) then V = sqrt(P x R), eg. 25 W into 8 Ohm gives V= sqrt(200) = 14 V

    (There are others on here who know a lot more about this than I: if you don't get what you want in Physics, try electrical engineering.)
  4. Feb 4, 2016 #3


    Staff: Mentor

    I envy you. It sounds like a fun experiment.

    A quick check on EBay.com showed numerous signal generators for less than $50 that might meet your needs.

    Also, a quick check for sonoluminescence on YouTube.com showed numerous home experiments and how-to videos that might give ideas.
  5. Feb 4, 2016 #4
    Thanks guys. Yes, SL is an exceptionally fascinating phenomenon. It's one of those rare things that can be recreated quite reasonably, yet is not well understood (particularly in terms of the underlying mechanism). If you are interested in looking into it more I can suggest the website of Dr. Putterman's lab at UCLA. He is probably the foremost expert on the subject and is credited with the shockwave theory, which seems at present to be the most likely and widely accepted mechanistic theory out there. His and some associated labs (esp. Taleyarkhan, et al.) have recently confirmed a dense plasma formation in SL, which suggests the photons are likely emitted due to Bremsstrahlung radiation (what a name, eh?). All this points to fusion (like the SUN - it turns out the nickname 'star in a jar' is probably more apropos than might have been expected) and Putterman's shockwave theory.

    For those interested, here's a very rough explanation of this theoretical mechanism. The collapsing bubble reaches a point where Van der Waal's forces suddenly overtake the (acoutic, mechanical) force driving the collapse, causing the interior walls to stop collapsing violently, at which point a shockwave (not subject to Van der Waal's forces) continues the collapse. The remaining (noble) gas atoms are accelerated with/by the shockwave until they too are forced to stop exceptionally quickly. Note: if using air in water the 1% Argon does the job while other constituents are turned to H2O2 and NOX in a 2 stage set of rxns under the intense heat/pressure. This is when Bremsstrahlung (photon emission) theoretically takes place. For those of you who, like myself, don't know German, Bremsstrahlung roughly translates to braking (as in car brakes) radiation. The acoustic waves don't really carry much energy compared to that required for Bremsstrahlung, or any of the other theoretical mechanisms for that matter. The condensation of energy is nothing short of incredible. I have one reference that claims it to necessarily be on the order of a factor of 1 trillion!

    I can also suggest chasing after some of the other theories that have come up over the years (teaser: a couple of my favorites have to do with lifesavers and Hawking radiation). It's quite the rabbit hole. But I think Putterman and the gang are getting pretty close to the rabbit so I'm glad I'm getting the chance to jump in on the action in my small way before the curtain is pulled down. I'm still working on it and will be going in bright and early to find out exactly what the load is as I have it set up now (kicking myself for not taking that measurement already).

    Sorry about the long post, but I get excited about this stuff... Now, for some clarification, if I measure the load at 8 Ohms, then a 25W amp will give me the 30V peak to peak that I'm after, but, since I want to run the amp at ~10% of its power rating, I should get one rated at ~250W?

    Finally, and I suppose I should have mentioned this in the first post, probably the main factor in rejecting most of the amps I'm seeing (other than maybe price) is that they are rated for a max frequency of around 20kHz (usually 20Hz-20kHz) and I will be working in the 25-30kHz range. Being outside the human hearing range, most audio applications really have no need to be in that frequency, I suppose. If I understand it right, the gain is liable to drop off dramatically outside the rated frequency response range. I have heard that some amps will work really well outside the response range, but that you have to test the gain at the target against the gain inside the rated range to find out if it is still doing its job and that's something I can't do until I buy it.

    So, maybe I should focus on finding a lower wattage with the correct frequency range and worry about matching my inductance? In any case it is proving troublesome trying to find anything rated to these higher frequencies. I did find this: https://www.amazon.com/Electronics1...=1454639286&sr=8-76&keywords=50W+amplifier+ab (wish I knew how to shorten that link up) which seems to have the specs I'm looking for, but it's gonna be almost 50 bucks with shipping, needs assembly, and has no heatsink or power supply and may need some other parts that I'm not even thinking of according to that amazon page...

    It's been a long day. Maybe tomorrow will bring an end to this search. In any case you guys have already been a huge help and I thank you again.

    EDIT: The link was automatically shortened a bit. Thanks PF!
    Last edited by a moderator: May 7, 2017
  6. Feb 5, 2016 #5


    User Avatar
    Science Advisor

    You state that you need 30Vp-p to drive the transducer.

    That's 10.8V RMS (or 30/2.828).
    Relating that to amplifier power (W=E^2/R) is 117/8=15W RMS, the "8" is the assumed speaker impedance for which the amplifier is rated.
    In that case a step-up transformer would not be needed. Or, since you can wind wind inductors, you can wind your own transformer using a ferrite core.

    Watch out for the way the amplifier is rated though. Rather than using RMS, most consumer audio amps are rated PMEP (Peak Music Envelope Power) which can be ridiculously high. I've seen some advertised as 10W output when powered by a 3.5 Watt wall-wart power pack.

    Another possibility for boosting the drive voltaage to the transducer is make it part of a tuned circuit. That is, use an inductor with the transducer so you end up with a resonant circuit at the driving frequency. The voltage amplification at resonance will be equal to the "Q" of the circuit. Just don't exceed the voltage withstand capability of the driving amplifier. This will also greatly reduce any waveform distortion that the transducer sees.

    You don't say how much power is actually needed into the transducer, this could be a concern if trying to use a signal generator directly as a source as they are typically very low power voltage sources.

    My suggestions would be select an amplifier with suffucient frequency response and power output then find or wind a resonant transformer to match the transducer. A possible source for a ferrite core would be the transformer core from an old switching power supply.

    If you prefer to build an amplifier you would probably need only a gain stage driving a push-pull or a totem-pole output stage. With a 50V power supply you could drive the transducer directly; although you may still want to resonate the transducer to cleanup any distortion.

    You also refer to "single bubble" sonoluminescence. If this requires a pulsed rather than a continuous signal, resonating the transducer could be a problem; it will ring long after the driving pulse stops. You would then need a high degree of damping, either active or passive.

    Looking forward to what you actually come up with. Please let us know the results.
    Last edited: Feb 5, 2016
  7. Feb 5, 2016 #6


    User Avatar
    Homework Helper
    Gold Member

    Quite! Input power is always a reality check on the wilder claims of the ad men.
    As the link already does. This is reminiscent of continuous wave radio transmitters. They used class C amplifiers, driving the amp with pulses at the required frequency - a near square wave - and relying on the tuned "tank" circuit (and low pass output filter) to ensure the output contained very little harmonic energy.
    Link says, "In fact typically only ONE watt is dissipated in the transducers/drive system during sonoluminescence."
    No. Again see link. (Though whether modulated drive has any advantages, may be an open question?) The load itself would provide some damping.

    Generally agree with your suggestions. I'd say resonance is mandatory to sensibly get power into piezo disc.

    And like you, I'll follow this with interest (though less optimism than others.) I don't know whether there's any way of having a thread about the physics of this phenomenon. It raises many questions in my mind, which are not addressed by anything I've yet discovered, so I'd love to see what some of our real physicists make of it.
  8. Feb 5, 2016 #7


    Staff: Mentor

    Pardon me for being skeptical about the need for spectral purity in the signal generator. The transducer will introduce some distortions, and sound reflections from the chamber walls will add others.

    Also, the voltage and power levels required are a function of the transducer. You need a matched set of components end-to-end as opposed to specifying one at a time.

    A frequency doubler circuit may be of interest. Unfortunately, all the commercial doublers I found were for RF frequencies. https://en.wikipedia.org/wiki/Frequency_multiplier

    Here is an audio signal generator that will work fine 10 hz - 100 khz. You may need to boost the voltage.
  9. Feb 6, 2016 #8
    Again, thanks for all the input, everyone.

    I went in yesterday and did some testing. I found some issues with some of the equipment/connections I was using so I tore everything down, got out the best pieces, and rebuilt the apparatus in a more streamlined configuration. Great results. Everything seems to be working the way it should have the first time. Still not entirely sure what the problem(s) was, but that isn't what's important now. The heatsink for my laptop arrived yesterday so I've been busy with that this morning, but I'm going in this afternoon to get back to work on this with renewed hopes. I may not even need an amplifier at all. It seems now (as I had originally hoped) that the wave form generator with properly matched inductors may give me the drive signal I need, just as that techmind link suggests. For whatever reason yesterday was just one of those days (in a good way for once).

    I won't be ordering the amp till Monday at the earliest so I'm hoping to find out this afternoon that I don't even need one. I'll measure my p-p output with the inductors I've got. From the preliminary tests yesterday it looked promising.
    Last edited: Feb 6, 2016
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook