Insect Eavesdropper: Creating a High-Gain Parabolic Microphone

My undergraduate minor was Evolutionary Biology and I always particularly loved my insect classes and learning how they evolve and coevolve with plants, animals, and predator/prey insects.  I've been a big fan of insect sonograms and love the sound of crickets, especially. 

This project builds a high-gain amplifier with a piezo microphone on one end and earphones on the other.  In between the magic happens.  In fact, if you replace the piezo mic with wire wrapped around a ferrite core you can hear magnets in your wall.  Or if you hammer a nail halfway through your wall and place the microphone you can hear conversations in the next room as clear as a bell (ahem, not that I've tried it).  I'll offer suggestions for improvements and other uses at the end of this instructable.

Step 1 Ingredients

I laughed when I used the word "ingredients" instead of "Bill of Material" or "Parts."  Not sure why.  Maybe I need to get out of the house more often.

Start with the following items and feel free to substitute similar items with what you have handy:

• All detergent container (or similar...just get the shape close to a parabola)
• Small (5"x5") pressboard (you can buy these in huge sheets for minimal scratch at homedepot and elsewhere)
• piezeo microphone like the one at radio shack and elsewhere...
• a few feet of 0.25" x 0.170" vinyl tubing (homedepot) [optional]
• some cardboard (you *can* substitute the cardboard for the pressboard if you want)
• headphones
• felt pads

Electronic Components BOM

• LM358 Single-supply OP Amp
• LM386 Audio Op Amp
• 10k OHM resistor (x3)
• 1k OHM resistor
• 0.1uF capacitor (x2)
• 100uF electrolytic capacitor
• 220uF electrolytic capacitor
• 470uF tantalum capacitor
• 1uF tantalum capacitor
• 100k OHM linear potentiometer
• 1/8" audio jack similar to here

Supporting items

• Soldering iron, solder
• hot glue gun
• paint
• tools
• compass (as in the thing you make circles with)
• Schematic cad software if you plan on changing things
• copper-clad or perfboard for pcb

That should be it. 

Comments

[lasagna]

Good idea!

I had trouble seeing the schema clearly so I've enclosed an enlarged one.

Your comment about battery life got me thinking... Since you're just driving an earphone or two from it, why not skip the power amp, which is what's eating up the power.

Try this experiment, connect the earpiece to pin-3 of the LM386 and see if you get a decent volume. Worst case is you change the 100k volume pot to a 1Meg one (the LM352 is good for another 10-20x gain). The 9-v should then last you a few weeks.

[nevdull]

For a better view of the schematic, click on the "i" in the upper left-hand corner of the pic and select "original image." It's pretty big and should be large enough for you to work with. If not, let me know.

As for your ideas, they sound great. I had originally designed it to work with an 8ohm speaker in a small handheld using the electret mic like the wongo-thingo in a star trek medical tricorder. Ha. I switched to headphones at a later stage, but you make a good point. I'll give it a burl and see how much output I get with just the power amp and the volume pot change-out sounds like a good plan B.

Thanks for the advice!

[foxwoodfarm]

I can't get the files you attached to this instructable to open. What program opens them? Also, I am using a small tape recorder instead of building the circuit you show on this page. I tried the 3'x5' coils of wire but the unit is basically a short circuit to the mike input of the recorder. Shouldn't it have a resistor to match the impedance of the input?

Jerry

[nevdull]

Hi Jerry,

The files that are actually attached to this project are *.brd and *.sch (board and schematic) files that are meant to be opened with CadSoft's Eagle Schematic and PCB Design software. That's what I use for my schematics and to lay out my PCBs.

Now about your impedance matching question...I apologize in advance if this is a review for you, but it may help other readers considering a clever project such as yours.

The impedance of a microphone describes how  much the mic resists the flow of an AC signal. Low-impedance is ~ 600 Ω.  Medium-impedance is typically 600 - 10K Ω while high-impedance ranges over 10K Ω. So, the first step is to determine what sort of mic your recorder is designed to be used with. This will establish what sort of impedance matching circuit you'll need, or even if you'll need one at all. You really have one of three choices that I can think of:

1. Dynamic microphone
2. Condenser microphone
3. Electret microphone

Dynamic Microphone
These are often found in PA systems, hi-fi, and recording apps. When alternating pressure (that is, compressed air from a sound wave) hits the diaphragm, it induces a voltage across the leads as the voice coil accelerates through the magnet's magnetic field. These mics perform well over a broad frequency range and have a low impedance output. Some, however, have switches that allow for setting high- or low-impedance output.

Condenser Microphone
These mics offer exceptionally crisp, low-noise sound and are used for high-quality sound recordings. It consists of two charged plates (supplied from an external power supply) acting essentially like a sound-sensitive capacitor. A low-noise, high-impedance amplifier is typically required and also to provide low output impedance.

Electret Microphone
This is the microphone I used in this design. They are ubiquitous, low-cost, and easy to design with. They can be viewed as a specific class of condenser mics in that they have two plates, but one of them is an electret and provides its own charge. Power is normally supplied through a resistor (see my schematic). They respond best to mid- to high-frequencies and not well at all to low or bass frequencies, making them most useful for voice apps. Also, as the charge on the electret degrades over time, so does its performance.

In general, it is advisable to connect a lower-impedance mic (source) to a higher-impedance input (destination, or just "load"), but inadvisable to connect a high-impedance mic to a low-impedance input. In the latter, a serious attenuation of the signal is likely occur. The commonly accepted heuristic is to allow the load impedance be about 10 times the source impedance.

Impedance Odds and Ends
So, is impedance matching necessary? Not typically when it concerns a low-impedance source connected to a high-impedance load. In the past, impedance matching was pretty important for power transfer reasons but with modern circuits and audio amps, what is most important is voltage transfer. For the most "bang for your buck" in voltage transfer, it was figured that the load should have an impedance of at least 10 times as big as the source, hence the de-facto standard I mentioned above. This is called bridging. Without bridging, matching same-impedance source and load results in about 6dB of attenuated signal loss, which is acceptable for most applications. However, in the opposite case of matching a high-impedance source to a low-impedance load, the signal attenuation would be equal to:

dB = 20 log₁₀ * R_Load / (R_Load + R_Source)

As you can see, the signal loss is significant in this case.

My suggestion is to try to get the specs for your recorder, if you don't have them. Determine what your source impedance from your chosen microphone will be and what the load impedance of the recorder input is. If you are "rolling your own" microphone like with the 3'x5' coil, calculate the impedance of the coil using the area and length of the wire and its resistivity using:

R = ρ(L/A)

Where ρ is the resistivity of copper, L is the length of the wire in meters, and A is the area of the wire (ie its AWG) in m². To keep you from looking up stuff from different places, here are the specs for your 3'x5' 26# coil I mentioned in this instructable. You may have to tweak some of the values below as I'm forgetting some of my spring physics this morning. :)

A = 26# AWG Area (m²): 252 * 5.067 x 10^-10 m²
L = ~ turns * π * D where D = 0.9144m → 6 turns ~ 17m, 8 turns ~ 22m (gross approximation: measure your wire by hand for best results)
ρ = 1.72 x 10^-8 Ωm

To ballpark it without doing the above, take note that 26# copper wire has ~ 41 Ω per 1000 feet so you're looking at something like 2 - 3 Ω if I did my math right (don't hesitate to double check me). Given this, if your recorder input/load impedance is 20-30 ohms minimum you probably don't need to impedance match.

Hope this helps. Don't hesitate to post back here, email me here or jamesbl at research.cs.colorado.edu. And of course, if you build your project be sure to post it here on instructables and let me know. I'd like to see it! Good luck!

Kind Regards,
Gian

[foxwoodfarm]

Gian
Thank you for all the information. I will have to take all this information and do some adjusting to my project. One thing I want to make is a devise to change various natural sounds that and above or below the range of human hearing and convert to so I can hear it.

Thank you, Jerry

[bobzjr]

foxwoodfarm,

I found your idea about shifting sounds into hearing range very interesting. Could we create some stegonagraphy here? No pun intended.

Back in the early 90s, I had a program that could embed files of any extension into a faxed page. The faxed page would look like snow - noise - giberish. Receiver would think that it was an error reception. But, if you knew what it was, you could simply scan into your computer and the file(s) could be recomposed perfectly. I wish I could find the program - believe I have a copy somewhere lost in my attic.

Anyway, could we do something like this with sound? That is - purposely float a message(s) in the inaudible range and then manipulate into useable form. Seems like would work best in the "hiding in plain sight" scenarios...

Could cell phones "hear" this inaudible range?

[nevdull]

bobzjr,

If I'm reading you right, you can modulate many types of waves outside of the 2Hz-20KHz audible range (aka frequencies, aka carrier wave) with an input signal in the audible spectrum as a communication mechanism (as we do with FM (Frequency Modulation) and AM (Amplitude Modulation) radio signals). By necessity, the carrier wave is usually more energetic than the "message" signal modulating it.

Something fun you might try is to take an AC signal in the audible range shift it up in frequency by mixing it with the signal from, say, a high-frequency oscillator. Physics states that when two frequencies are mixed you always get two signals: one at the sum of the frequencies and one at the difference. These frequencies are heterodyne pairs and the process is heterodyning, a common solution to shifting signals up or down in frequencies.

If you have a random mess of logic IC's laying about your workshop, you may even try doubling the signal frequency by rectifying the AC with some diodes then shift the frequency back down into audible range with something like 4017 decade counter IC.

All kinds of fun you can have! :)

Gian

[nevdull]

Hi Jerry, It sounds like you definitely have a fun project in mind. I'm not very knowledgable about infra-sonic transducers, but I have some experience with ultra-sonic transducers and frequency shifting and heterodyne frequency mixing to convert ultrasonic frequencies into the human-audible spectrum. I made an ultrasonic version of my "Insect Eavesdropper" to pick up cricket chirping in the > 20 KHz range (mid-band ~ 40 KHz). I never published it here on Instructables.com, though, as I got busy and it just sorta dropped to the wayside. I've included two pictures, one of the complete schematic and one of an unrouted PCB board in hopes that these might give you some brainstorming ideas for your project.

You've got a very fun and interesting project ahead of you from the sound of it. As always, best of luck to you!

-Gian

[nevdull]

oh, p.s. if these images are too small to make out the details and you have the Eagle schematic and PCB layout software, if you like, I'll be glad to send you the actual schematic and board files you can load directly into the layout editor.

Cheers!

[tomaradze]

On the bill of material you wrote LM368 but it's actually a LM386. The link is correct, though.

[nevdull]

Sharp eyes! Thanks for noticing that. I've corrected it in the instructable now.

Cheers!

[DisappearingOak]

Hi, I'm making this right now, working with the black circuit diagram on my phone. could you tell me if polarity matters for the 0.1uF caps, they don't seem to be marked in the photo nor in the two caps i bought? also, for the 358 chip which is the right side up? sorry, beginner here.

[nevdull]

Hi @DisappearingOak,

No worries about being a beginner. We all start somewhere and part of the fun of being a beginner is all that is open and new for you to explore!

Re: the 0.1uF caps, they have no polarity, so you can solder them in any way you like. As for the 358 IC, there should be a circle or dot (or maybe some other symbol) in one of the corners. That symbol indicates that when you orient the chip with the circle/dot in the top left corner, the top left pin is pin 1.

Hope this helps and good luck with your project!
-gian

[bears0]

i had a really old mic and i plugged it into my computer and i amplified it by like 30 db because it was really quiet and i could hear stuff on the radio, unfortunately it had a lot of static so you couldnt really understand what they were saying.

[Tommunbeig]

I went to Teacher Training College as a very mature student when I was forty to do Technical Teaching course. I didn't stand out quite as I expected for  Ravenscraig Steelworks and Rolling Mill and a few ancilliary mills had been closed by The Great Thatcher so that there were more of my age than fresh faced youngsters just in from School. I enjoyed my four years having free run of several machine shops and laboratories. A second year metalwork project was making a lamp and special tools and patterns such as could be made in schools. I made a wall mounted lamp from polished aluminium shaped into a parabola along its length, held in perspex end mounts with large hollowed aluminium screws I cut on the lathe so that it would give a good light from two bulbs to illuminate my drawing desk. It now is above my bed. And there is where I would appreciate guidance from experienced parabola designers. For the lamp was not nearly as good at focussing light as it was at directing heat. This I found out after accidentally switching the lamp on when I was asleep and waking up part roasted and asking to be turned over so the other side could be cooked. Sound is a much longer wavelength than heat and travels slower whereas light and heat travel at the same speed differing in wavelength red being at the longer wavelength. Is there any way I should have corrected for the difference in the calculation of the parabola's dimensions to make it more light effective?

[watergeorg]

A parabola allways follows the formula y=x↑2, no matter if it is used for light or sound (a wave is a wave, no matter what).
The ends of your constrution do not look to me as a parabola. If you know how to draw a function using the formula above, it would be easy as 1-2-3 to have the right curve you are looking for.

Good luck!

[CobaltBlue]

May I suggest scissors? Another tool which is great for poly bottles and even things as heavy as paint pails, is the PVC pipe cutter intended for plumbing. With a can-opener action you can get smooth edges without sawdust or melting.

[nevdull]

That's true: you do get a good bit of melting when cutting the container and the pipes. That looks like a good thing to try Thanks for the tips!

[qubic]

The whole destructive-interference thing sounds really hard to tune correctly, since you'd have to get both the frequency and the phase exactly right. Have you considered a notch filter? I found a potentially useful circuit at http://www.kennethkuhn.com/electronics/ (the Hall Network); it's a notch filter you can tune with a single potentiometer.

[ranaakamarth]

After reading all the uses for it I really want one! It is so cool all the normally undetectable forces you can "hear" with this

[LWS]

Hello, Could you recommend a high gain amp kit? Thanks.

[nevdull]

While I'm sure there are those kits out there (maybe Jameco or Velleman?), I am not familiar with any. Sorry I couldn't help!

[Mark Rehorst]

For anyone interested in going further, check out some of the work of Dr. Rex Cocroft from U of Missouri. He discovered that bugs make much more sound that we can't hear by vibrating the plants that they are sitting on. He devised several ways to record that sound including using accelerometers and phonograph cartridges attached to plants. Here is some of his stuff:
http://www.npr.org/templates/story/story.php?storyId=5424281

Do a web search and you can find a lot more.

[nevdull]

That's very awesome. There's also a book that's available called The Songs of Insects by Elliott and Hershberger that includes sonograms of tons of crickets and grasshoppers along with a CD so that you can hear each of them. It`s well worth it if you like this sort of thing.

Also, another point from your comment, is that an addition of an ultrasonic piezo with driver and receiver might be a good area to investigate. I've built the ultrasonic driver and receiver but haven't connected it into the parabolic design before. Would be a good project!

Thanks!

[jeff-o]

So, am I correct to assume that the better the parabola, the more amplification you can expect? Would a salad bowl work even better?

[nevdull]

Yep, the better the parabola and the more closely you've put the mic to the intersection of the symmetry axis, the more focused the sound waves will be on your microphone, allowing you to pick up fainter sounds than without the parabola. I think a salad bowl would be an excellent candidate :)

[jeff-o]

Cool, that's what I figured. Hmmm, I wonder if my 3-year-old would like a bug listener? It would have to have two headphone plugs, of course...

[nevdull]

I had considered putting the microphone on a metal rod and placing it in the ground and using RF wireless (say, sub-ghz range) to transmit the sounds back but my transmitter/receiver pair wasn't fast enough for real-time sound (maybe a sample-and-hold circuit?). The reason I had thought of this in the first place is that crickets get deadly silent when you walk into the middle of them with your Doctor Destro Insect 'Eavesdropper.

My brother and I were on the front patio pointing it at cars in the dark last night and trying to guess what type of car it was. I could hear snow tires, squeaky belts, etc.

The concern I would mention re: 3-year old is that the volume/squelch thingo on it can belt out some volume, especially at a full 9V. if you have the volume up high and listen to something close to the microphone it will clip slightly and you'll get a stuttering sound. It doesn't show up under normal use...it's just feedback. I'd consider putting a volume limiter or a window comparator to only allow volumes within a certain range.

[jeff-o]

Well that's why I'd have two headphone ports. One for her, and one for me - and I'd be controlling the volume!

You make a good point about frightening the bugs, though. She is less than sneaky when it comes to that sort of thing!

But a remote listening station sounds like a really cool idea! I wonder how it could be best implemented...

[nevdull]

I've had some thoughts about this remote listening device...if you're interested in walking your (our) wits about a design shoot me an email. I've seen plenty of video over RF sub-gig so I can't imagine that it'd be too tortuous.

[jeff-o]

Nah, it should be easy enough.

I can't spend too much time on design at the moment though. I've just started up something epic and it'll require most of my attention...

[nevdull]

Ah nice! I look forward to seeing this epic "something" sometime in the near future (if it's an instructable, that is).

[jeff-o]

It definitely will be an instructable - I just hope I can pull it off!

[nevdull]

i have confidence in you...you've pulled off some pretty spectacular things so far, oh ye of little faith :)

[jeff-o]

Thanks. :)

[lukeyj15]

have you guys heard of infrared headphones?
something like this but cheaper
http://www.dse.com.au/cgi-bin/dse.storefront/4c99a3e9019f978c2741c0a87e010710/Product/View/C4270

[nevdull]

That would require line of sight, which I was thinking that the mics on posts would probably be down in the weeds (say, a foot off the ground or maybe a little more), but that's not a bad idea at all. It would make the wireless reception a non-design issue and seems like it would be worth throwing another foot or two on the post with the IR transmitter on the top. I'm guessing it connects via an 1/8th jack or something?

I'll keep my eyes open for a cheaper pair of these things. Thanks for the heads up!

[rimar2000]

Very interesting. Your device is perfect to listen wood-worms!

[nevdull]

Now that would be interesting to hear!

[cypher073]

This is awesome. I know what I'm doing when I get home :)