Making an Infrasonic Microbarometer to Listen to Ocean Sounds From Miles Away

Amateurs have frequently followed in the footsteps of professional scientists to build and refine DIY sensors to detect natural phenomena. Over time I’ve built a number of those myself. While there are hundreds of Sensor-based project examples, most of these focus on sensing a short list of classic phenomena: weather, astronomy, seismic, environmental, health, etc. Over time what has changed has been the sophistication of the implementations, from early vacuum tubes and hand made components of the past, to integrated circuits, MEMS sensors, and microprocessors today.

I thought it would be interesting to make a sensor for a less explored phenomena. I remembered reading a Scientific American article long ago about sensors, called microbarometers. These could detect very low-frequency acoustic waves that could travel great distances. The phenomena was first recorded when the Krakatoa volcano exploded in 1883. The pressure wave was so strong it was recorded on early barometers thousands of miles away in London. Twenty years later, in 1908, when another giant explosion, the Tunguska asteroid, it was once again detected on distant barometers. This infrequent natural phenomena related to the pressure waves of violent explosions was called infrasound, because the frequency of these inaudible waves were below the 20Hz limt of the human ear.

Infrasound remained a scientific curiosity until the 1950s when scientists were looking for a sensor to detect distant nuclear explosions during the Cold War. An unanticipated side effect of the research was noticing background signals that were not associated explosions or any other human activity. Eventually, it was determined that the Earth’s atmosphere is constantly vibrating with extremely low-frequency pressure waves caused by the violent turbulence of crashing ocean storm waves (now called microbaroms), weather systems, earthquakes, volcanic eruptions, and falling meteorites. Now called infrasound, these signals are far below the range of human hearing, yet they travel thousands of miles across the globe. These studies established the link between ocean wave activity, seismic motion, and atmospheric pressure fluctuations. By the late 1950s, these signals were recognized as a persistent natural phenomenon.

Today you can buy highly refined and sensitive professional-grade infrasonic detectors, but they cost thousands of dollars. This project describes the design and construction of a DIY version capable of detecting global atmospheric phenomena. The goal was to attain a high level of sensitivity, while keeping the construction relatively simple and inexpensive. The final system has two major components, a custom mechanical hardware sensor and a electronic hardware/software signal processor. Because the sensor construction is fairly involved, the emphasis in this Instructible is on the design and construction of the custom mechanical sensor hardware component using a combination of a special diaphragm-based pressure sensor and strain gauges sensors. It essentialy uses a specialized combination of microphone and barometer designs specifically tuned to low frequencies.

My primary experience is on software and electronic hardware side. While the sensor mechanical hardware appears relatively simple, it needs to be based on a principled design to achieve the high sensitivity. Not being a mechanical engineer, I thought this would be a good opportunity to test using ChatGPT as an expert assistant. That turned out to be extremely helpful. It was also my first experience using AutoCad for the mechanical components and in using small foil strain guages.

My hope is this Instructable will provide the design for a sensitive infrasonic hardware sensor that can be duplicated by other Makers who will then add their own version of electronic sensor processor.

Aluminum Disk (Diaphragm), 4" diameter, 1/32" thick, 6061-T6 aluminum

Aluminum tube, 3" diameter, 1.5" tall, 1/16" wall thickness

Aluminum Diaphram Clamp Ring, 4" diameter, .25" thick, 6061-T6 aluminum (see step file)

Aluminum Base Plate, 6"x6", .25" thick, 6061-T6 aluminum

8 × 6-32 Stainless Steel Screws, 2" long

Strain Gauges (4) 350 Ω Foil gauges: BF350-3AA 350ohm High Precision Pressure Resistance Foil Strain Gauge, https://amzn.to/4v076ar

Minature 28 guage, 4-wire shielded cable: https://amzn.to/4lWdCe8

Instrumentation Amplifier Moduile: AD623 AD620 Instrumentation Amplifier Module

Passive components (resistors, capacitors) as required

Infrasonic sensors have some special requirements that make their design different from the classic a barometer and an accelerometer. A traditional barometer measures absolute atmospheric pressure using a sealed reference chamber and a relatively stiff diaphragm, so it responds mainly to slow weather changes. An accelerometer, by contrast, measures motion (acceleration) of the device itself and becomes insensitive at very low frequencies. Microbarometers sit between these: like a barometer it uses a diaphragm to sense pressure, but instead of a fully sealed reference it employs a leak or vent that acts like a high-pass filter, allowing slow pressure changes to equalize, rejecting long-term drift while remaining sensitive to small, rapid pressure fluctuations in the infrasonic band. In this way, it functions as a very low-frequency microphone, optimized not for sound we can hear, but for atmospheric pressure waves with periods of seconds to minutes.

I found several DIY infrasonic sensors on a web search. They were all based on using a large diameter woofer speaker for a diaphragm and sensor. While this will work for the more intense types of nearby infrasonic events (thunder, explosions, etc.) because they are basically a velocity based transducer. So they fall of quickly at lower frequencies.

A review of the range of professional grade sensors showed that they fall into three classes that differ in the diaphragm and how they measure the diaphragm deflection. The most sophisticated professional sensors use capacitive or optical sensing. But they require extremely precision tolerances, down to the micron level, and a thin fragile diaphragm. They are the most sensitive, but also the most expensive, in the multi-thousand dollar range.

MEMS pressure sensors are at the other extreme. They are redialy available and much less expensive, but they have a small diaphragm diameter that is not matched to the lowest infrasonic freqencies. They have been used in DIY projects, but limited to detecting nearby events, such as thunder storms.

The third commercial option uses strain guages on a thicker metal diaphragm, priced in the $200-$500 range. They are a pressure-coupled infrasonic sensor, not a sealed barometric sensor. This puts them in a middle class that is less sensitive but more practical for a DIY version than capacitive, and more sensitive in the frequency ranges desired and sensitivity than MEMS sensors. This design strategy using strain sensors on a metal diaphragm in a vented cavity, shown in the figure above, that will be used on this project.

The core mechanical sensing element consists of a flexible aluminum diaphragm that sits on top of a small, hollow cavity. Changes in atmospheric pressure cause the diaphragm to bend slightly, producing a tiny strain that can be measured using bonded strain gauges. The slow daily variation in atmospheric barometric pressure is thousands time larger than the pressure change from the infrasonic signals. A small vent is added to the cavity t0 allows the diaphragm to respond mainly to more rapid infrasonic pressure changes rather than the slow daily barometric pressure drift. The small cavity leak actually acts as a pneumatic RC filter.

The cavity is designed to have a resonant frequency of ~200Hz, which is safely above infrasound signal frequncies. The thickness of the disk has to be designed to provide deflections that are detectable by strain guages. Taking these into account, led to the following component sizes:

An aluminum disk, 4" diameter, 1/32" thick, 6061-T6 aluminum acts as the sensor diaphragm.

This disk is centered on top of a hollow aluminum tube, 3" diameter, 1.5" tall, 1/16" wall thickness.

The aluminum tube has a threaded hole on side for a 3/16” Sintered Bronze reference vent (or capilary tube)

The tube sits on top of an Aluminum Base Plate, 6"x6", .25" thick, 6061-T6 aluminum.

The diaphragm disk is edge-clamped on top of the tube by an aluminum clamp ring, 4" diameter, .25" thick, 6061-T6 aluminum (see step file). This clamp ring has 8 equally space clearance holes for 2" stainless Allen head screws that screws into 8 corresponding threaded holes in the base plate.

Once these screws are tightened, a hollow cavity is formed under the diaphram that has a volume of ~100cc. The clamped diaphraghm becomes a pressure-sensitive membrane.

The strain guages will be added in the next step.

Four tiny foil strain gauges will be used to measure bending strain on the diaphragm from the infrasonic signals. The tiny guages have two wire leads, as shown in the picture. If you have not installed foil guages I suggest you check out one of the many good instruction sites online. It is suggested to do some surface preparation on the diaphragm first. Lightly polish the surface with600–1000 grit sandpaper to remove oxidation. Then degrease it with alcohol. Handle the gauges with tweezers, since your finger oil can contaminate the bond.

It turns out the region of maximum deflection is at the edges of the diaphragm. So it is important to place the guages approximately 1/2" from the edge of the diaphragm for maximum sensitivity. You can use SuperGlue to attach the guages once carefully placed. But use a very small drop of glue. Place two of the guages so they are lined up long-way on opposing circle radius, as shown in the drawing above. These will sense the bending strain on the diaphragm. Place the other two guages at right angles, perpendicular to the radius. These will compensate for temperature shifts.

As described in all the online instructions, and shown in the picture above, it is wise to put a tiny loop in the wire leads to allow flexibility. The guages have tabs that are placed on the edge clamp ring for lead attachment. This is a very delicate step, so take your time. You probably have extra guages, so don't be surprised if you make a mistake on one of your first placements.

Once the guages are secured and soldered to the tabs, small wires can be used to electrically connect them to form a resistive Wheatstone Bridge (see Wikipedia for explanation). Following the diagram, two branches of the bridge will go to the exatation voltage and the other two are the sensed output that will go to the the output ampliier. Be sure that the guages are wired according to the diagram above.

At this point you can use a 4 conductor shielded cable to connect the guages to the excitation and sensor amplifier (next step), or you can put the amplifier on top as shown in the next step.

Congratulations, you've finished the most difficult part of the construction.

Strain gauges require connecting their outputs to instrumentation amplifiers because they produce extremely small, differential voltage signals that are usual ly masked by noise and high common-mode voltages. Instrumentation amplifiers are specialized, high-gain differential amplifiers designed to extract these tiny signals (often in the microvolt or millivolt range) from a Wheatstone bridge and amplify them for use by data acquisition system. An instrumentation module is used in this design to save some implementation complexity. A connection diagram for the AD650 is shown above, so just follow the instructions about connecting in strain guage input signals and power supply. The output can be connected to follow processing in an Arduino or similar signal processing circuits. I found it convenient to mount the AD623 directly above the strain sensors, as shown in the diagram in the introduction. I also found it helpful to place the finished sensor in a metal box to provide addition RFI shielding, as shown above.

This completes the basic implementation of the infrasonic sensor. Because of the overall complexity of this sensor construction, this Instructable is leaving the follow on options open in the hopes that it will lead to creative infrasonic sensing projects.

Sensor Articles

A robust, low-cost and well-calibrated infrasound sensor for volcano monitoring https://www.sciencedirect.com/science/article/pii/S0377027319302665

How to detect low frequency acoustic waves in the atmosphere. https://www.belljar.net/microbar.htm?utm_source=chatgpt.com

Wheatstone Bridges, Strain Guages and Instrumentation Amplifiers

https://www.electronics-tutorials.ws/blog/wheatstone-bridge.html

https://www.allelcoelec.com/blog/A-Complete-Guide-to-the-AD623-Instrumentation-Amplifier-IC.html

https://www.engineering.com/why-strain-gages-need-instrumentation-amps/

Instrumentation Amplifiers

https://en.wikipedia.org/wiki/Instrumentation_amplifier

https://docs.cirkitdesigner.com/component/2a9ef085-a643-41d6-9cd5-766a10137208/ad620-transmitter-high-precision-microvoltmillivolt-voltage-amplifier

Commercial Sensors

infrasound sensors are available from Hyperion, Chaparral, and CET. These use piezoelectric, differential

capacitive, and linear variable differential transformer (LVDT) sensing technology, respectively [3-5].

B. J. Merchant, "Hyperion 5113/GP infrasound sensor evaluation," Sandia Report SAND2015–

7075, Sandia National Laboratories, 2015.

B. J. Merchant and K. D. McDowell, "MB3a Infrasound Sensor Evaluation,"; Sandia National Lab.

(SNL-NM), Albuquerque, NM (United States) SAND2014-20108; Other: 547438 United States

10.2172/1165050 Other: 547438 SNL English, 2014.

G. W. Slad and B. J. Merchant, "Chaparral Model 60 infrasound sensor evaluation," Technical

Report, pp. SAND2016-1902, 2016.