Infrasound Translator

Many people worldwide are troubled by very low frequencies (infrasound) that's impossible to hear but can (and does) cause problems.  The issues vary with individuals, but can range from no effect whatsoever right through to physical illness.  Some affected individuals will feel nauseous when the infrasound is present, others suffer sleep deprivation, and others may suffer anxiety attacks.  Conventional 'wisdom' says that we can hear from 20Hz to 20kHz, but this is simply the normal frequency range that is covered by the definition of 'audio'.  While only the very young can hear above 20kHz, many people can hear (or feel) frequencies below 5Hz.

The range of effects is wide, and contrary to popular belief, these effects are not limited to humans.  Large wind farms in Scandinavia have caused significant problems for mink farmers ^[ 1 ], with minks being born prematurely, deformed and/or dead.  I have been to an affected property in Australia, and have seen photos of chicks still in their shells, deformed beyond anything I have ever seen before.  These are not isolated problems - similar reports come from all over the world, and are most likely when there is a defined source of infrasound nearby (typically within 1 - 5km).

Wind farm operators and others who generate low frequency noise with machinery such as large ventilation fans (used for underground mines) generally manage to continue because legislation in almost all countries insists that noise measurements will be performed using A-Weighting.  This simply removes all the low frequency noise, and the measurement is completely worthless.  I have published some information on this topic, and also have a project for an A-Weighting filter - See Project 17 for the details.

There is a truly vast amount of information on-line, and this issue is being studied by people the world over.  There is little consensus, and the wind farm industry in particular fights every claim tooth and nail.  'Tame' acoustical consultants are always presented as 'expert witnesses' in the many court cases that have been held worldwide, and attempt to discredit anyone who claims that the infrasound is causing harm.

The whole subject of infrasound is mired in controversy.  There are countless experts who claim that people are not affected, and a roughly equal number who state the opposite.  The real issue is that if people are exposed to high levels of continuous low frequency energy, it ultimately must have an effect.  It's simply non-sensible to assert that people can't be affected, and doubly so when animals are affected as well.

In case you were wondering, the name 'Infrasound Translator' was used because it translates the normally inaudible low frequency energy picked up to a higher frequency so the effects and 'character' of the sound can be heard.

Because (by definition) we can't hear infrasound, there's no point trying to pick it up with a microphone and amplifying it.  It's easily recorded if one has microphones (and preamplifiers) that can get down low enough, but that doesn't help if one is investigating infrasound, or trying to track down the source.  This can be especially irksome, because the source of very low frequencies cannot easily be located by our hearing mechanism.

Some time ago, Silicon Chip magazine in Australia published a circuit for an infrasound monitor ^[ 2 ].  It works well enough, but it also requires a PIC that's programmed to do most of the work.  Although it has more options than the unit described here, it doesn't work as well in practice (yes, it was directly compared and failed rather miserably).

There are also many other circuits, some good, some bad, and some indifferent.  A web search will show countless variations, but the unit here is dedicated to the task, while many others are designed for recording only.  The vast majority cannot produce an audible tone that is related to the infrasound that's picked up.  Some are very elaborate, and one of the very best recording systems created so far was designed by me, and is now in production in New Zealand.  Because it was developed for a client, the details will not be made available on the ESP website (sorry).

Photo Of PCB Prototype Infrasound 'Translator'

The design shown here is (almost) unique.  Although a roughly similar design is referenced above, the unit described has facilities not offered by any other system.  In particular, you can use a meter to (literally) see the fluctuations of air pressure, you can send the amplified signal to a recording device (such as a USB microphone adaptor - with modifications), and you can listen to a frequency modulated version of the waveform.  By far the hardest part is the microphone, unfortunately.

Please note that the PCB shown is no longer available.  I have had a few enquiries, but not in sufficient numbers to warrant production of more boards.  If this changes I will add the board to the pricelist.

To be able to detect if there is a persistent LF noise, the entire frequency spectrum needs to be raised.  Some readers may have been bemused by the appearance of a voltage controlled oscillator as a project (see Project 162 for details), and this came about during the development of the project shown here.  If we can't hear the noise, all that's needed is something that moves the frequencies up the spectrum to where we can hear it.

This seems simple enough, and in theory, it is simple.  One of the biggest problems is finding a microphone that can respond to the very low frequencies.  Ideally, it will be able to pick up sound down to 1Hz or less.  The Panasonic WM61A electret mic capsule was pretty good - it didn't get to 1Hz, but most managed around 4-5Hz fairly well.  Unfortunately, these are no longer made, and those you can buy that claim to be WM61A mics are not - they are Chinese copies that are nowhere near as good.

Figure 1 - Infrasound Translator Schematic

The complete schematic for the translator is shown above.  Several of these were built, and it works exceptionally well with the right microphone.  Because there's rather a lot going on, it needs to be explained in some detail.  The microphone is seen at the top left - getting one that works well will cause some grief, but they exist.  In some cases, a 'lesser' mic may be able to be modified to improve its LF performance.  In general, the larger the electret mic capsule, the more likely it will have good low frequency response, but you will have to check and be prepared to buy several likely candidates before you find one that works.

The first stage (U1A) is a preamp, which has a gain of 11 (100k and 10k feedback resistors).  The output then passes through a low-pass filter (U1B) that removes frequencies above around 18Hz.  These cannot be monitored, because they just make an awful noise at the output.  Since the higher frequencies are not infrasound, they aren't part of what we are trying to hear anyway.  The filter has a gain of 2, so the total gain is now 22.

U2A is a variable gain stage, and the gain can be varied from 5.5 up to 101, giving a total gain range from 121 up to 2,220 (41dB to 67dB).  Note D5 from the +5V supply to C6+.  This diode is needed to make sure that C6 charges quickly, and without it the circuit can take almost a full minute to stabilise and start working normally.  This is because otherwise (without the diode), C6 can only charge via R11, R12 and VR1, which takes a long time.

U2B provides power to the LED, and detects the signal level.  If the signal is too high, it will turn off the LED.  The LED should flicker occasionally to indicate that the signal level is within the limits.  There is provision for a recording output, that can be fed to a suitable USB sound card.  Note that most require modification though, because the input coupling caps aren't big enough to pass the very low frequencies.  Although it's not shown, it's possible to use the 'translator' to convert a recorded sample to be made audible - disconnect the mic and couple the audio from the recorder into the mic input.

The last option on the upper section of the schematic is a connection for a meter.  This can be used as a visual indicator of infrasound.  Because the frequency is so low, most moving coil meters can move quickly enough to show the waveform being picked up by the microphone.  A centre-zero is needed because the meter is designed to deflect positive for an infrasound pressure 'wave', and negative for a rarefaction.  Since centre zero meters are now few and far between, TP1 is used to set the pointer of a normal 100µA moving coil meter to the middle of the scale.

The amplified signal now goes to the VCO (voltage controlled oscillator).  See Project 162 for the details - there's no point duplicating the complete description here.  VR2 is used to set the centre frequency, and it will normally be adjusted while the gain is at minimum, and set for a comfortable frequency.  Advance the gain until you can hear the frequency changing in response to the incoming infrasound.  A positive pressure increases the frequency, and a negative pressure (rarefaction) reduces it.

The triangle wave output from the VCO is converted into something more like a sinewave by the diode clipping circuit (D1 and D2), and then to a volume control so the headphones can be set to the desired level.  The headphone driver is an LM386 amplifier, and it can drive the 'phones to very high levels if desired.  Obviously this isn't recommended because it will cause hearing damage.

Ideally, the unit will be powered from a 11V (nominal) Lithium-Ion battery.  A 3S (3 series cells) pack will deliver 11.1V which is more than adequate.  You can use a 9V battery, but if you do so, the TL072 opamps should be changed for TL062 because they work better with low supply voltages.  Note that if you use a Li-Ion pack, you must use a proper balance charger - never attempt to charge any lithium chemistry pack without a balancing circuit.  Don't leave lithium batteries charging unattended - there have been any number of news stories to show what happens if proper precautions aren't taken.

There is provision for two DC inputs.  Typically, DC1 would be connected to the internal battery (via an off-board switch) and DC2 is provided so an external DC power supply can be used.  The voltage should be 12V, and it only needs to provide around 100mA so any adaptor you have handy will work fine.  Schottky diodes are used so that reverse polarity doesn't destroy the circuit.

To record the output, use a USB microphone adaptor and a PC.  There are countless USB mic units available, many for less than $10 or so.  However, the unit you get must be modified to allow a recording to get below 20Hz.  I can't provide specifics because there are simply too many different units available, but the general type of unit is sold as a combination microphone input and headphone amplifier.  Most are fairly easy to modify to allow very low frequency operation - it's mainly a matter of finding the input capacitor.

The cap is most often a surface mount ceramic, but since the translator shown has an output coupling capacitor the mic interface cap can usually be bypassed - it may not need to be replaced, depending on the adaptor itself.  It might be necessary to provide more attenuation for the recording output - as shown the output level will be around 250mV.

It's important to understand that the recording will not be calibrated.  This means that it will not be possible to extract the SPL of any signal(s) revealed in the recording.  Calibration is a comparatively expensive process at the best of times, but it's made harder with this unit because the signal is deliberately limited to frequencies below 20Hz.  Most microphone calibrators work at 1kHz, and that won't register at all.

This circuit has been proven in the field, with an affected family easily able to hear the frequency change in exact sympathy with their sensations of LF energy.  With this device, others (including me) were easily able to identify that a very low frequency signal was present, although I felt no sensations.  A potential problem may be assumed to exist if the tone from the VCO varies rhythmically.  Random variations will always be present because our atmosphere is never perfectly still.  Even the faintest breeze causes small localised fluctuations of air pressure, and the 'translator' will detect them if the microphone is good enough.

From the information that I've been able to gain during several years working with (and designing) low frequency detection equipment, so-called 'micro-baroms' (very small variations in barometric pressure) are not a problem for people, but this can change when the variation is steady, representing one or more low frequency tones.  These low frequency signals are often also accompanied by amplitude modulation, so the tone is not at a consistent level, but varies rhythmically.

As noted in the introduction, many people involved in infrasound research have found that there is a direct correlation between the presence of rhythmic infrasound and physical sensations - including nausea or other symptoms.  The key word here seems to be 'rhythmic', where there are one or more distinct frequencies present, and usually including amplitude modulation of higher frequencies.  These can be up to 20 or 30Hz, but the actual numbers are still being investigated by a number of very serious professionals.

I have seen enough evidence to accept that there are people who are genuinely affected by infrasound.  The corrupt 'consultants' who will use every possible trick to discredit people (some of whom have abandoned their homes because they can't live there any more) have to be stopped.  The continued use of the fatally flawed A-Weighting for sound level meters also has to be stopped, because the filter attenuates any low frequency energy so much that the end result is utterly meaningless.  A (clearly audible) rumble at 20Hz is attenuated by nearly 50dB by the filter, so it's no surprise at all that the readings taken with sound level meters don't show a problem.

Hopefully, this project will help anyone who is affected or who knows someone else who might benefit from being able to reveal that rhythmic very low frequencies do exist in their locality.  With the option to record the signal detected, it may help to prove that they are, in fact, not simply making it up (as is so often claimed).  Analysis of the recording is another matter entirely, and it may be necessary to submit the recording to someone who is experienced in acoustic analysis and can measure the frequencies of the recorded infrasound.

I would like to say that PCBs are available, but this is not the case at present.  Apart from the prototype pictured above, there are no PCBs left, and there is no detailed construction information available.  If anyone is interested, please contact me and it may be possible to work something out.  The design was produced for a customer, but he is no longer interested in pursuing the project.

1   Windfarms: 1,600 miscarriages | World Council for Nature
2   Infrasound Detector For Low Frequency Measurements March 2013, Silicon Chip Magazine
3   Waubra Foundation - A collection of infrasound related information ¹

Note 1: I have had considerable direct contact with one of the principles of the Waubra Foundation, so I must disclose this as a direct influence on my opinions about infrasound and its effects on people.

Main Index

Projects Index