Project: Audio Induction Loop Receiver (Part 1)

In daily life, you might have come across signs like this, with a picture of an ear, and the letter T on them. Maybe you didn’t think any more about them, as an average person with perfectly good hearing – after all, they are an assistive technology intended for the hard of hearing.

But as an electronics and signals enthusiast, this really piqued my interest. Everyday, we are literally bathed in signals we don’t notice, and don’t understand! I’ve seen these logos on auditoriums, in trains (e.g. picture above), in train ticket windows amongst other places. I’ve always wondered how these things worked, and what it would take to receive such transmissions. After some brief research, I got some of the answers I was looking for.

Audio Frequency Induction Loop Background

The audio frequency induction loop is a system by which audio can be transmitted to hearing aids through the use of magnetic fields. This system does not modulate the signal on top of a subcarrier, instead relying on an alternating magnetic field in audio frequencies. This system appears to have come about due to the parallel development of telecoils in hearing aids, which were designed to pick up the stray magnetic field leakage from telephone handsets to improve intelligibility for hearing aid users. The T on the signs indicate to users to switch over their hearing aids manually to the telecoil operation position.

By leveraging these telecoils for longer distance transmission of audio, it is possible to transmit audio to hearing aids without relying on a bulky receiver, and improve the quality of the audio over that picked up by the integrated microphone. Systems to transmit the audio can be as simple as a loop of wire hooked up to a regular amplifier.

Unfortunately, such a system also causes electromagnetic interference by spewing moderate to high levels of electromagnetic fields (as that is how it works). It is also vulnerable to electromagnetic fields, which cause interference and marginal audio quality.

What Does That Mean?

You know when people say they wished they had superpowers? Well, and I mean this with all the respect, those with telecoils in their hearing aids actually have one that regular humans don’t!

Normal hearing relies on detecting the compression waves in air. When audio is generated by a speaker, the speaker is converting the electrical waves into magnetic fields which then drive the cone via the voice coil motor to create the compression wave.

In a telecoil system, they are shortcutting this, and instead, users with hearing aids can directly perceive alternating magnetic fields the audio range. The induction loop system exploits this, essentially acting as a loud magnetic-field only speaker …

Because they can perceive alternating magnetic fields in the audio range, they can also “hear” many annoyances which interfere with the audio – such as emissions from switching power supplies, radio frequency transmitters, inverters, etc.

Do It Yourself Receiver

Given the standardization of induction loops here, I would have thought receivers would be cheap, plentiful and designs widely available. Strangely, this was not the case. It seems that the system is generally relegated to assistive uses, and thus non-hearing aid users don’t get to benefit from the system.

My interest was to not only access such telecoil services, but also to try and perceive the world in the electromagnetic audio region. Think of this as an artistic venture, similar to how the guys who implant magnets in themselves do so to try and perceive the world’s electromagnetic fields – except this one is generally painless.

Phase One Design – Passive Coil

Seeing as it’s an alternating magnetic field, it should be fairly simple to pick up if the field is strong enough. I decided to grab any scrap beefy inductor with many windings (to improve the inductance) and placed it in series with a blocking capacitor (to prevent any phantom power flow which would saturate the inductor). Attach some wires and a 3.5mm plug and you’ve got something basic.

The trick was trying to find a highly amplified sensitive input. I tried mic inputs, and others. But even plenty sensitive inputs, even on the Zoom H1 (which I used to do almost all of my recordings) were very noisy. They are so hissy that I won’t even bother uploading any of them.

It was clear that some amplification was needed.

Phase Two Design – Amplified by Rail-to-Rail Op-Amp

This time, I decided to get a little more sophisticated, opting for the use of a rail-to-rail opamp to provide amplification (as I abhor trying to build, and carry 15v dual-rail supplies). In this case, the device consisted of a three-AAA carrier with switch, the board with the circuit and the 3.5mm audio plug. The whole device was encased in glue to improve resistance to conducted stray charges which affect performance.

It looks pretty simple, but stupid me made so many mistakes along the way. It’s clear I didn’t do any analog electronics for a long time and I’ve literally forgotten some important practical considerations.

I started with the Microchip MCP6273 2Mhz Rail-to-Rail Op-Amp, a 1200uH inductor, and an inverting amplifier design with virtual ground provided by a resistive divider. Trying to save some power, I used some fairly high resistances in my resistive divider, which resulted in an unstable virtual ground that caused oscillation. Get it together Gough! You can’t screw up an inverting amplifier!

After I scoped that one out, thanks to the Picoscope, I still had oscillation of a different sort. I decided to make the gains adjustable by trimpots, and I found the oscillation was pretty bad for most gains. Why? Why!

Well, the Picoscope again gave me the answer – the coil is a very efficient receiver of broadcast AM transmissions, and these signals were getting into the opamp, distorting non-linearly, and creating the resulting tones. In fact, I only recognized that when, in spectrum mode, I could see the AM sidebands. Then it hit me. So, uh, remember to band-limit your signals. Just because you’re interested in one sort of signal, doesn’t mean that you won’t get others “leaking” into your receive chain. I decided to go for a simple R-C filter (which doesn’t do much, as the cut-off was quite high due to a lack of spare components), and you’ll see later, I managed to stuff that up too.

But it still didn’t satisfy me. I tried to push the gain, and at some point, it would just collapse and die. Then I remembered the rule of Gain Bandwidth Product. You would think that the 2Mhz rail-to-rail opamp is fine for an audio frequency project, but alas, when you want 500-1000 fold gain, then the bandwidth drops dramatically. The bandwidth is given for unity gain, duh!

Therefore, it is much better to go for a two-stage construction, with the first opamp doing some of the amplifying, and the second doing some more. But I only had a limited number of them at the time, so I decided to conserve.

Then … after having sorted through all of that, I found out (the hard way) that the circuit was very sensitive to component choice. A design I breadboarded just a moment ago, built using supposedly identical components, was not functioning properly. As it turns out, I was pushing the opamp so much to its limits, that the other unit I picked to build onto the veroboard just didn’t have the same characteristics.

In the end, I ended up with something that looks like this. But then I realized, I made a royal goof with the R-C input filter, which should have the 2n2 capacitor looped back to the virtual ground point. Ah … insert expletive.

But funnily enough, this design worked. And it did, in part, because the opamp was pushed so hard that its frequency response fall-off was acting as a natural filter! I think by doing this, I’ve managed to refresh a lot of the things I should have known – components aren’t as ideal as you would like them to be.

Phase Two Design – Audio Samples

I decided to carry this around with me, for a few days, as I went about my regular business and checked what the recordings showed. A lot of interesting sounds were received, with all of them provided as .wav files, as I dislike compression.

• Riding on a Bus – it’s likely the alternator on the bus engine is putting out these whines.
• A different bus – this one shows a strange pipping noise as well.
• In a car – a petrol car’s ignition system gives a tick every time the spark plug fires.
• Next to a lift – the lift inverter produces some rather harsh noises, but the background “hum” is endemic to the public announcement system at the train station.
• Passing train – a passing train seems to make a strange buzz, but only at certain carriages. It’s likely those carriages carry the chopper/motor drive circuitry.
• Under a power line – a hum, but not the sort I expected.
• A GSM 2G mobile – the familiar bipping of the slotted TDMA transmission envelope. How annoying!
• Alighting from a Warratah Train with Flashing LED – the LED drivers make an interesting noise.
• Riding along the rails, and again – you get to hear strange noises, some of them alternating, some of them steady tones. These, I believe, are related to audio frequency track circuit impedance bonds – I only found out about these when I stumbled across the NTSB presentation about loss of train detection in WMATA. Some others may be related to the RFID detection systems and their power envelopes.
• A pelican crossing – probably my proudest moment was when I decided to put the transducer up to the vibrating part of the push button padestrian crossing. Normally, when at a crossing with an audio recorder, you get this. Instead, now you can get the “signal” cleanly, without the ambient noise. Even better, if you employ DSP, you can even clean it up!

Of course, this is not all. By holding the device up to monitors, light switches, power supplies – you can tell if they’re on or off, and what sort of loading they are running. It’s even possible to hold this up to a phone’s earpiece and get a recording of the speaker audio (the basic purpose a telecoil should fulfil).

Time to visit some real audio induction loop systems … see that in Part 2!

Phase Three Design – Improvements and Fixes

For the third design, I decided to opt for a much higher (80Mhz) bandwidth opamp (which is overkill) and not bother with having the two-stage design I would have otherwise gone for. Component values for the R-C filter have been changed to narrow the filter response, and since I was out of 3.6v 3xAAA cell holders, I opted to go for USB, instead using an LDO linear regulator to hopefully remove most of the ripple noise. I left a gain trimpot in there as well, so I didn’t have to settle for fixed gain. The design looks like this:

Unfortunately, and as I had predicted, the RF noise emissions from the power bank seems to have an influence on this one, causing popping and clicking if you use a poor quality power source. However, it seems the quality power banks do make for a quiet result! I can always build a linear USB power supply if I need to, or adapt the connection.

Conclusion

Audio frequency induction loops are an assistive technology that allows for the broadcast of audio to hearing aids using magnetic field coupling. Receivers for these systems are not common, however, it appears simple to design and build your own.

In the process of attempting to do so, I’ve reminded myself of how many basic realities of components I have forgotten, and it’s been a re-education exercise to some extent. However, I did eventually achieve what I set out to achieve, through sheer persistence and logical troubleshooting. In turn, I have been rewarded with the ability to perceive alternating magnetic fields inside the auditory hearing range.

It’s important to remember that, as the inductors used are not designed for picking up such emissions, they are probably shaped to reduce stray field leakage which means a low signal collection efficiency which reduces the signal to noise ratio. Real hearing aids are likely to see better quality reception in that regard. Proper hearing aids are now starting to provide digital signal processing noise reduction on these inputs, thus continuous tones, and hiss noise is probably quite significantly reduced to make it more intelligible. Furthermore, they may operate with automatic gain control and bandwidth-limiting systems which would alter what they would perceive.

However, by having one of these devices, it is now possible to receive audio from such systems without the associated echo and room noise.