DIY Stereo Microphones for my GoPro

A while back, I got a Sony ECM LV1 Lavalier Microphone, which is really two separate microphones in a common housing in a stereo configuration. On paper, at least, this is a good fit for my GoPro Hero 9 with a Media Mod, as it is capable of stereo recording. But every time I rode at highway speeds, the audio track would be full of pops and clicks, leaving the audio unusable. I made my own inline attenuator, which did drop the volume of the recording, but the audio still had pops and clicks at highway speeds. This had me blaming the mic, so I decided to build my own dual mic setup.

One goal was to have the ability to do stereo recording, like I could with the Sony mic. When recording inside a helmet, one could argue that the stereo effect will be minimal, but if GoPros and other action cameras can do stereo recording, why not do stereo recording? I wanted some distance between the microphones, with each on it's own side of the helmet to maximize the difference between what each mic records.

I also wanted a microphone that could handle a high volume of sound. It's difficult to find maximum SPL ratings for electret microphone capsules, but after some research, I found that the PUIAudio AOM-5035L-HD3-R is rated for a maximum SPL of 135 dB. This means that if what I'm hearing while riding isn't hurting my ears, the microphones should be ok. What I'm not sure of is whether that rating is how loud the sound can be before damage occurs to the mic, or if that's how loud a sound can be with good-enough audio quality. (See the datasheet for the AOM-5035L here.)

After ordering the microphone capsules, I wanted to house them in something that was long enough to make mounting the microphones easy. I also wanted to limit how much the wires could move near the solder joints to prevent cracking those joints. I decided to use a couple 1/4" copper plumbing fittings. The capsules did not fit initially, but were close. After a little bit of drilling, they fit well enough.

These capsules are the three-terminal variety with an internal JFET. This means the actual electret condenser element is internally connected between ground and the gate of the JFET. The external terminals are ground, the source of the JFET, and the drain of the JFET. You can find a good description of 2-terminal and 3-terminal electret capsules here. 

The data sheet recommends a source follower configuration, and so does the article linked in the previous paragraph. However, the GoPro Media Mod input circuitry seems to be setup for a common source configuration where the Media Mod provides the drain resistor and output capacitor for each audio channel, with the JFET being provided by each microphone capsule. In other words, the GoPro setup expects two-terminal electret microphones.

This means the recommended source follower setup is not an option. However connecting the source terminal directly to ground basically converts a three-terminal electret capsule into a 2-terminal one. This was my initial plan.

However, I didn't want to simply bridge the two terminals with solder. A common-source gain stage typically has much more than unity gain, and if that lead to clipping, I wanted to be able to add a resistor between source and ground to lower the gain. If the ratio of the drain resistor to the source resistor is one, unity gain can be achieved (or possibly less than unity gain.) More on this later. Also, having a wire for each terminal could mean wiring to source follower stages that I build myself in an external amplifier with output characteristics for using the line-in setting on the GoPro.

When wired together with the output jack, I wanted to be sure to be able to distinguish between the drain terminals of the left and right channels, so I chose yellow for left and red for right. Since I was starting off with the ground and source wires connected together, there was no need for color coding between the ground and source terminals of the two channels.  

After soldering, I added some hot clue for some mechanical support to the solder connections. I definitely want to avoid bending at those connections, which could lead to broken solder joints.

I then put each capsule into its "housing", filled the inside with more hot glue, soldered the wires to the output jack, and heat shrinking everything along the way.

I then positioned the bundled microphone setup along the inside of the front of my helmet with the left mic along side the front edge of the left cheek pad and the right microphone along side the front edge of the right cheek pad. I used the microphone from my communication system to hold it all in place.

When I first tested this setup, the audio was HORRIBLE - much worse than with the Sony Lavalier mic. The pops could be heard at anything above walking speed. I tried various things to make improvements - made shorter cable, closed vents on my helmet, etc. These helped a little bit, but still any audio I recorded was unusably bad at highway speeds.

I knew I had to reduce the gain to eliminate clipping. I knew that with a source-follower topology suggested by the data sheet, unity gain would probably be best. I tried various resistors between source and ground for each microphone, and using 5.1kΩ resistors worked well. 

When looking at the diagram below, keep in mind that the gain stage depicted is composed of elements within the microphone capsule, including the internal JFET, and components within the GoPro media mod. D, S, and G represent the labels on the solder pads on the microphone capsule. Rs is the 5.1kΩ soldered inline in the green wire for each mic. There's no way to know for sure what the circuit topology is in the GoPro media mod for its portion of each input gain stage.

If I were to build another one of these, I would connect the source pin and the ground pin with the 5.1kΩ source resistor as close to the capsule as possible for each mic, eliminating the need for a long green wire in the bundle. I might even use a chip resistor, soldering it directly to the source and ground pins. I would probably also shorten the wires a bit, with the output jack not sticking so far outside of the helmet.

I might would also consider doing just a single microphone supplying audio for both channels for mono recording. It seems that the stereo effect is minimal, and would be better with a separate audio recorder. With such a setup, one would probably only care about the GoPro audio for the voice of the rider, maybe even excluding that when the rider isn't speaking. 

I've since learned that GoPro 9's are notorious for having clicking/popping noises, and maybe the Sony mic was not at fault. Just adding an inline resistor for each microphone element might have solved the problems with the Sony Mic. But, of course, the very thin wires used for consumer audio products make such modifications very difficult. Building my own stereo microphone setup made me feel like I had more control over the end result.

Hopefully soon, I'll post a video demonstrating how well the microphones work. 

Here it is plugged into the GoPro with a short cable I made.