USB-C DIY Microphone

Hey you, yes you. Are you looking for a really high quality microphone but keep getting foiled by Bose's pricing scheme?

Well, look no further, as I have the solution to your problems. Well, rather Theon Gray has the solution; I'm just ripping it off and changing it to my needs. The official website for this project is here. Other people have actually recreated it and posted their results online, and it's only left me more in awe.

Something to note is that I'm going to follow his approach of making the interface with USB-C, but keep in mind that this is a definite downgrade from a pure analog signal and proper audio equipment.

I also generally recommend reading up on USB-C protocols, since it is the future of ports basically. He also made a simple video on the topic; however as an EEE it's ideal to go a bit deeper and actually understand the power and data protocols properly.

Project Overview

Now, I'm always in need of good microphones. I join a lot of online calls, conferences, and I make a lot of videos and media projects that need high quality audio. I remember a few years back, me and my sister were looking to make one podcast project at the time with my dad. We were browsing microphones in a store and were just floored at the absurd prices for a decent microphone. In the end, we decided to use our phones, but man it would have been so nice to have a high quality, clip-on microphone instead.

So, when I stumbled upon a dude who already did 50% of the legwork for me with research and components, I knew that I needed to give it a go and make one myself, and if it turned out well, I was planning to make another one as a gift for my sister.

Mechanical Design Plan

So, my main idea is centered around modularity. I don't have that much money to make several mics each geared for a different purpose. The best I can do is one microphone module that is detachable from the desk, and used as a clip-on, or maybe even a suspended microphone.

So, something like this, which can have a built-in clip and the hooks can serve as attachment points for other mounts.

The main two challenges in my opinion with regards to off-the-desk usage is damping vibrations so they don't leak into the microphone capsule output, and ensuring that my electronics are small enough to make it practical to use off-the-desk. So, I'm going to have to design proper mounts for off-the-desk usage; mainly a clip/stick on, and a suspended mount of some sort.

Some improvements I can pick up straight away is the lack of a pop shield which can reduce any plosives picked up (a foam cover between speaker and the microphone, or some fluff like the gopro's mic which fits over the capsule cage itself). The other improvement is through software, where a proper EQ profile can be setup to maximise audio quality.

Electronic Design Plan

So, from the electronics-side, I need to decide if I want everything integrated into one PCB; with the microphone attached on the end:

The other choice is similar to the video where I would want to separate the microphone module and electronics aside. Either way I will require a custom PCB to massively reduce space using SMD components and also allows USB-C to be integrated straight into it without a breakout.

I'm also playing around with the idea of adding wireless features, maybe a wireless control knob instead of wireless audio transmission since I have no belief in my abilities to losslessly transfer audio data wirelessly.

After some thinking and research, my overall system should vaguely look like this:

And yes, this is what ACSE students do, I'm definitely not biased; pick EEE.

Ok, beautiful. This system diagram isn't set in stone, and I'll probably have to change some things as i go along. Let's start with actual electronics design and shelve the mechanical plan for future me to deal with.

Preamp Electronic Design

So, Theon established a baseline circuit in the video which saves me the pain of having to make one myself, find components, prototype them and pray they work:

The entire circuit makes sense, except for the weird ass potential divider he adds on the ±15V rail. For the life of me, I can't figure out the purpose of including these two resistors. My only theories are regarding stablising high frequency ripples? But this is honestly non-standard for a DC-DC converter output, so it must be a special purpose for this circuit that I don't know of.

There are a few problems that prevent me from just copying it over though. Firstly, most of these components are much chonkier than they need to be since he designed it to be stripboarded. If I want to make my end result portable-ish, I'm gonna need to downsize all of this unnecessary space wasted. So, I'm going to have to port all/most of the components to SMD packages.

Now, this porting to SMD really isn't necessary at all and literally the only components that need changing are the bulky decoupling caps for the boost output and of course the NMA boost converter itself. But, I have to justify going through the pain of this degree somehow; so I'm doing it anyways. The main components to replace are:

2N4416 - Front-End Amplifier
THAT1512 - Preamp Chip
NMA05155C - Isolated Boost Converter
Remainder of passive through hole components (resistors and capacitors)

2N4416 Replacement

Good news is the 2N4416 doesn't need to be SMD-ified since it's stuck straight onto the capsule, however bad news, it is unfortunately an obselete component, and I'm not in the mood to pay like 12$ for a tiny ass FET. So, I had to start looking for an appropriate substitute. I went diving on digikey and mouser and came up with a list of potential substitutes; I was mainly focused on finding a similar turn on voltage for the gate, and base drain current at that voltage so I can most closely mimic the amplification characteristics of the 2N4416. I also looked for a low noise figure value as a nice bonus:

I ended up choosing the J111, although I reckon the others are close enough to also work. Choosing the J111 also saved me about 9$ which was a minor plus as well. Fingers crossed it works as expected.

THAT 1512 Replacement

Now, the next step was the THAT1512 preamp IC. That one is a surprisingly decent preamp for this usecase; since we're using a shitty ADC later on for digital USB audio, and not the analog XLR signal. And although it really is a good preamp chip, it needed a ±15V rail to function and if I could find a way to cut that out then it would save me the need to setup a boost converter for both rails.

Now unfortunately, voltage rails are an inherent part of any audio system. If you go back to your op-amp theory notes, you'll know that distortion is something that can often happen due to a small voltage rail. The signal's roof is always capped at the supply rails of the op-amp meaning that if it goes higher or lower than the supply voltage, it will just clip to that rail and you lose out audio range and quality. I did briefly look into low-voltage preamp chips like the TL072, OPA1641, and SSM2167, but none of them offered the same low-noise and distortion characteristics as the one by THAT.

So, I begrudgingly stuck with the THAT1512. It was pretty much in stock only on mouser and farnell; and who tf buys from farnell unironically. Mouser didn't have the SMD 8pin package; only the 14-pin package in stock. I wasn't too fussed since the 3D height is damn near identical for both the SMD and Through-hole packages. So, I ended up going for the 1512P08-U, which is the through-hole 8-pin package also used in the video.

NMA 05155C Replacement

Now, to find an appropriate boost converter for the voltages we want. The max supply ratings on the THAT1512 is ±20V, and the datasheet encourages the use of ±15V, as it leads to an overall 26.6V output swing; which is really good for overall range. So, my decision is straight away made for me. I did briefly consider the implementation of 48V phantom power for a less noisy and distorted mic signal but again, what's the point if I'm just converting my preamp output to a digital signal and no XLR output? I don't have any fancy audio mixers or whatnot that I need compatibility with, so it's essentially useless for my application. I'll shelve the idea for a future upgrade [1][2][3].

With those considerations, I ended up going for the TI-DCPA10515DP; which is an SMD isolated boost converter that takes in 5V and outputs ±15V at 33mA.

Schematic Creation

Something that was interesting to me is that the video schematic barely followed the application circuit in the datasheet at all, which confused me to be honest. But I can only assume he chose a really simple schematic to make it easy for the viewer to recreate. He essentially just connected the preamp inputs & output to some DC-blocking capacitors and called it a day.

After looking through the datasheet properly, I ended up deciding to go for a circuit based around their 'basic' recommendation:

RG is for gain control, and I planned to replace it with a sliding or rotary pot. I also planned to add a switch shorting the audio signal to ground in order to function as a mute switch.

I debated adding DC-blocking caps like Matt did in his schematic, and decided to include their footprints as well although the manufacturer datasheet didn't specify the need for any at all.

The only change I did was get rid of C1, C2, and C3 since I didn't understand what their point was and assumed that their removal wouldn't significantly impact noise quality. My running theory is that they're some sort of common high frequency filtering for both inputs (although why even include C3 at this point??). Whatever, moving on...

That brings me to the current schematic (still a work in progress) which ended up looking like this:

PCB Layout for Preamp Circuit

LTSpice Simulations (especially for C1, C2, C3)
Use a LogPOT instead of a normal linear pot.
Justifications for layout, and low noise considerations.

Audio to USB Interface Design

DIYing this section almost made me blow my head off. I'm sure if I invested enough money and time I could eventually make one work cleanly at a low-level. The thing that makes this so difficult and expensive to properly make is the several complicated stages involving both hardware and software expertise to pull this off.

So, firstly the preamp output is an analog signal and needs to be converted to a digital signal. To do this without losing out on a lot of quality, you need a really good ADC (Analog to Digital Converter), and those chips can get really expensive quickly without economies of scale backing you up. After that, you need to know what the hell to do in terms of the actual USB interface; this means the process of first establishing a connection with the host device (computer) and then handling the transfer of the now digital signal to the computer.

Nope, not done yet. To get a good digital audio signal, it'll need to go through even more digital signal processing (e.g. audio equalisation).

Ok, so DIYing it properly is off the table due to a lack of skillz. Does that mean I just have to buy an off-the-shelf interface?

Well, yes, the true plug-and-play solution is to simply buy a commercial USB Soundcard or audio capture card and transplant the PCB inside and connect your inputs and outputs through it; which is what Matt did.

So, that is a valid option to consider especially since they're not terribly expensive either and can be quite small. It does ruin my whole plan of condensing all the circuitry onto one tiny PCB, but at least I'd be guaranteed a damned good audio signal with a commercial audio capture card.

So, after a ton of research on audio capture chips that would be suitable to transplant for this project, I came across some VERY helpful forum threads for this project (here, and here), and Matt admits in a reply that the perfect audio interface for this microphone comes from Behringer. So, I went ahead and ordered literally the cheapest one I could find which was a "Guitar Line 2 USB". Matt said it uses the same chip, so should have the same effect in the end. So, I bought it.

And then literally the next day, I stumble upon this video, and after a tiny bit more diving, I also found this video.
For fuck sake, the Teensy microcontrollers have built-in ADCs and DACs, and can interface with computers through USB; and of course being Teensy boards, they're easy as fuck to program compared to some low-level chip.

Page updated

Google Sites

Report abuse