This post is about microphone back-end design: it shows how a microphone’s interoperability requirements—which specify which types of devices it must work with—drive key aspects of its design.
- The microphone back-end
- How interoperability requirements drive microphone back-end design
- A modular approach to DIY microphone back-end design
- The interface between the microphone front-end and microphone back-end
- The interface between the microphone back-end and the microphone input of the downstream device
- Advantages of the modular approach
The microphone back-end
In Part 1 of this series (Basic DIY microphone design—Part 1: the microphone design process), I discussed the overall microphone design process and how a microphone can be decomposed into a front-end (that largely determines performance) and a back-end (that largely determines the types of devices the microphone can be used with). This post is about microphone back-end design, which addresses the items shown in blue in the following block diagram:
How interoperability requirements drive microphone back-end design
Even if our microphone meets all our performance requirements, it will be useless if it’s incompatible with the devices we plan to use it with.
If we’re designing a microphone to work with only one specific device, then interoperability isn’t a big issue: we can just design the microphone’s power supply, electrical interface, and physical interface to match the device’s input characteristics. The following chart shows how the interoperability requirements drive the microphone back-end design:
But more typically, you’ll want your DIY microphone to be usable with a variety of devices, such as DSLRs, field recorders, audio interfaces, laptops, smartphones, etc. Unfortunately, no single microphone design will directly work with all devices that have a microphone input. So the approach that makes the most sense is to design the microphone for compatibility with the type of devices you expect to use the most, and rely on properly-designed adapters for when you have to use the microphone with other devices.
There are five characteristics of a device’s microphone input that have implications for microphone design:
- Balanced-versus-unbalanced input configuration
- Device input sensitivity (or, equivalently, full-scale signal input voltage)
- Input impedance
- Connector type
- Provided microphone power, if any
Fortunately, a microphone with a balanced output will work with both unbalanced and balanced inputs, and a microphone with a low output impedance will work with both low and high input impedances. So, assuming we design our microphone to have a low-impedance balanced output, we only have to worry about three design variables from the standpoint of interoperability:
- Full-scale signal microphone output voltage (or, equivalently, the downstream device’s full-scale input voltage)
- Connector type
- Provided microphone power (if any)
Full-scale signal input voltage
Audio inputs can be grouped into three categories based on the full-scale signal voltage they expect:
- Low-level microphone inputs expect a full-scale input voltage of the Order Of Magnitude (OOM) of 1 mV. This is the typical OOM output level of dynamic microphones.
- Standard-level microphone inputs expect a full-scale input voltage of OOM 10 mV. This is the typical OOM output level of the powered microphones which represent the bulk of the microphones in use today. These include microphones based on small, low-voltage Electret Condenser Microphone (ECM) elements, larger diameter, Externally-Polarized Condenser (EPC) elements, and Micro-Electro-Mechanical Systems (MEMS) microphone elements.
- Line-level inputs expect a full-scale input voltage of OOM 1 V. Actually, consumer-grade line-level inputs, called Aux inputs, expect a bit less (around 0.3 V), but still within the same OOM.
The vast majority of devices we’re going to want to drive with our DIY microphone will have a standard-level mic input or a line-level input, so there’s isn’t much need to design a DIY mic to intentionally produce a low-level output signal. That leaves standard-level and line-level output voltages.
Not surprisingly, most microphones are designed to produce an output voltage compatible with standard-level microphone inputs. But a microphone capable of producing a line-level output offers a significant advantage: since the voltage is 20 dB greater than a standard-level output (1V versus 10 mV), any electrical noise picked up on the wiring between the microphone and the downstream device will be suppressed by 20 dB due to the reduced voltage sensitivity of the line-level input. However, even with a 10 mV output, the noise at the downstream device is usually dominated by ambient acoustic noise rather than electrical noise picked-up on the cable. So, the added 20 dB of suppression from a 1V output level typically has a negligible impact on performance.
Also, there are three disadvantages associated with designing for a line-level output:
- not every audio input device has a line-level input;
- we’d need a more capable preamplifier to generate the line-level output; and
- if the mic is battery-powered, the battery life is going to be shorter (or we’re going to need a larger battery) than if the output were just 10 mV.
Based on the above considerations, designing for a 10 mV nominal output voltage generally makes the most sense. One exception would be an ultra-low-noise microphone intended for use in a quiet studio, when the downstream device is known to have a line-level input; in that case, designing for a 1V line-level output can maximize the overall signal-to-noise ratio (with careful design of the buffer/preamp built-in to the microphone).
If we exclude low-level and line-level inputs, most of the devices we’d want to drive with a DIY microphone will have one of three types of mic input jack/pinout:
- 3.5 mm (1/8-inch) Tip-Ring-Sleeve (TRS) phone jack with Sennheiser/AC’97 pinout. This extremely common format has the microphone signal (optionally with a DC bias for Plug-in Power, which we’ll discuss below) on the Tip, and the signal ground on the Sleeve. The Ring carries only the DC bias, if it’s present.
- 3.5 mm (1/8-inch) Tip-Ring-Ring-Sleeve (TRRS) phone jack per the Cellular Telephone Industries Association (CTIA) standard. This has the microphone (with optional DC bias) on the Sleeve, and the ground is on the Ring closest to the Sleeve.
- XLR3 (3-pin) jack. This has ground on pin 1, and balanced signals on pins 1 and 2.
Balanced versus unbalanced signals and connector configuration
High-performance microphone inputs use a differential input stage that expects a pair of balanced microphone signals. Balanced signals have the same amplitude but opposite polarity, and the differential stage produces an output that’s the difference between the two signals. The benefit is that any common-mode noise picked-up in the input cabling—which is noise that affects both signals equally—is subtracted-out in the differential input stage.
There are two requirements to obtain this benefit:
- Both signal paths from the microphone to the differential input stage must have the same impedance. This has nothing to do with the type of connector, and is affected only by the design of the input stage and (to a lesser extent) the cable.
- Both signal paths must be equally well (or equally poorly) shielded from electrical noise. This does depend on the connector configuration to some extent (as well as the cable). However, with the right cable configuration, the type of connector has only a small effect on noise pick-up and rejection.
All audio devices that have XLR3 inputs also have differential input stages that expect balanced signals, and all XLR3 connectors and cables are configured to ensure the best possible balance between the two signals to maximize common-mode noise suppression.
Conversely, many audio devices that have phone-jack microphone inputs do not have differential input stages or balanced input impedances. However, some do. And while a phone-jack interface isn’t ideal for carrying balanced signals, there’s nothing to preclude it from carrying balanced signals—and with the right type of input stage and cable, it will provide almost as much common-mode rejection as an XLR set-up. So any DIY microphone that uses a phone-jack interface can (and probably should) be designed for balanced signals.
Which connector to design for?
Of the microphone input-connector types shown above, the most common across all devices is the 3.5 mm TRS type—and it’s also the smallest and least-expensive type of wide-used microphone connector. However, the most common microphone input-connector type across high-performance audio devices is the XLR3.
Fortunately, it’s possible to use adapters to connect a 3.5 mm TRS plug with an XLR jack, and vice-versa—but it can be tricky. That’s because the connectors sometimes supply power to the microphone (as well as accepting the microphone’s output signal), and the two types use differing power standards.
As has been the case for many decades, most of the microphones in use today require external power: microphones using ECMs and MEMS elements require a few volts, while those using EPC capsules require a few tens of volts (dynamic mics per se don’t need any power, but they need a preamp for use with standard-level inputs…and if the preamp is built-into the microphone to minimize noise pick-up, then the mic will need power for the preamp).
If we include include a battery, USB, or mains-powered supply in our DIY microphone, then we can eliminate microphone power as an interoperability issue. On the other hand, having to worry about batteries or plugging-in a USB or line cord can be inconvenient.
Fortunately, most microphone inputs today provide some form of microphone power: either Plug-in Power (PiP) or Phantom Power (PP). The power is supplied by the same cable and connector that accepts the microphone output signal, which is very convenient.
From a design standpoint, another convenient fact is that any device that supplies PiP almost always has a 3.5 mm connector, and any device that supplies PP almost always has an XLR connector. So, if we design for one type of external power, we only have to design for one type of connector.
However, relying on either PiP or PP does present potential performance issues (as I’ll discuss in the next section).
So, selecting the power source for a DIY microphone comes down to a trade between performance (which favors a dedicated microphone power supply) and convenience (which favors either PiP or PP).
Plug-In Power (PiP)
Almost all consumer-grade microphones use ECM capsules, so almost all consumer-grade audio inputs provide a few volts of what is known as Plug-in Power (PiP).
PiP is very convenient. However, that convenience comes at the expense of performance, for two reasons:
- Every mic capsule that can use PiP power expects a certain PiP supply voltage and load resistance, whereas the actual PiP voltage and load resistance vary from device to device. Therefore, it’s unlikely that any given mic capsule used with a given PiP supply will see the optimum voltage and load resistance. And if a mic capsule doesn’t see the optimum supply voltage and load resistance, it won’t provide its full specified performance (sensitivity, dynamic range, SNR, maximum SPL capacity, and THD could all be affected).
- PiP voltage can sometimes contain significant noise from the digital circuitry in the supplying device, which can raise the noise floor if the device doesn’t also have a differential input stage to suppress common-mode noise.
I’ll discuss these issues in more detail in my post on why you should avoid Plug-in-Power for a high-performance DIY microphone; as the title implies, it’s best not to rely on PiP power to power a DIY microphone if you care about performance. Fortunately, just because a device supplies PiP doesn’t mean a mic has to actually use it.
Phantom Power (PP)
Externally-polarized condenser elements are used only in more expensive microphones, so the tens of volts they require is provided only in prosumer or professional-grade audio equipment, via what is known as Phantom Power (PP).
Unlike PiP, the supply voltage and load resistance of PP is pretty consistent from device to device: although there are three voltage standards (12, 24, and 48V), each with different load resistances, almost all devices that supply PP are compliant with the 48V standard (which specifies 6.8K load resistors). So a microphone that expects PP can be safely designed for 48V/6.8K with the expectation of wide interoperability.
Also, while PP can contain noise generated by the device that supplies, PP is usually much quieter than PiP, and all devices that supply PP have differential input stages that suppress common-mode noise.
However, PP can also present a couple of challenges for the DIY mic builder:
- Most DIY mics require much less than 48V, so the mic will need to include some form of voltage regulator.
- Most modern devices that supply PP are capable of only sourcing up to 10 mA of current…and some aren’t even capable of sourcing that much.
The first issue is easily overcome, but the second could be a significant problem for DIY mic designs that need more than 10 mA. Fortunately, it’s not hard to stay within 10 mA unless you need a very specialized design.
A modular approach to DIY microphone back-end design
I build a lot of microphones of different types…but I rarely use more than two microphones at the same time (and I suspect that many other DIY microphone builders fall into the same category).
Because I need to drive only a couple of microphone inputs at the same time, it doesn’t always make sense to build the back-end functionality into every microphone. Instead, it can be much more efficient to put the back-end functionality in a separate back-end module, and build only enough such back-end modules for the number of microphone inputs I expect to drive simultaneously. This greatly reduces cost and fabrication labor.
This approach requires standardizing on the interface (the type of connector and its pinout) between the front-end and back-end modules, and between the back-end module and the microphone input of the downstream device.
The interface between the microphone front-end and microphone back-end
All standard microphone interfaces use two or three pins. However, a modular microphone configuration is better-off with a non-standard interface.
A four-pin interface provides maximum flexibility
All standard microphone connectors provide only an unbalanced (single-ended) output signal, and/or share one of the conductors for both signal and power. For maximum flexibility, a high-performance modular microphone is better off with an interface that has separate power, ground, and differential signal conductors:
- Keeping the power separate from the signal (unlike Plug-in Power or Phantom Power) eliminates the need for relatively large DC-blocking capacitors in the front-end module and facilitates powering MEMS microphone elements and active circuitry.
- Differential outputs allow a downstream audio device that has differential inputs to greatly suppress common-mode noise picked-up on the microphone cable.
Such a four-pin interface can be implemented with the same kind of 3.5mm TRRS connectors used in smartphones and/or 4-pin XLR connectors used in intercom headsets. Of course, the pin assignments will be inconsistent with standard applications for such connectors.
A non-standard three-pin interface is appropriate if all your microphones use the same ECM element
Many DIY microphone builders use nothing but ECM elements in their microphones…and often standardize on a favorite make and model of ECM element. In that case, the design of the front-end modules can be simplified with a non-standard three-pin interface:
- Two pins handle the two terminals (drain and source) of the ECM’s FET buffer.
- The third terminal is the signal ground on the back-end side. This allows a shielded cable to be used between the front-and and back-end modules to provide some noise immunity. The ground is connected to the shield on the back-end side, but it must be isolated from either of the ECM element’s terminals on the front-end side. It can either be left unconnected on the front-end side, or connected to the ECM element’s body (if the body is isolated from either terminal) or a metal housing.
The interface between the microphone back-end and the microphone input of the downstream device
I’m interested in capturing sound with high fidelity, so most of the microphone inputs I drive with my DIY microphones are of the XLR type that supply Phantom Power. So, it’s a no-brainer to standardize on that interface for the output of my back-end modules.
When used with a 4-pin interface between the front-end and back-end, the result is the modular configuration shown in the following figure:
I do have some miniature field recorders that have 3.5mm TRS inputs that supply Plug-in (not Phantom) power (and therefore don’t have differential signal inputs), so I’ve also built a few back-end modules that work with those inputs, too:
Advantages of the modular approach
The main advantage of the modular approach shown above is that you don’t have to build a whole new microphone for each different type of microphone front-end you need, or for each different type of audio input you need to drive.
For example, suppose you need a lavalier microphone with an isotropic response, a studio microphone with a cardioid response, and an ultra-directional parabolic microphone—all for use with the same high-performance field recorder. Without the modular approach, you’d need three copies of all of the functional blocks of Figures 7 and 8. But with the modular approach, you’d need to build three microphone front-ends, but only a single copy of the interface/power module.
Similarly, suppose you’ve built a parabolic microphone that needs to work with both a high-performance audio interface with XLR inputs that supply Phantom Power, as well as with a miniature field recorder with 3.5mm TRS inputs that supply Plug-in Power. Without the modular approach, you’d need two complete parabolic microphones, but with the modular approach you’d only need two interface/power modules.
In summary, the advantages of this modular approach are that it enables a DIY microphone to use a variety of different microphone front-ends (with different types of microphone elements, and with and without onboard signal-processing) and to be powered from a variety of sources…and by using different output adapters, it can be used with virtually any device that has a microphone input.
I’ll go into the detailed design of the microphone, power, and adapter modules for this modular approach in upcoming posts.
As we’ve seen, a microphone’s back-end architecture flows directly from its interoperability requirements, and defining it is a key part of the overall microphone design process (as described in Part 1 of this series). No single back-end design is optimum for every application, but a modular approach like the one I use can maximize convenience, performance, and interoperability.