A Good-enough Microphone Preamplifier (Part 2)

 Introduction

Following on from the first part of this series of articles, we take a look at some of the design choices involved in creating an actual implementation.

Decisions, Decisions

The two candidate designs were very similar in performance, so choosing one was somewhat arbitrary. In the end, I selected the THA1510 circuit as it potentially offered better common mode rejection.

The  next issue that arose was how, physically, to construct the hardware. The active circuitry itself is sufficiently straightforward that it could easily be built on strip board. However, the transformers and audio and power connectors also need a home and it seemed preferable to come up with some sort of PCB. Given that ordering one PCB typically means ordering several, I decided to create a PCB sufficiently generic to suit a number of different potential projects. You can read more about how it was created here, but it provides for two XLR connectors, two transformers, two relays (for switching), a power connection (either USB B or 4-pin DIN connector) and an audio jack (either 3.5mm or 6.35mm). The rest of the PCB is designed as a breadboard, providing flexibility at the expense of additional wiring. The size of the board was determined by the size of the aluminium case I had available to house it (10cm x 8cm).

A further consideration was how to control the gain of the amplifier. The application notes suggest a potentiometer in series with a large capacitor to minimise changes in the DC offset of the output signal. In practice, although the offset does vary slightly with changing gain, the overall level is low (a few mV) and the change even less, so it's unlikely to warrant the cost and bulk of a large electrolytic. If it were a problem, for most purposes it would be easier to AC couple the output as the likely input impedance of the connected device would mean a much smaller capacitor would be required.

The gain is determined by the value of the gain resistor R as:

1 + 10,000

           R

 

which means the gain increases dramatically for small values of R. For repeatability using a given set of microphones, I felt it was better to be able to set the gain deterministically rather than rely on setting a potentiometer to approximately the right place, but multi-pole switches are typically quite large and would not fit in the case I had to hand. Fortunately, I discovered some miniature multi-pole switches from Adafruit that would serve as the basis of a switched resistance:

A further choice was whether to build a separate power supply, or incorporate the necessary supplies (+15V, -15V and +48V) within the same case. The PCB would permit either option (3 rails via a 4-pin DIN connector or a single 5V rail via the USB connector with on-board power conversion). I wanted the flexibility of using a USB power bank, so it made sense to avoid a (further) separate power supply unit if possible. Unfortunately, it would be tight to fit both of the DC-DC converters described in Part 1 into the case. Fortunately, another and smaller converter offering a dual-rail 15V supply came to my attention:

Like many modern power converters these operate at a relatively high frequency. Despite a discrepancy of around half a volt between the magnitudes of the positive and negative supplies, the voltage regulation is quite good and the ripple voltage is low and well outside the audio frequency range:

Final Schematic

The final schematic differs very little from the prototype - the addition of some decoupling capacitors and the addition of the gain switch:

The gain available in the various switch positions is (approximately) as follows:

1	2	3	4	5	6	7	8
0 dB	12 dB	18 dB	24 dB	30 dB	36 dB	42 dB	48 dB

Some of the resistances are created from series pairs - this stems from a desire to use only E12 values where possible: the 39R values used in positions 7 and 8 are the exceptions. 

Also to be noted is that a different "ground" is shown for pin 1 of the XLR connector - exactly what should be grounded where will be discussed in more detail in a further post on power and grounding.

Physical Layout

The construction can be seen in the image below:

The 15-0-15V converter can be seen immediately below the 6.35mm jack socket.Rather than modify the separate DC-DC converter (right, red) with higher-voltage capacitors, I simply turned the output voltage down to around 42V to give a bit more margin: the microphones worked perfectly well on the slightly reduced supply. The switched resistor chain plugs into header pins adjacent to the IC. 

For initial testing, rather than using a matched pair of resistors to feed phantom power to the XLR connector, I used a single 3k3 resistor connected to the centre tap of the input side of the transformer. More discussion of phantom power issues will follow.

Performance

In initial testing, the pre-amplifier has worked well with no obviously-audible defects. The frequency response is rather wider than I would have expected for a transformer design - the apparent ringing around the lower and upper ends of the audio spectrum seem to be a measurement artefact:

Otherwise the behaviour is inline with the observations from Part 1.