Published 20 February 2013

Building a microphone preamplifier
Using a THAT1512 IC

The original mic amp which I had been using with the Behringer ECM8000 microphone was based on an INA118 instrumentation amp IC and did not have the best noise figure. When I was given a THAT1512 to try, the v.2 mic amp project was started. Data sheet for the THAT1512 can be found here. This IC is designed from the ground up to improve on existing integrated microphone preamps by offering lower noise at low gains, wider bandwidth, higher slew rate, lower distortion, and lower supply current. The IC includes ESD overload protection on all critical pins.

My design evolved a little (some call it scope creep) to include some high-pass filtering, low pass filtering and a simple bar-graph level indicator. I decided to run the unit from +12V and so the design includes a 48V converter for the phantom power and balanced 15V rails for the audio stages. Most of the sub-sections are based on manufacturer data sheets or are off the shelf boards. This is however, not a step-by-step constructional project although experienced constructors should have no problem replicating the final design.

Block Diagram

mic amplifierOn the top, you can see the gain selector and the two filter selection switches. Front of the box; not visible, has the female XLR input connector and the small 3-way slide switch for selecting phantom power. At rear, is the balanced output socket, dc input socket, unbalanced audio output and the LED bar meter. Three tiny LED's show which gain range is selected.

block diagram

The balanced output driver is useful to drive long cables between the amp and the measurement soundcard or in my case, the Liberty Instruments Praxis module. I made the connector a 5 pin, so that dc could be run up the same cable for convenience. The phantom supply for the Behringer is 48V, however, for greater versatility, there is also a 5V phantom option, so that a simple electret mic could be used.

The sub-sections

PCB1: THAT1512 board

This is built simply on 0.1" pitch veroboard and is largely as per the data sheet. THAT1512 schematic

Differential input resistance of this board is 2.2Kohm. The three gain selection resistors R3, R4, R5 are parallel combinations in each case in order to get the exact gains required. You may consider the gain options of 10dB, 28dB, 40dB are a little low, however, the primary use for the preamp is for loudspeaker testing, and high gain is not required. Given the specified sensitivity of the ECM8000 is 10mV/Pa, then at 94dB SPL the 40dB gain range will produce 1 volt at the output (0dBV).

PCB2: High pass filter board

High pass filtering is very useful to reduce wind noise or unwanted low/infrasonic frequencies. The three-way selector allows choices of no filter, 15Hz cutoff or 60Hz cutoff; both with 36dB/octave attenuation. For this unit, I used the design and pcb from Elliot Sound Products called 'Subsonic filter; project P99'. I have no permissions to reproduce this, but here is a link to the project.

The board is designed as a stereo subsonic filter, but for my purposes, I made each channel a different cutoff, paralleled the inputs and selected one or other output with a switch. The critical capacitor values C1, C2, C3, C4, C5 and C6 were 180nF on the 15Hz side and 47nF on the 60Hz side. NE5532 op-amps were used.

PCB3: Low pass filter board

It was decided to add low pass filtering partly as insurance against there being little effective anti-alias filtering in recording devices or sound cards and as protection against recording unwanted supersonic sounds. While I understand most sound cards use oversampling techniques which makes alias frequencies very high, there is no guarantee that it is done properly. In environmental recording, there is often little value in recording sounds above 8-10kHz and so sampling rates of 22.05kHz can be used to reduce recorded file sizes. In this case, nothing above 11kHz should reach the A/D converter. For this unit, I used another subsonic filter; project P99 pcb from Elliot Sound Products and modified it to behave as a low pass filter, although I made these sections only 24dB/octave. While this rate of attenuation is no substitute for a full anti-aliasing filter, it will supplement such filtering in any A/D converter. What were R2,R3,R5 and R6 are now capacitors, while C2,C3,C5,C6 are now resistors. A 51kohm resistor is added across the input terminals.

Again, each channel was made a different cutoff, the inputs paralleled and either output is selected with a switch. The 10kHz LPF uses 11kohm resistors with C=1nF, while the 20kHz LPF uses 12kohm resistors with C=470pF. NE5532 op-amps were used.

PCB4: Balanced line driver board

This is a design originally published by Elektor magazine although it is based entirely on the data sheet for the Analog Devices SSM2142 IC. I had some of these chips and an Elektor pcb making this section rather simple.

SSM2142 driver

Normally one sets potentiometer P1 to maximum; providing x2 gain, but here, it is set so that the balanced output voltage is identical to the applied input voltage.

The audio detector

This makes use of the Analog Devices AD8307 high frequency log detector. While there are many ways to make an audio detector, some AD8307's were to hand. It may seem odd to use a 500MHz RF demodulator for audio, but this IC works well at audio frequencies. I had used one of these as a ultrasonic range finder receiver and that worked very well.detector

For operation at audio, the input capacitors are 100uF. A simple 5V 3-terminal regulator is on board. The dynamic range used in this application is 45dB and the detector is rms responding, although only up to crest factors of about 3. A square wave and a sine wave of identical rms value, produce identical dc output, however, a rectangular wave of 1:4 M/S ratio and pink noise which are set to the same rms value as the sine wave, read about 1dB low. This isn't of great concern here as the LED array steps are 5dB apart.

The level indicator

This is based around a 10-LED bar array and LM3914 driver. For this unit, I used the design and pcb from Elliot Sound Products called 'LED Audio VU meter; project P60' which itself closely follows the National semi data sheet. Since I have a log detector, I needed the LM3914 (linear) driver as opposed to the LM3915 log driver as used in the ESP project. The schematic as used here is shown below with the actual resistor values.

LED driver
The LED steps are 5dB apart and the board is calibrated so that the top LED represents +5dBV at the output. Total displayed range is then 45dB. The THAT1512 just starts to distort at +9dBV (2.82Vrms) out, so there is a good margin. Assuming the ECM8000 mic does deliver 10mV at 94dB SPL, then the top LED represents 100dB, 112dB, 130dB SPL using the gain ranges of high, med, low respectively. Behringer do not specify a maximum SPL for the ECM8000, however it is unlikely to cope with 130dB SPL. So, using the low gain range I would need care to limit the peaks to LED No.8. The THAT1512 is fairly heavily loaded in this design, which does compromise the maximum output, however the undistorted output is high enough for this application. If this situation needed attention, it would be necessary to use another pole of the filter selector switches to simultaneously switch the inputs to the filter boards.The potentiometer is a high (1Mohm) value due to the high output resistance of the AD8307 detector. However, this is workable since the LM3914 chip has a very high input resistance. I have left pin 9 unconnected to operate the display in 'dot' mode, since 'bar' mode could draw significant current at 10mA per LED.

Power supplies

My original v.1 preamp used a 20Vac input and created the 48V by voltage doubler techniques. This method works well but you do need an ac input. In this version, it was considered that portability could be useful, which dictates a 12Vdc input. So, included in the design is a balanced 15V dc/dc converter and a 48V dc/dc converter. A low dropout 12V regulator was used ahead of the two dc converters partly as protection and to ensure the 15V rails do not vary with input voltage.

The low dropout 12V regulator is a simple implementation of the datasheet for the LM2940, so I have not reproduced the schematic here. Similarly, the 15V rails are produced by the Murata NMH1215SC. This IC can deliver plus and minus 15V at 67mA. The total draw on the 15V rails here is 45mA.

The 48V converter uses the Linear Technology LT1172 switching regulator. Schematic for this is below.

LT1172 schematic

Using these in boost mode means the there is no effective current limiting possible. The ECM8000 mic draws around 4mA only and is fed via 47 ohms on board, 100 ohms at the phantom selector switch and then 3.4kohms to the XLR socket. Therefore, no external short circuit can cause a problem. Inductor L1 is a powdered toroidal core of 180uH inductance about 12mm diameter. Any on board short will be limited by the LM2940 LDO regulator and if that fails, the PTC fuse will open.

The concept of using a dc input with inbuilt dc-dc converters has one advantage and that is there are no 50/60Hz hum products or harmonics of these on the audio output. The switching frequencies of the two converters are both above 50kHz. Total dc current drawn is 210mA at 13.4V input. There is an initial surge up to 300mA as the two dc-dc converters start up.


The frequency response on the 28dB gain setting with the 60Hz HPF and the 10kHz LPF switched in are below. Without the filters, the unit is flat from 5Hz to at least 48kHz on all three gain settings. preamp response

The flat noise floor is between 7dB and 10dB lower than from my INA118P based version and there are no hum or any other unwanted products up to 48kHz.

Any comments or questions please email: Comments may appear below if you request it.

Axino-tech Consulting & Services , February 2013.



There are no comments yet