For a recent project we wanted to capture signals from a couple of microphones using the Digilent Analog Discovery. If you are not familiar with it, the Analog Discovery is a very cool (and cost effective) data acquisition system. It features a network analyzer function which is perfect for making basic audio measurements.
One problem: Microphones typically have a very low output signal. The Analog Discovery has limited gain options and our mic signals were too low to get good data.
Solution: A level matching preamplifier. In order to get better signals from microphone outputs we generally need to do “level matching” or amplify the signal in order to match the downstream circuitry. In this article I will describe how to modify the AudioEQ to provide a microphone preamp function. Some key benefits to this solution are:
- Fixed gain solution for repeatability
- High quality/low noise signal conditioning
- Compact design
- Very simple modifications to an off-the shelf design
The AudioEQ comes as a unity-gain audio buffer. The input signal is filtered to reject signals below 20Hz and above 20kHz. Signals from 20Hz to 20kHz are passed through at the same level as received. This is a perfect fit for a microphone preamp. Since it is a completed and tested device, we can focus on changing the gain of one of the intermediate stages. A very minor change and we are done. It also has an option for an industrial style case, so if you need something like this for lab use, the AudioEQ is a perfect choice!
The Digilent Analog Discovery data acquisition conversion takes a 2Vpp differential input as full scale. Our goal is to get the microphone signals to that level.
The microphones we are measuring are INVN421 analog mems microphones. In the test environment we are seeing about 40mVpp output from these microphones. An overall gain of 40x (32dB) would get us to a 1.6V pp signal which is near to full scale while leaving a bit of headroom in order to avoid clipping. Since we will drive the Analog Discovery differentially, we only need half this gain in our preamp, so a gain of 20x (26dB) is what we will design in.
We will modify the final eq stage to get our gain: Since this is a non-inverting amplifier, the stage gain is 1+ R307/R308. For a gain of 20, R307/R308 = 19. I want to keep my impedance high so I will change R308 to 10kohm. Then R307 = 19* R308 = 190kohm. The nearest EIA 1% resistor value is 191kohm which should work just fine. We also need to add a jumper in at C308 in order for this gain to take effect. (C308 could also be a large valued capacitor, eg C308>>42nF. A capacitor here reduces DC offset, but is not essential for this application).
For two channels we make the same changes at U5B. A summary of changes then is:
- R307, R407 -> 191k ohm
- R308, R408 -> 10.0k ohm
- C308, C408 -> 0 ohm jumper
Additional changes we made for this project were:
- Installed J3 & J4 for single ended inputs.
- Installed 2 jumpers on each of J9 & J10 for Differential output.
- Wired the outputs to a header to match the Analog Discovery inputs.
- Added a 9V battery clip and taped the battery to the Analog Discovery for portability.
Other things you could do:
Change the op-amps to a precision type op-amp such as the OPA2227U.
Bypass the first EQ stage for a very slight reduction of noise:
- Remove: C24, C25, R301, R401
- R103, R203 -> 0 ohm jumper.
- R305, R405 -> 10k ohm
- Run a jumper wire from C24 to R301
- Run a jumper wire from C25 to R401
Our finished result worked great and we were able to get get repeatable response measurements using the Analog Discovery and an AudioEQ based preamp. Here is our finished assembly ready for measurements. Not pretty, but it was compact and effective. This is a great application for the AudioEQ board, and can be applied to many other level matching scenarios. For a more professional application we recommend using the industrial enclosure and a wall-pack power supply.