One of our experiment rooms, implant
where some experiments are sensitive to sound, health is shared with some other groups. To make matters worse, the room is periodically cleaned and maintained by lab technicians. As a result, sometimes it is hard to tell if recorded EEG signals are due to neural activity of the subject caused by some internal mechanisms or if it is due to external sound stimuli. To help alleviate this problem, I designed a simple narrow-band microphone amplifier to plug directly into our acquisition system.
To be clear, the sounds that we are most worried about are short, loud noises like doors slamming, hands clapping and objects being dropped on the floor. Things that we do not care so much about are computer constant noise sources like fans or quiet conversation. The amplifier should be compact and accept a standard pc (electret) microphone. Finally, the device should run off of the +/- 5V power system that is in place and be able to plug directly into our 2000Hz sampling system.
To do this, I first set up a stage to provide bias current to the microphone with AC coupling, then feed the signal through a high gain amplifier. Here, R7 can be increased from 0 Ohm to provide even greater gain. The input stage is similar in design to that of an early SoundBlaster board. The next stage provides DC to RMS conversion by way of an AD736. Because of previously mentioned design requirements and the desire to use non-electrolytic capacitors of a reasonable size (and based on what was lying around in the lab at the time) the high pass, or averaging frequency was set to 300Hz. Because the acquisition system is running at 2000Hz, a low pass filter was added with a corner frequency of 1000Hz.
( schematic )