Batto MkII
The MkI is fine but it has a few issues. The first is that it’s pretty difficult to get a Raspberry Pi to record more than two audio channels. Also it uses quite a bit of power and long recording times are tricky without a big battery. There’s also the issue of interference caused by DC power supply ripple generated from various sources, including writing to the SD card and WiFi. I managed to reduce the impact in the MkI by writing to an external USB drive and turning off WiFi etc but there’s still interference, and a lower power device would be good.
I considered various hardware possibilities for the MkII. The requirements were:
Ability to record at least 4 audio channels (i.e. two SPI inputs) at high data rate
Ability to record to SD card
Low interference
I looked at various Arduino-like solutions and also considered getting an Audiomoth and adapting it. This seemed like a sensible option as I think the ultimate solution might be an extension board that could be plugged in to an Audiomoth v2 when it comes out. In the end though I went with a Teensy 3.6 which uses the same type of micro controller as an Audiomoth but is easier to tinker with. I also decided to use the same microphones as I used in the MkI but just add two more. I really need to get some mics with higher frequency coverage for better bat detection but that involves building circuit boards and doing some reflow soldering, which is all a bit challenging (for me) at this stage and I can still test out the quad-mic concept with the lower frequency versions.
The other nice thing about the Teensy is that the audio library is very nice and the microSoundRecorder library is already pretty close to what I want.
So here it is, the Batto MkII:
The Batto MkII in the field. The main board is in the centre of the picture and the 4-microphone array on the left. Not a thing of beauty, but it works :-)
I spent quite a bit of time trying to get the self-generated interference under control. When the Teensy is taking samples from the microphones it generates a square-wave clock signal and this leaks back into the DC. This DC ripple appears as interference in the recorded audio. I experimented with RC filters but couldn’t really eliminate the problem entirely. In the end I decided to give the mics their own power supply (from a pair of AA batteries) and that made a huge improvement.
So I now have a recorder :-). I’ve now taken quite a lot of data and it works well at sampling rates up to 130 kHz or audio up to 65 kHz (that’s as much as the mics can take, not the Teensy, and I reckon better mics will allow sample rates up to 384 kHz).
I’m now grappling with the data and attempting to get reliable direction measurements. More on that next time…