eConvergence's Belfry Bat
Frequency Division Detector

Vincent Goffin

Last updated Sat Aug 12 14:19:36 EDT 2000

The Detector

The Belfry bat detector shows up early when a web search is made for "bat detectors" and that's what initially drew me to it. It's a very simple frequency division detector, available in kit form for $49.95. The Belfry is actually very similar to the Simple bat detector. This last one has the advantages of more detailed explanations and a kit available for $35. I estimate that these kits contain $20 to $25 worth of components.

The unit is not hard to assemble, it has so few components! The lack of a circuit diagram makes understanding how it works a bit of a challenge. (A schematic diagram is provided below). The belfry PC board is also incomplete. One of the capacitors has to be secured in the lid of the unit.

Schematic diagram and parts list.

My current Belfry detector is assembled on a small RadioShack protoboard, housed in the original case. Something that is more suitable for tinkering.

My expectations for this detector were rather low at first, so I was pleasantly surprised when it detected bat sounds on my very first expedition. The fact that there are no volume control or other adjustments makes it really easy to handle. I have come to regard this type of detector as a hard-to-beat, simple, efficient and elegant way to enter the world of bats!

Recordings

July 12, 2000 early nightfall at Lake Surprise, Union County, NJ.

Our first search for bats, and there was no need to worry: the bat calls really leap out from the detector and are unmistakable. For the most part the bats were invisible among the trees, but we did have one unambiguous sound and sight confirmation as a bat flew from the trees over our heads and over the lake where it was easy to spot against the open sky.
We recorded echolocation calls (pulse trains at approx. 10 hz), pulse ultrasounds (around 35 khz) and buzz-stop calls (buzzing around 80 khz).

0712-1 amplitude and frequency plots.
0712-1 sound (4s, 30 kb mp3) pulse train, followed by a buzz-stop.
0712-1b sound the buzz-stop only.
Discrete pulses are emitted around 9 per sec, and each pulse is a burst of 35 khz sound. The final buzz contains frequencies around 80 khz.

0712-2 amplitude and frequency plots.
0712-2 sound (14s, 112 kb mp3) several consecutive buzz-stops, followed by a pulse train.
0712-2b sound the pulse train only.

0712-3 amplitude and frequency plots 50s of multiple bat calls immediately following 0712-2.
0712-3b sound (mp3) echolocation
0712-3c sound (mp3) echolocation
0712-3d sound (mp3) echolocation

0712-4 amplitude and frequency plots.
0712-4 sound (0.8s, 7 kb mp3) an isolated hiss. Frequencies are as high as 100 khz.

July 23, 2000, Lake Surprise, Union County, NJ.
Some pulses are quite different, they are actually triple pulses.

0723-1 amplitude and frequency plots.
Including clear evidence for pitch lines (and hence vocal cords).
0723-1 sound (3.3s, 26 kb mp3).

0723-2 amplitude and frequency plots.
Each pulse can be seen splitting in 3, as frequency drops.
0723-2 sound (3.2s, 26 kb mp3).

Aug 05, 2000, Lake Surprise, Union County, NJ.
During this walk we detected a steady source of ultrasound pulses and, turning on a flashlight, discovered we were nose to nose with a little brown bat (Myotis lucifugus) hanging from a tree trunk. Unfortunately my tape had already run out! We detected our first pulse-zap, a new type of broader (> 6x) pulse and new types of much faster (20, 100, 170 and 200 hz ) pulse trains.

0805-1 amplitude and frequency plots.
A new type of fast pulse train (100 hz) followed by a regular (10 hz) one.
0805-1 sound (5s, 44 kb mp3).

0805-4 amplitude and frequency plots.
A new type of wide pulse followed a little later by a regular pulse train.
0805-4b sound (1s, 8 kb mp3) a leading single broad pulse (0.06 sec).
0805-4c sound (3.5s, 29 kb mp3) followed by a train of regular pulses (0.01 sec each).

0805-5 amplitude and frequency plots
A clear pulse zap, some echolocation and again one of those new fast (> 100 hz) pulses.
0805-5b sound (1s, 8 kb mp3) the pulse zap.
0805-5d sound (1s, 8 kb mp3) the 170 hz pulse train.

0805-7 amplitude and frequency plots.
10 seconds of sound featuring different pulse frequencies (7 hz, 20 hz and 100 hz).
0805-7b sound (mp3) 7 hz train.
0805-7c sound (mp3) 20 hz train.
0805-7d sound (mp3) 100 hz train.

0805-10 amplitude and frequency plots.
15 seconds of sound with 2 fast echolocation calls (200 hz) and two regular echolocation calls (8 hz)

Notes

Recording

The sounds were recorded on an inexpensive ($29) Panasonic mini cassette voice recorder. The rest of the processing was done on a PC running Linux (Red Hat 6.2). The tape was converted to a wav file (32 k samples per second, mono) and edited using linux-snd and then compressed to mp3 using bladeenc. The images are linux-snd screenshots captured with gimp. I use freeamp to play the samples from the web. While a PC is of course quite expensive, all the software is available at no charge.

Snd is a sound editor for UNIX/Linux.
BladeEnc is a cross-platform (Windows/UNIX) mp3 encoder.
Gimp is an image editor for UNIX/Linux.
FreeAmp is a cross-platform mp3 player.

Comparisons