Primary Testing: Choosing Shapes

Figures (a), (b) and (c) show our primary working recording hardware. With reference to (a), the Arduino MKR1000 was mounted on a Printed Circuit Board (PCB) (green). Connected to the Arduino was the microphone at the bottom and the SD card on the right (Figure (a)). The microphone was inserted into the neck of the echo chamber. With reference to (c), we used a commercially available Littmann diaphragm held at the bottom of the cone-shaped echo chamber. The Arduino was plugged into a laptop via USB A and the recordings were controlled (start and stop) with the Arduino software (Integrated Development Environment, IDE).

Figure (d) shows how the stethoscope is placed at the aortic region of a horizontal and flat chest (the honorary test subject is Jia Hao, a member of our team). At this region, we wanted to listen for sounds of the aortic heart valve. The hardware was kept as stable as possible while the subject breathed regularly. The diaphragm was pressed firmly but with normal pressure against the chest.

Using the Arduino IDE on a laptop, we managed to generate audio amplitude graphs of amplitude (as unsigned 12-bit values which ranged from 0 to 4095) against time. After starting the heart sound recording, we waited about 5 seconds to obtain a relatively stable spectrum. We then recorded the developing spectrum on the laptop screen with an iPhone camera for 7 to 8 seconds to obtain the video (on loop) shown above. This video shows the spectrum obtained with the reverse horn echo chamber. As can be seen, there are distinct pairs of peaks that we believe to have corresponded to the ‘lub’ and ‘dub’ sounds of the heart. In the biological context, the ‘lub’ sound is created when the mitral and tricuspid valves close, and the ‘dub’ sound when the aortic and pulmonary valves close soon after. Here we placed the stethoscope at the aortic region, but sounds from the chest cavity were picked up even from regions further away. Coming back, the aforementioned peaks took amplitude values that fluctuated between 300 and 600 units. This fluctuation was in fact one of the most stable values we could obtain, the other being that generated by the cylinder echo chamber as shown below.

A similar peak range of 300-600 units was observed in this case. Compared to the reverse horn and cylinder echo chambers, the normal cone and shallow cone shapes gave generally lower amplitude sounds (100-300) of the heart recorded under the same conditions, as illustrated in the 2 videos below in that order.

Though at lower amplitudes, the peaks obtained from the cone shapes were still relatively recognisable. The dome-shaped echo chamber, on the other hand, gave rather chaotic signals.

This could be explained by the sound concentration property of curved surfaces, similar to that widely used in theatre hall where sound is focused to the center of the volume, and the performance is best enjoyed at center seats (Vercammen, 2013). Since our design specifies sound capturing at the neck of the echo chamber, these recordings were not unexpected. However, this could also be due to mechanical problems with the microphone when we inserted the microphone into the narrow neck.

Secondary Testing: Choosing Ratios

With these results, we proceeded to a second testing stage with only the cylinder and reverse horn echo chambers. In light of a compact design for a wearable design, we attempted to reduce the size of the chamber. We managed to find Littmann diaphragms at 2 diameters, 4.5 cm and 3.5 cm, so we went for smaller echo chambers that are in line with this measurement, instead of the base diameter of 2 cm as initially intended. Following the base diameter of 3.5 cm, we 3D printed 2 echo chambers (1 for each shape) with the same aspect ratios as our first prototypes, with the exception of the neck which remained the same (where the microphone was inserted). We hereby called these short cylinder and short horn. 2 additional chambers were made where the base diameter were reduced from 4.5 cm to 3.5 cm but the height was kept constant at the original 2.85 cm, and called these tall cylinder and tall horn.