After determining the best suction channel geometry through suction strength testing, the best suction channel was fabricated and attached to the mask body.
The full prototype was tested and evaluated using a sealed container with three isolated flows in ideal conditions which were 1. CPAP machine, 2. Vacuum, and 3. Environment. The CPAP machine was connected to the top of the mask with direct access to the headform to simulate assisted breathing. A vacuum generator was connected to the vacuum channel on the mask, which pulled air from the suction channel through a separate part of the mask from the CPAP flow. A flow meter was connected to the top of the air-tight container to measure any change in airflow in or out of the chamber.
Under perfect conditions, the flow meter should read 0 Liters/sec because the CPAP machine and vacuum should create their own isolated flows.
With the mask placed on top of the headform and the vacuum generator set to 200 mmHg, the CPAP pressure was slowly increased from 0 cm H2O. The airflow out of the test chamber was recorded with the flow meter. The CPAP pressure and the airflow out of the test chamber are plotted on the left.
As seen in the figure to the left, the mask does not significantly allow any airflow out of the test chamber until 242 seconds which corresponds to a CPAP pressure of 11.6 cm H2O. At this point, the airflow out of the test chamber rapidly increases indicating a complete loss of seal. Around 100 seconds and up to 242 seconds, there is airflow into the chamber of approximately 0.1 Liters/sec. This could be an indication of an imperfect outer seal resulting in the mask pulling air from the environment into the vacuum. This result would still satisfy a successful seal because the CPAP flow is still isolated which is the main importance.
Being able to hold a seal up to 11.6 cm H2O is a success because this pressure is in the range at which CPAP machines and ventilators are used. The mask is expected to be able to hold a seal with a higher CPAP pressure with the use of the straps around the head.
Additional testing was performed just to see how a patient may interact with the mask.
A human subject put on the mask with a conventional ventilation setup and a pneumotachometer was attached in a junction in front of the ventilator to measure the patient's breathing rate. The figure below shows the results of the subject's breathing pattern.
The subject was asked to talk and move around with the ventilator on. As a result, small moments of leak were induced, as shown in the data above at ~40 and ~80 seconds. The mask quickly resealed though, validating the performance of our mask. The mask was taken off fully at ~130 seconds and then put back on at ~160 seconds to conclude the test.