2 Test Methodology

A custom test enclosure was designed to evaluate water ingress detection under realistic conditions. The enclosure was built to be robust across a range of temperatures and humidity levels, flexible enough for various test scenarios, and representative of a typical vented electronics housing.

One additional scenario was evaluated before the main experiment with the sensor outside of the enclosure, where a water droplet was placed directly on the HDC3020 sensor cavity. In such a case, the sensor reading momentarily drops to 0%RH for a few seconds, then rises above the original value until the water evaporates or is removed (after which normal readings resume). The drop to 0%RH is nearly instantaneous. This distinctive signature unambiguously indicates water reaching the sensor cavity, independent of the variables tested within this study such as ambient conditions, enclosure type or size, water volume, and so forth. In cases where liquid water is expected and such behavior must be avoided, an alternative device option is the HDC3022, which comes with a waterproof IP67 rated filter.

The test enclosure was used to emulate a semi-open system with a waterproof vent at the top to allow air exchange. The vent served two purposes: this mimicked real-world enclosures (which are rarely hermetically sealed) and this prevented pressure buildup during testing when conditioned air was pumped in. Without the vent, introducing humid or dry air from an external chamber can over-pressurize the box. The additional pressure can change the AH in a way other than water entering the system.

Inside the enclosure, a test PCB held the HDC3020 sensor, which was mounted on a small raised board so this cannot be submerged by any incoming water. Water was introduced through a funnel, simulating either a slow drip or a rapid pour. A heatsink and heater were added next to the PCB to vary the temperature inside the enclosure. This allowed the box to be heated to elevated temperatures without needing to place the entire enclosure inside of a heated environmental chamber. Humid and dry air was supplied through a pneumatic tube connection to a humidity chamber directly into the test enclosure.

Testing was performed under a variety of ambient conditions to make sure the detection algorithm works across the range of scenarios in which water ingress can occur. In automotive and industrial applications, leaks can happen in cold and dry climates as well as hot and humid ones. Accordingly, the system was evaluated at several representative conditions.

One drop of water was measured to be approximately 0.023mL, while three drops of water was approximately 0.07mL. For the drop tests, after acclimating the box to the set humidity level, the humid air tube was removed and water was deposited into the enclosure through the brass fitting hole with a dropper. For full flow (100mL) of water tests, a funnel was used to pour the water into the same fitting hole. After water was deposited into the enclosure, the brass fitting hole where the humid air can be piped in was closed, to prevent the evaporating water and humid or dry air from escaping.

All test conditions were evaluated with three drops of water to simulate a minor water leak. The "full flow of water" tests were designed to simulate a much larger volume of water breaking into the system. The overtemperature tests were not done with the larger volume of water, because pouring lots of water into the enclosure requires extensive clean up between tests and often requires a disassemble of the test enclosure. Furthermore, using a smaller volume of water is the worst-case test for the humidity sensors to detect because much less water vapor can be evaporated into the air, representing a stronger test of the proposed RH slew rate threshold method. Lastly, a one-drop test was performed for test 1 to push the boundaries on how fine of an Ingres event the HDC3020 can reliably detect at the given test enclosure size.