5 One Shot Heater Correction

One-shot heater recovery is the appropriate mode when the sensor experiences an occasional high-humidity event rather than a regularly recurring one. The experiment in Table 1 emulates that use case: measure initial RH error, soak the devices at 25°C and 90%RH for seven days, measure the shifted RH error, then apply one full-power heater pulse of varying duration and measure RH error again across 30%RH, 60%RH, and 90%RH. The results captured in Figure 5-1 are not captured over time, but over RH setpoint. Because the high-humidity exposure is temporary, the results in Figure 5-1 must be treated as a recovery experiment, not merely a sequence diagram.

90%RH was chosen because it is a known high-humidity condition that induces RH drift in a matter of hours without risk of condensation occurring during the experiment. This was done for seven days since that is long enough to induce enough RH drift to push the HDC3020 out of the RH accuracy specification. The HDC3020 has configurable heater power levels, so the full power setting was chosen to achieve the largest temperature rise possible. A larger temperature rise increases the diffusion rate of the water vapor molecules, hence providing better RH drift mitigation. One minute is usually sufficient to achieve a steady-state temperature rise. To explore the question of how long the heater must be run at steady state, some devices were run for longer to see how much benefit was gained.

Figure 5-1 One-Shot Heater Correction: RH Error vs RH Setpoint

In Figure 5-1, RH error is plotted against RH setpoint. This is not showing how the RH is changing from the initial result over time but instead showing how the RH error has changed across RH when the different steps are applied. Each line represents an average of the results for 17 HDC3020s. The RH error was measured at 25°C across multiple RH levels. The green line shows the initial time-zero RH error. The red line shows the RH error after 7 days of exposure at 25°C/90%RH, with no heating done to remove the RH drift. The blue & purple lines show the RH error after running the heater for variable durations.

One minute of full-power heating already pulled the averaged HDC3020 much closer to the original performance, but two minutes delivered a materially better result at 60%RH and 90%RH. Five minutes only marginally improved the 90%RH result further. The strongest engineering takeaway is therefore simple: on this board, 2 minutes at full heater power is the best starting point for one-shot recovery, while 5 minutes is better treated as an optional high-RH refinement when power and cycle time are less constrained. If judging solely at ambient RH levels between 30%RH and 60%RH, 1 minute of heating offers almost the same level of correction as the longer heating times while saving power consumption by running the heater for at least half as long as the other options.