Possibly defective HTU31D

[dzaharia]

Hello. I recently purchased an HTU31D temperature and humidity sensor. I am using it connected to a PyPortal to report both temperature and relative humidity. The temperature readings are fairly accurate, but I'm finding the humidity readings to be suspiciously low. I have a digital humidifier in the same room, and it reports readings as much 50% higher than the HTU31D.

Here are the readings from my humidifier's humidity sensor:

Screenshot_20230328-140410.png (100.49 KiB) Viewed 73 times

...and from the HTU31D over the same period:

Screenshot_20230328-140437.png (84.82 KiB) Viewed 73 times

I explicitly set the heater to off in my code to ensure that that wouldn't interfere with the readings. Any ideas of what the issue could be?

[adafruit_support_carter]

Could it be location related? Somehow one sensor is seeing more moisture than the other?

[dzaharia]

I thought about this, but even when I put the two sensors directly beside one another, the readings can be vastly different. I originally experimented with proximity because my first suspicion was that the humidifier (being, you know, a humidifier) is by default in the most humid part of the room. But even when the humidifier itself is off for hours, therefore giving moisture time to diffuse throughout the room, the two sensors have radically different readings.

[adafruit_support_carter]

In the two plots above, they at least seem to be trending in the same way. Same general spikes, etc. in the data trend with time. Does that seem to always be the case?

[dzaharia]

Hm not really. Here are the two plots for today’s data from 12 midnight until now. That short spike to 40% on the sensor plot happened while I was holding my palm over the sensor to see if it would react to the humidity of my hand. Note, especially the inverse trends on each plot at midnight, when it was 80% humidity according to the humidifier. I will also note that the outdoor humidity where I live at that time was somewhere around 90%, so the humidifier reading is more logical than the HTU.

Screenshot_20230328-153120 Medium.png (57.49 KiB) Viewed 64 times

Screenshot_20230328-153040 Medium.png (72.13 KiB) Viewed 64 times

[millercommamatt]

Relative humidity is temperature dependent. Are you sure the temperature is accurate for the sensor and your humidifier? If one is reading a temperature too low, the humidity will read too high and if the temperature reading is too high, the relative humidity will read too low. The fact that the sensor and your humidifier show the same patterns suggests that the sensor is working. The cooling effect from evaporation from your humidifier might bias it's relative humidity assessment high. Self-heating in the sensor on your breakout might bias it's relative humidity measurements low.

Can you convert to dew point and compare those numbers?

Anecdotally, I've found all solid state humidity sensors to be wildly variable from one another regardless of what the manufacturer specifications state.

[dzaharia]

Ah that’s super helpful context! Can you tell me more about dew point? I don’t know too much about the science behind this topic. Based on the temperature sensor’s readings compared with other temperature sensors in the same room, I believe that it’s pretty accurate. When you say self-heating, do you mean warmth emanating from other components around the sensor?

[millercommamatt]

When you say self-heating, do you mean warmth emanating from other components around the sensor?

That and the sensor itself. This breakout is designed to minimize this effect.

https://en.wikipedia.org/wiki/Dew_point