0

SHT40 question
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

SHT40 question

by MarkEmK on Sun Jul 25, 2021 5:03 pm

I recently bought a couple of these temperature/humidity sensors to experiment with. Rather than use the Adafruit SHT40 library, I just used the Wiring library. It wasn't too difficult to get the sensor and Teensy LC talking via Wiring, but I noticed something regarding the temperature value vs. chip heater settings. Even with the shortest, lowest-power setting (20mw for .1 second), the indicated temperature was about 7 degrees hotter than readings done with no heater.

Since my main interest in using this sensor is to evaluate the performance of various homebrew A/C schemes, this kind of temperature error is significant. I'm wondering if any other users of this chip (and similar ones) have noticed the same thing.

After reading the data sheet, I had wondered how the sensor readings could be all that accurate when using the heater, and, apparently, they aren't if used incorrectly. Perhaps a two-sequence command set is needed -- one with the heater (to dry out the sensor), followed by a short delay so the chip temperature is closer to ambient, and then a final readout with NO heater.

I guess one way to check this would be to write some code to (1) apply a heater pulse and (2) monitor the temperature over time using no-heat commands.

Mark

MarkEmK
 
Posts: 12
Joined: Sun Jul 25, 2021 4:46 pm

Re: SHT40 question

by MarkEmK on Sun Jul 25, 2021 6:16 pm

Here's a plot showing chip temperature vs. time. The first measurement was done with the heater (lowest power, least time), followed by a 250ms delay, then 500 no-heater samples in approximately 10millisecond intervals (not counting the time for the I2C transaction). The 250ms delay is needed to properly acquire the first measurement.

https://drive.google.com/file/d/12n16DIKzQ31_laIr15IsvrIjF3JTzTBF/view?usp=sharing

There is little doubt that the temperature measurement is significantly perturbed when using the heater. So if it is necessary to use the heater users will have to perform a follow-up measurement that's done without the heater, or characterize their system and code to compensate for the effect; or live with the inherent error produced by the heater.

If the sensor is not exposed to liquid water or extremely high humidity, it probably is not necessary to use the heater. One thing to beware of is condensation on the sensor.

Mark

MarkEmK
 
Posts: 12
Joined: Sun Jul 25, 2021 4:46 pm

Please be positive and constructive with your questions and comments.