Topic: The Reliability of the SHT-31 Humidity Sensor & What Psychrometer Should I Buy?

[jgentry]

After the last two days of rain, rare, heavy fog enveloped southern AZ last night. Didn't quite get to W0X0F (at least while I was awake), but very close. My SHT31 refused to get closer than 0.1F in temp/dew spread, resulting in 99% humidity for almost seven hours til 10 this morning. Not stating anything new, just wish it would pull that 100% trigger. This sensor is about two years old. Anyone getting consistent 100% out of newer ones?

With the WeatherFlow Tempest SHT-31, I hit 100% easily. With the Davis, naw. But I haven’t used a brand new sensor. The last time I’ve texted Ryan, he told me Davis hasn’t changed anything so the new ones would probably perform about the same.

[CW2274]

Harrumph. That's what I figured. Thanks. 

[drew1021]

Mine did but only for a short period after I installed new. Now 97% is the highest I’ve seen it since. Sensor is about a year old

[Mapantz]

The SHT31-7346-070 has been bang on for me.

Now we've hit the damper end of the year, it will go to 100%. November several times, and December. Quite a few 99% days too.

[jgentry]

I just installed a brand new Davis VP2 station at Petals From The Past in Jemison, AL. So far the humidity sensor is performing well with no wet bias. 

You can see the comparison here with the nearby ASOS

Petals From The Past https://www.weatherlink.com/embeddablePage/show/66eab9495ce5496ba281d83872dcb2bf/summary



Shelby County Airport (ASOS) https://mesowest.utah.edu/cgi-bin/droman/meso_base_dyn.cgi?stn=KEET&unit=0&time=&hours=24&hour1=00&day1=11&month1=12&year1=2023&radius=25&past=0&order=1

[CW7491]

That’s great to hear. Let us know how it does over time. I’ve been initially encouraged to only see the usual wet bias return. Seems it’s just a limit of the technology. Fortunately I have an old analog sensor 7346.029 that is really very accurate. It hits 100% easily and dries out very well. If anything it tends to go to saturation too easily.

I’ve written both customer service and tech support at Davis within the last month about a potential improved sensor. Unsurprisingly they would neither confirm or deny any potential new temp/humidity sensor product.

[jgentry]

Would there be an improved sensor out on the market that Davis can use that wouldn’t cause a huge increase in the cost of the product?

[CW7491]

That’s a great question I don’t have the answer to. I asked them in the same email if they had any information on the supplier of the analog humidity sensing element they used in the 7346.029 and they also declined to answer. Funny how sometimes the older technologies seem to work better than the new. I know there has been speculation that Davis could no longer get a supply of that part.

Sensirion is on to the SHT4X series now, but presumably the those sensors will suffer from the same shortcomings not to mention the communication protocol problem for Davis with I2C, which I presume will be the expensive aspect for them. Also, if you look at the Sensirion website, of all the applications they advertise using their humidity sensors for, weather observations/stations is not among them.

[JCA433]

CW, it is unfortunate Davis declines to answer your questions in your email.   The best solution right now is to calibrate the humidity sensor using the Cumulus software or other good calibration software. With Cumulus I can reduce the wet bias in humidity readings and get the dew point reading within +- 2F of my local reference.  It is not perfect though and there will still be error.


The trouble with the calibration is that my sensor reads humidity accurately between 75% and 93% and reads too low above 93% and too high below 75%. so my corrected humidity reading is too high between 75% and 93%.  It is quite good below 75% and above 93%.

[CW7491]

Thanks, yes, that’s exactly what I would do when I had to use the Davis SHT31 sensor. I’d use offset and slope to deal with the inherent biases. It works fairly well, but it’s not a perfect fit. The trouble is in the midrange there is still a wet bias and when the humidity really drops, the correction can be excessive. I’d use 1.071x-6 and I found that to be about as good as you could do.

I found two of the older analog sensors and after some refurbishment they are superior to the SHT31. These sensors are almost 20 years old and had a dry bias. Davis would not help me when I asked if rinsing the humidity element in distilled water would bring it back into tolerance. I figured I had nothing to lose and tried it on one of my sensors. It brought it right back into spec. Amazing how well they work given how old they are. Unfortunately when these go, that’s it. Hopefully Davis has a better solution by then