Author Topic: The Reliability of the SHT-31 Humidity Sensor & What Psychrometer Should I Buy?  (Read 106702 times)

0 Members and 1 Guest are viewing this topic.

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Sensor mishandling according to the manufacturers handling document is not uncommon. If you inquire they will just say the handling guidelines are overkill. It's like every time I step foot into a weather office and the computer tech has computer memory modules just thrown around on his work bench. If you inquire about why don't you keep the memory in an ESD bag, use an anti-static wrist strap with an anti-static work mat, you will get told, "Been doing it this way for 20 years. Never had a problem." You'll get the same response with sensor handling. Unless they see a high failure rate, it's business as usual. [emphasis mine]

Except in this instance, they ARE having a problem. We have documented it pretty thoroughly in this thread. It was you, in fact, who stated you had never seen "drift" like what exists in these sensors.

I'd like someone to explain, for instance, how putting a pool or drain filter style cap over a sensor that needs microscopic protection won't predictably destroy that sensor's ability to give accurate readings in relatively short order. Then go ahead and square that with Davis' literature stating you shouldn't need recalibration for humidity within 5 years.

You can say "well that's just business as usual". And I can say you get what you expect. It works both ways.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline kcidwx

  • Forecaster
  • *****
  • Posts: 330
Just for fun, I still have my three SHT-31 eval boards. In the humidity chamber this morning they were reading:

Stability Time: 42 minutes
Ref Humidity: 45.0%
SHT-31-1: 46.31%
SHT-31-2: 46.20%
SHT-31-3: 46.97%

Number 3 went outside for 6 hours in the shade without a filter on it. And now......

Stability Time: 48 minutes
Ref Humidity: 45.0%
SHT-31-1: 47.12%
SHT-31-2: 47.03%
SHT-31-3: 50.02%

Number 3 is not happy.  :lol:


 
Meteorological Technician
NWS Certified Aviation Weather Observer
B.S. Meteorology

Offline kcidwx

  • Forecaster
  • *****
  • Posts: 330
Just for fun, I still have my three SHT-31 eval boards. In the humidity chamber this morning they were reading:

Stability Time: 42 minutes
Ref Humidity: 45.0%
SHT-31-1: 46.31%
SHT-31-2: 46.20%
SHT-31-3: 46.97%

Number 3 went outside for 6 hours in the shade without a filter on it. And now......

Stability Time: 48 minutes
Ref Humidity: 45.0%
SHT-31-1: 47.12%
SHT-31-2: 47.03%
SHT-31-3: 50.02%

Number 3 is not happy.  :lol:
Two "points" I see: (1) 48 minutes to stabilize means "diffusion" or extremely SLOW response time; and (2) the OEM filter probably is inhibiting contaminates.

Stability time just means the amount of time they were in the chamber. I always give the sensors a minimum of 30 minutes in the chamber before I read them. I provide that because people will always ask, "How long did you leave them in there before you took a reading?"
Meteorological Technician
NWS Certified Aviation Weather Observer
B.S. Meteorology

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
As promised, here are my initial measurements with my new Kestrel 5000 meter calibrated at the factory for +/- 2% and +/- 0.9F Temp.  All readings are compared against an 8 month old SHT31 installed in the field in a Davis 24hr FARS shield (about half it's life was spent in a passive shield though). The SHT31 sensor has the Davis filter over it.

I took a few readings between 80 and 90% humidity and, as expected, it runs cool (dry) in the dew point but only between 0.5 - 1.0 degF. That's 1-2% in the lower to mid 70s F. Cannot compare humidities directly because of solar error in the air temp during the day.

Around 75% it seems to be accurate (0.3 - 0.5F within Kestrel DP). So this must be near the crossover point in the curve if we were to draw one modeling the non-linearities of the sensor.

In the 60 percentiles is when the Davis sensor starts running noticeably wet. For example, at 82.5F and 61% humidity it ran 1.4 degrees warmer (wetter) in dew point compared to the Kestrel. That's 3% high in humidity which is out of spec. By comparison, my 15 month old SHT31 was running 4.8 degF warmer (wetter) in DP than the Kestrel with a humidity error above 8% at 61% reference humidity.

I will continue testing and post back if anything interesting comes up.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
Alright testing. More the merrier  =D>

I'm in the frustration mode after finding out my sensors aren't really new just not aspirated yet.  ](*,)


 
Randy

Offline kcidwx

  • Forecaster
  • *****
  • Posts: 330
When I requested my eval boards from Sensirion, I requested them with the filter membrane.

https://www.sensirion.com/fileadmin/user_upload/customers/sensirion/Dokumente/0_Datasheets/Humidity/Sensirion_Humidity_Sensors_SHT3x_Datasheet_Filter_Membran.pdf

So on the Davis sensor PCB's, there's absolutely nothing but the Davis cap and filter?
Meteorological Technician
NWS Certified Aviation Weather Observer
B.S. Meteorology

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
So on the Davis sensor PCB's, there's absolutely nothing but the Davis cap and filter?
Yes.
I wish I had seen this filter option earlier, putting on that tiny cap is going to be fun.... ](*,)

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
Just for fun, I still have my three SHT-31 eval boards. In the humidity chamber this morning they were reading:

Stability Time: 42 minutes
Ref Humidity: 45.0%
SHT-31-1: 46.31%
SHT-31-2: 46.20%
SHT-31-3: 46.97%

Number 3 went outside for 6 hours in the shade without a filter on it. And now......

Stability Time: 48 minutes
Ref Humidity: 45.0%
SHT-31-1: 47.12%
SHT-31-2: 47.03%
SHT-31-3: 50.02%

Number 3 is not happy.
  :lol:

Seems to get worse as you get toward the middle (50% humidity). It doesn't take much exposure to knock them out of calibration. This is appalling, though.
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
Wouldn't it be wonderful if this was our sensor   \:D/
 
The SHT3x membrane is a PTFE foil, dedicated to
protect the sensor opening from water and dust
according to IP67. It will thus allow for sensor use under
harsh environmental conditions, where spray water and
high exposure to dust is challenging for accurate sensor
performance
. The membrane cover is attached to the
sensor package by a double faced adhesive tape, which
is designed to stay on the package over the whole
sensor life time. Due to the minimum enclosed air
volume, the response time of the RH signal is the same
as the value achieved by the uncovered sensor


Oh this probably cost too much maybe extra .50 cents per sensor. Our sensors don't even come with a .20 cent bag to keep fresh and protected.   #-o What was I thinking.
Randy

Offline JudinNorman

  • Senior Member
  • **
  • Posts: 65
What is Davis reply so far if any of you has contacted them?

A little over a year since installing the Sht31. Last summer my dewpoints were tracking the nearest weather service station but this year my dewpoints are usually a few degrees higher.

Looks like my sensor is out of calibration to.

Not that I can't afford a new one, it's the principle that the sensor didn't last or perform per their specs.

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
They admit nothing.. 
I know how they roll from experience when I contacted them about firmware issue on Vue console and VP2 anemometer calibration they denied any issue existed even when I made a video.

Then they come out with console update a few months later, first update in years and issue was fixed. Still no acknowledgement...
They admit nothing...And won't here either.

Our best hope is they start packaging and handling humidity sensors correctly, but I wouldn't count on it.
Randy

Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 1237
    • WUnderground
What is Davis reply so far if any of you has contacted them?

...

As stated previously, I contacted the Davis support person (Brett Lane) and alerted him to the discussion in this thread and provided him a link hereto. He acknowledged receipt of my email but nothing more. If I were Davis I would be very careful posting anything in this Forum, rather I would post something on the Davis website, then post in this thread a link to the Davis website. My guess, Davis is talking internally whether or how to address these issues on its website. If I was a betting man, they will not publish anything; rather just consider our comments should they decide to replace the horribly aging VP-2!

Unrelated to the above, I have been reading this Forum for about 5 years and lately contributing, and there has been considerable fluff in many threads. Not this one. I just spent several hours re reading this thread and identified several useful takeaways. When I finish my list, I will seek editing review by contributors to this thread then post so folks don't have to spend hours reading this thread--I will call it the Cliffs Notes of SHT 31 issues. Are Cliffs Notes still published? Setting the record straight by a very knowledgeable poster, kcidwx, has been most refreshing.
« Last Edit: July 13, 2018, 10:13:19 PM by WheatonRon »
Davis VP2 with SHT31 (3 complete VP2 systems—2 with a daytime fan and 1 that has a 24 hour fan); CWOP--CW5020, FW3075 and FW4350; WU--KILWHEAT17, KILWHEAT36 and KILWHEAT39; WeatherCloud.net; CoCoRaHS--IL-DP-132; and Weatherlink 2.0

Offline JudinNorman

  • Senior Member
  • **
  • Posts: 65
They admit nothing.. 
I know how they roll from experience when I contacted them about firmware issue on Vue console and VP2 anemometer calibration they denied any issue existed even when I made a video.

Then they come out with console update a few months later, first update in years and issue was fixed. Still no acknowledgement...
They admit nothing...And won't here either.

Our best hope is they start packaging and handling humidity sensors correctly, but I wouldn't count on it.


Most consumers don't test like you guys do on here. If we call Davis they just would say weather is not same even a mile away and get away with it as the consumer would agree the weather in my backyard is just that, my weather.

Why couldn't Davis recall the sensors and fix or replace them like auto industry does when parts are found defective ??

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
They admit nothing.. 
I know how they roll from experience when I contacted them about firmware issue on Vue console and VP2 anemometer calibration they denied any issue existed even when I made a video.

Then they come out with console update a few months later, first update in years and issue was fixed. Still no acknowledgement...
They admit nothing...And won't here either.

Our best hope is they start packaging and handling humidity sensors correctly, but I wouldn't count on it.
Why couldn't Davis recall the sensors and fix or replace them like auto industry does when parts are found defective ??
That's when it's a safety issue, the DOT makes them under US law.

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
What is Davis reply so far if any of you has contacted them?

...

As stated previously, I contacted the Davis support person (Brett Lane) and alerted him to the discussion in this thread and provided him a link hereto. He acknowledged receipt of my email but nothing more. If I were Davis I would be very careful posting anything in this Forum, rather I would post something on the Davis website, then post in this thread a link to the Davis website. My guess, Davis is talking internally whether or how to address these issues on its website. If I was a betting man, they will not publish anything; rather just consider our comments should they decide to replace the horribly aging VP-2!

Unrelated to the above, I have been reading this Forum for about 5 years and lately contributing, and there has been considerable fluff in many threads. Not this one. I just spent several hours re reading this thread and identified 11 useful takeaways. When I finish my list, I will seek editing review by contributors to this thread then post. Setting the record straight by a very knowledgeable poster, kcidwx, has been most refreshing.

Thanks Ron! This forum has been very educational for me. Hopefully Davis will take consideration to the information on this thread and make the necessary changes. Thanks gang for all your input!  Please keep them going.
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
I should add just because they (Davis Inst) probably aren't going to admit any wrong doing something good may  still come from thread.
It brought to everyone's attention humidity has issues and proper handling of these sensitive sensors is important. Davis did fix the problem with Vue just didn't admit the problem existed.
Randy

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
I'm testing three of these evaluation boards right now. RH spec: 1.5%

https://www.idt.com/products/sensor-products/humidity-sensors

I'll run them against my three SHT-31 evaluation boards as well.

How did those sensors perform?
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Just reading the spec sheet, Davis would do well to get their transmitter to take I2c and switch over to IDT’s HS 3001. That’s IMO.

But aren't SHT35 and HS3001 pretty much on a par? Actually SHT35 looks better in the 90-100% region. Where would HS3001 score over SHT35?

Not saying there aren't differences - they're just not too obvious from the specs, other than details like 0.1% difference in nominal tolerance which may well get lost anyway in slightly different interpretation of the specs - eg what is the difference between tolerance and accuracy or nominal vs max - and differences in eg binning. And yes Davis would need to move away from Sensibus to modern I2C if these were to be relevant to use in VP2 units.

johnd- After carefully reading through the datasheet of the HS3001, I think you’re correct in your overall thinking. At first, I thought the HS3001 was a more “durable sensor” and it maybe is (when it comes to performance while the sensor is contaminated with dust or debris) but that’s about it. Other than that, it behaves basically the same as the Sensirion sensors. Here is what the HS 3001 datasheet says:

High humidity conditions:

“Important note: The HS300x series sensors are optimized to perform best in the more common temperature and humidity ranges of 10°C to 50°C and 20% RH to 80% RH, respectively. If operated outside of these conditions for extended periods, especially at high humidity levels, the sensors may exhibit an offset. In most cases, this offset is temporary and will gradually disappear once the sensor is returned to normal temperature and humidity conditions. The amount of the shift and the duration of the offset vary depending on the duration of exposure and the severity of the relative humidity and temperature conditions. The time needed for the offset to disappear can also be decreased by using the procedures described in sections 10 and 11.”


“Storage and Handling
Recommendation: Once the sensors are removed from their original packaging, store them in metal-in antistatic bags.
Avoid using polyethylene antistatic bags as they may affect sensor accuracy.
The nominal storage conditions are 10 to 50°C and humidity levels within 20% to 60%RH. If stored outside of these conditions for extended periods of time, the sensor readings may exhibit an offset. The sensor can be reconditioned and brought back to its calibration state by applying the following procedure:
1. Bake at a temperature of 100°C with a humidity < 10%RH for 10 to 12 hours.
2. Rehydrate the sensor at a humidity of 75%RH and a temperature between 20 to 30°C for 12 to 14 hours.”


Soo.... I think the best thing for Davis is to update their transmitter board to take pure I2c and switch to the SHT 35. But even more importantly, be sure that all sensors are packaged correctly.
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
We really could use the special coating to protect from contamination.
As mentioned humidity sensors are not designed to be continually aspirated.
Not taking these special precautions will surely end sensor life sooner than later.
And yes handling and storage and not open air is very important. Open air starts the end of live for sensor prematurely who knows how old they are when we get them. 12 months 24?  Who knows. 
Randy

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
I’m curious to know why Belfry Boy hasn’t make his version of the SHT-31 sensor like he did the 15?
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline kcidwx

  • Forecaster
  • *****
  • Posts: 330
I'm testing three of these evaluation boards right now. RH spec: 1.5%

https://www.idt.com/products/sensor-products/humidity-sensors

I'll run them against my three SHT-31 evaluation boards as well.

How did those sensors perform?

The LCD screen died on one of the eval boards so the testing is on hold until I get a replacement.
Meteorological Technician
NWS Certified Aviation Weather Observer
B.S. Meteorology

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
We really could use the special coating to protect from contamination.
As mentioned humidity sensors are not designed to be continually aspirated.
Not taking these special precautions will surely end sensor life sooner than later.
And yes handling and storage and not open air is very important. Open air starts the end of live for sensor prematurely who knows how old they are when we get them. 12 months 24?  Who knows.

But man, it’s definitely frustrating to get a “Aged” sensor when you paid for a fresh new one
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
But just thinking about it, I bet this is the reason why RainWise hasn’t produced a FARS version of the MK III.

But I do believe that the RW stations have a similar problem in terms of humidity as the Davis except for in the 90 percentile humidities.

Just looking at some of RW’s data on the company’s website, their temperature readings have a warm bias (most likely due to the shield they use).

In response to the ASOS being warmer than the Davis: I think the ASOS stations may have a warm bias. There is a MMTS at the Extension office in my county and they also have the Davis VP2 locates inside the same fence. The Davis (24hr FARS might add) reads sometimes warmer or colder than the MMTS. Typically by 1°
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6362
    • Valentine Nebraska's Real-Time Weather
But just thinking about it, I bet this is the reason why RainWise hasn’t produced a FARS version of the MK III.

But I do believe that the RW stations have a similar problem in terms of humidity as the Davis except for in the 90 percentile humidities.

Just looking at some of RW’s data on the company’s website, their temperature readings have a warm bias (most likely due to the shield they use).

In response to the ASOS being warmer than the Davis: I think the ASOS stations may have a warm bias. There is a MMTS at the Extension office in my county and they also have the Davis VP2 locates inside the same fence. The Davis (24hr FARS might add) reads sometimes warmer or colder than the MMTS. Typically by 1°

The ASOS here for sure plus bias +2° on average. Don't see that with nearby AWOS or any of the Mesonet stations in area. All follow real close with aspirated Davis. 

You know who believes the ASOS are those with thermometers on the porches.   :lol:

I said this in jest my neighbor lady says the airport matches her porch thermometer...I told her airport thermometer is in middle of open field nothing but grass and hay no heat source at all and your porch thermometer is 2" off siding of house on southside of all places. What's wrong here? 
« Last Edit: July 13, 2018, 10:03:11 PM by ValentineWeather »
Randy

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6730
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
But just thinking about it, I bet this is the reason why RainWise hasn’t produced a FARS version of the MK III.
Even if that's true, which I highly doubt, I don't care if fans are bad for the humidity portion of the sensor, I'll never quit using one til something comes along that's superior.
The benefits a FARS to me far outweighs the potential negative.

 

anything