Author Topic: The Reliability of the SHT-31 Humidity Sensor & What Psychrometer Should I Buy?  (Read 31537 times)

0 Members and 1 Guest are viewing this topic.

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
Issue isn't linear at all.  Everything is hunky dory just a nice typical summer day mid 88 and DP 67° on average with heat index only 91° around here. 
It's always near high heat index advisor levels above 100° the issue arises for me.

Once temps get into 90's and dewpoint 70-80.
 
Example Actual T93°,  DP 75, Hum 56% ,Heat Index 103°

SHT31 will show T93°, DP 78, Hum 62% Heat index 109°

That's the issue.
Randy

Offline CW2274

  • Forecaster
  • *****
  • Posts: 3989
    • Conditions @ CW2274
Issue isn't linear at all.
That's the issue.
I get that. I supplied info on what's going on with my conditions now, and regarding issues, baseless or not, that have been brought up with this sensor, simple as that. Not saying there isn't a problem, merely sharing what I observed. My ISS will be accessible til I decide to put it back out to pasture, and I'll continue testing, if nothing else but for myself. If I find something that may be relevant, I'll share.

Offline Dj1225

  • Member
  • *
  • Posts: 35
Issue isn't linear at all.  Everything is hunky dory just a nice typical summer day mid 88 and DP 67° on average with heat index only 91° around here. 
It's always near high heat index advisor levels above 100° the issue arises for me.

Once temps get into 90's and dewpoint 70-80.
 
Example Actual T93°,  DP 75, Hum 56% ,Heat Index 103°

SHT31 will show T93°, DP 78, Hum 62% Heat index 109°

That's the issue.

Yes! That was the issue I was having and like I said in an earlier post the humidity down on the Gulf Coast is unbearable but when you start noticing the heat index becoming outrageous and laughable (close to 120°) when around the area it's between 100°-108° I knew that something wasn't right, then I came across this topic. My new sensor I received Friday showed better results than my 2+ year old sensor, instead of mid 60s-low 70s humidity percentage while in the low to mid 90s it has been showing mid 50s to low 60s percent range. As far as the 100% issue in 2 days it's topped out at 97% which I hadn't seen with the old sensor in almost a year. It's just very frustrating.

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
Issue isn't linear at all.  Everything is hunky dory just a nice typical summer day mid 88 and DP 67° on average with heat index only 91° around here. 
It's always near high heat index advisor levels above 100° the issue arises for me.

Once temps get into 90's and dewpoint 70-80.
 
Example Actual T93°,  DP 75, Hum 56% ,Heat Index 103°

SHT31 will show T93°, DP 78, Hum 62% Heat index 109°

That's the issue.

Yes! That was the issue I was having and like I said in an earlier post the humidity down on the Gulf Coast is unbearable but when you start noticing the heat index becoming outrageous and laughable (close to 120°) when around the area it's between 100°-108° I knew that something wasn't right, then I came across this topic. My new sensor I received Friday showed better results than my 2+ year old sensor, instead of mid 60s-low 70s humidity percentage while in the low to mid 90s it has been showing mid 50s to low 60s percent range. As far as the 100% issue in 2 days it's topped out at 97% which I hadn't seen with the old sensor in almost a year. It's just very frustrating.

Glad you found the thread,
Now you know you aren't alone.  :sad:
Randy

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
I get that. I supplied info on what's going on with my conditions now, and regarding issues, baseless or not, that have been brought up with this sensor, simple as that. Not saying there isn't a problem, merely sharing what I observed. My ISS will be accessible til I decide to put it back out to pasture, and I'll continue testing, if nothing else but for myself. If I find something that may be relevant, I'll share.

Thanks, yes please do.
Keep those sensors even here they work good enough maybe average 2-3% high humidity winter but come summer and high dew points, watchout...
Randy

Offline CW2274

  • Forecaster
  • *****
  • Posts: 3989
    • Conditions @ CW2274
I get that. I supplied info on what's going on with my conditions now, and regarding issues, baseless or not, that have been brought up with this sensor, simple as that. Not saying there isn't a problem, merely sharing what I observed. My ISS will be accessible til I decide to put it back out to pasture, and I'll continue testing, if nothing else but for myself. If I find something that may be relevant, I'll share.
but come summer and high dew points, watchout...
Well, considering the monsoon is here..... As I look at this very moment comparing my dews with the "local gang" there's two 64F's and two 66F's. BTW, I'm using the 2 year old sensor at the moment. ;)

Offline jgentry

  • Forecaster
  • *****
  • Posts: 351
Issue isn't linear at all.  Everything is hunky dory just a nice typical summer day mid 88 and DP 67° on average with heat index only 91° around here. 
It's always near high heat index advisor levels above 100° the issue arises for me.

Once temps get into 90's and dewpoint 70-80.
 
Example Actual T93°,  DP 75, Hum 56% ,Heat Index 103°

SHT31 will show T93°, DP 78, Hum 62% Heat index 109°

That's the issue.


I deal with the same issues and it’s frustrating!!!
Davis Vantage Pro2. SHT-75. WU: KXALJEMI2 & KALTHORS2. CWOP/APRS: C6353 & E6358

  

Offline jgentry

  • Forecaster
  • *****
  • Posts: 351
Seems like jerryg sensor with the sf2 filter cap is performing well and the DPs appear to aligns with the Airport reports
Davis Vantage Pro2. SHT-75. WU: KXALJEMI2 & KALTHORS2. CWOP/APRS: C6353 & E6358

  

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
I'm looking at ASOS stations and they are doing the same. Skipping over 98 and 99% and jumping to 100%. Looks like they have same option using software like Cumulus where if sensor hits 98 jumps to 100%.

Omaha sensor can't make 98% like many SHT31's peaking at 97% with dense fog and mist.

Some images OMAHA,
North platte and Valentine




Randy

Offline jgentry

  • Forecaster
  • *****
  • Posts: 351
I noticed the same thing with the Airport that’s 15 milesnorth of me. Think I only seen 1 Airport station ever report 99% humidity.
Davis Vantage Pro2. SHT-75. WU: KXALJEMI2 & KALTHORS2. CWOP/APRS: C6353 & E6358

  

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
I noticed the same thing with the Airport that’s 15 milesnorth of me. Think I only seen 1 Airport station ever report 99% humidity.

Even the vaisala. I realize these are 5 minute averages but still, no 98 and 99 very suspect. I went back and downloaded 60 days of KVTN data and nothing.
Randy

Offline jgentry

  • Forecaster
  • *****
  • Posts: 351
I noticed the same thing with the Airport that’s 15 milesnorth of me. Think I only seen 1 Airport station ever report 99% humidity.

Even the vaisala. I realize these are 5 minute averages but still, no 98 and 99 very suspect. I went back and downloaded 60 days of KVTN data and nothing.

I think the accuracy of it in the 90 percentile is around +/-1.7%. SHT-31 is +/-2%
Davis Vantage Pro2. SHT-75. WU: KXALJEMI2 & KALTHORS2. CWOP/APRS: C6353 & E6358

  

Offline jgentry

  • Forecaster
  • *****
  • Posts: 351
My SHT-31 at my friend’s place has reached 100% for the third time in a row. It’s basically a new sensor but it does have the Davis filter cap.
Davis Vantage Pro2. SHT-75. WU: KXALJEMI2 & KALTHORS2. CWOP/APRS: C6353 & E6358

  

Offline Intheswamp

  • Early on one frosty morn'...
  • Forecaster
  • *****
  • Posts: 1385
  • Sure...I know enough to be dangerous, now.
    • Beeweather
Well, not to be one to rush things... 

I bought a new Sensiron SHT31 sensor from Ryan back in April of 2016.  It's been enjoying a dark, cool existence in a closet ever since. :roll:

I'm getting ready to replace the old sensor (and the rain gauge) and have been following this thread.  I've seen my heat indexes go out of sight for a while now, just not sure how long it's really been doing that.  My records only go back a year and supposedly the latest readings are the highest so maybe(?) this started since last year(?).  It seems I don't really remember the super-high heat index numbers (120'ish) until this year.

Anyhow, it will be interesting to see how things compare with the new(er) sensor installed...when I get it installed.  #-o

A question here (yeah, I need someone to hold my hand)...  What is the "Davis filter cap"???

Thanks!

Davis VP2, v3.12
Cumulus v1.9.4  Build 1079
Today's weather beats no weather at all!
www.beeweather.com

Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 892
    • WUnderground
The “white house” (filter cap) covering the sensor on the PC Board—see photo attached.
« Last Edit: July 16, 2018, 01:29:34 PM by WheatonRon »
Davis VP2 with 24 hour FARS, SHT31 (3 complete systems-2 for uploading to the internet the other system for test and play); CWOP--CW5020 and FW3075; WU--KILWHEAT17 and KILWHEAT36; WeatherCloud.net; CoCoRaHS--IL-DP-132; Rainwise 111

Offline hwcorder

  • Contract Weather Observer at KCLT
  • Senior Member
  • **
  • Posts: 73
Hi guys. Don't post here a lot but have been following this thread closely the last day or two since I have been a loyal Davis customer since 1993 when I got a WMII. Still miss that thing sometimes!

I am also a Contract Weather Observer at KCLT so I have a good bit of experience with ASOS. The reason you do not see 98 or 99 percent readings with ASOS is that it directly measures the DP then calculates RH values using ambient temp where Davis directly measures RH then calculates DP.  For example if ASOS measures 70/69 it calculates a RH of 97%. IF the DP moves up to 70 then of course the calculated RH is 100%. You've got to take the differences in the method of measurement into account when comparing your readings to airports.  As for the sensors reading up to 100% even the expensive ones like the HMP155 can have issues. Back a few years ago our ASOS would rarely hit 100% even in fog with visibilities of 1/4mi. Then they changed the sensor and it hits 100% with relative ease.  I used to obsess about getting to 100% too but don't anymore especially when I can adjust the higher readings using Weather Display. 

As for the SHT31 is was initially excited about its improved accuracy but quickly learned it suffered from the same amount if not more drift towards a wet bias as its predecessors.  I think its more of a moisture problem than a dust problem.  I bought several sensors and when I first installed them I would track  about 1-2 degrees higher than surrounding ASOS stations two if which are within 5-7 miles of me (KCLT, KAKH). I do have a lot of vegetation around my house so I take that into account when comparing DP readings to other places. I also have a Viasala HM46 probe I do spot checks with. RH values would be generally the same < 3% difference.  However I would notice after the first few very humid spells where we would have a day or two of RH values remain 90% or above the sensor would never quite come back to its reading prior to that. Differences in RH between the SHT31 and the Vaisala would be anywhere from 4-8% with the SHT31 higher.  I would buy another sensor and the same thing. 

Doing a little research I learned this is really a technological problem.  Electronics and weather specifically moisture do not mix. Therefore when using electronics to measure moisture there are a lot of problems to overcome to get accurate readings over the long term of say more than two years.  The DTS1 ASOS uses to measure DP comes with features to increase sensor longevity.  Its heated in high RH environments to protect it from condensation.  It also undergoes a high heat cycle I believe once a day to cleanse any chemical vapors that it may encounter. Even so, as has been mentioned prior in this thread the NWS replaces this very expensive sensor every 18 months.  I have recently come across a memo that NWS is looking to upgrade a few sensors for ASOS and the dew point sensor is on that list.

Anyway, bottom line is that if you really want to keep accurate readings in humid climates you need to replace these sensors every year to 18months.  I actually went back to the sht75 as they are very easy to switch out and actually seem to drift slightly less than the board mounted sensors. I wired in a female connector then mounted it through a slot I cut into the sensor plate in the FARS unit. Now when I change out the sensors I simply pull out the old one and replace it with the new about every 18 months.  I still use the Davis filter and replace that every year.  Senserion is phasing out the SHT75 however so I think I'm going to use Jerryg's idea and buy a 31 and replace the Davis filter with the SF2 using plumbers putty to see how that works out.

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
Well, not to be one to rush things... 

I bought a new Sensiron SHT31 sensor from Ryan back in April of 2016.  It's been enjoying a dark, cool existence in a closet ever since. :roll:

I'm getting ready to replace the old sensor (and the rain gauge) and have been following this thread.  I've seen my heat indexes go out of sight for a while now, just not sure how long it's really been doing that.  My records only go back a year and supposedly the latest readings are the highest so maybe(?) this started since last year(?).  It seems I don't really remember the super-high heat index numbers (120'ish) until this year.

Anyhow, it will be interesting to see how things compare with the new(er) sensor installed...when I get it installed.  #-o

A question here (yeah, I need someone to hold my hand)...  What is the "Davis filter cap"???

Thanks!

It's the stock filter Davis uses. The thinking is because of aspiration the sensor gets dirty and makes sensor less accurate. Sensirion has its own filter with better filtration but isn't designed for Davis sensor board.

What I've done is double up on stock Davis filter material and looks good in response time, no difference I see. All new sensors come with filter so just added old filter after washing along with new filter. This doubles filtration and hoping sensor will last longer.
I'll probably change sensor each spring anyway as long as price doesn't take off. Don't think it will the actual sensor is $4 and like electronics may actually trend down in price.
« Last Edit: July 16, 2018, 12:29:30 PM by ValentineWeather »
Randy

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
Hi guys. Don't post here a lot but have been following this thread closely the last day or two since I have been a loyal Davis customer since 1993 when I got a WMII. Still miss that thing sometimes!

I am also a Contract Weather Observer at KCLT so I have a good bit of experience with ASOS. The reason you do not see 98 or 99 percent readings with ASOS is that it directly measures the DP then calculates RH values using ambient temp where Davis directly measures RH then calculates DP.  For example if ASOS measures 70/69 it calculates a RH of 97%. IF the DP moves up to 70 then of course the calculated RH is 100%. You've got to take the differences in the method of measurement into account when comparing your readings to airports.  As for the sensors reading up to 100% even the expensive ones like the HMP155 can have issues. Back a few years ago our ASOS would rarely hit 100% even in fog with visibilities of 1/4mi. Then they changed the sensor and it hits 100% with relative ease.  I used to obsess about getting to 100% too but don't anymore especially when I can adjust the higher readings using Weather Display. 

As for the SHT31 is was initially excited about its improved accuracy but quickly learned it suffered from the same amount if not more drift towards a wet bias as its predecessors.  I think its more of a moisture problem than a dust problem.  I bought several sensors and when I first installed them I would track  about 1-2 degrees higher than surrounding ASOS stations two if which are within 5-7 miles of me (KCLT, KAKH). I do have a lot of vegetation around my house so I take that into account when comparing DP readings to other places. I also have a Viasala HM46 probe I do spot checks with. RH values would be generally the same < 3% difference.  However I would notice after the first few very humid spells where we would have a day or two of RH values remain 90% or above the sensor would never quite come back to its reading prior to that. Differences in RH between the SHT31 and the Vaisala would be anywhere from 4-8% with the SHT31 higher.  I would buy another sensor and the same thing. 

Doing a little research I learned this is really a technological problem.  Electronics and weather specifically moisture do not mix. Therefore when using electronics to measure moisture there are a lot of problems to overcome to get accurate readings over the long term of say more than two years.  The DTS1 ASOS uses to measure DP comes with features to increase sensor longevity.  Its heated in high RH environments to protect it from condensation.  It also undergoes a high heat cycle I believe once a day to cleanse any chemical vapors that it may encounter. Even so, as has been mentioned prior in this thread the NWS replaces this very expensive sensor every 18 months.  I have recently come across a memo that NWS is looking to upgrade a few sensors for ASOS and the dew point sensor is on that list.

Anyway, bottom line is that if you really want to keep accurate readings in humid climates you need to replace these sensors every year to 18months.  I actually went back to the sht75 as they are very easy to switch out and actually seem to drift slightly less than the board mounted sensors. I wired in a female connector then mounted it through a slot I cut into the sensor plate in the FARS unit. Now when I change out the sensors I simply pull out the old one and replace it with the new about every 18 months.  I still use the Davis filter and replace that every year.  Senserion is phasing out the SHT75 however so I think I'm going to use Jerryg's idea and buy a 31 and replace the Davis filter with the SF2 using plumbers putty to see how that works out.

Thanks for this information very good.
You might consider just doubling the Davis filter to avoid any chemical fumes that can damage sensor. SF2  filters are very small so whatever putty or glue used will surely get into sensor while drying.
Randy

Offline kcidwx

  • Forecaster
  • *****
  • Posts: 329
Daughters WMII pegged out all night long at 100% humidity in the fog. I really like this weather station. Too bad you can't get parts for it. I'd really like to know what brand the humidity sensor is. It tracks pretty well with my Vaisala HMT337. Either that or she just got lucky with this one.

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

Offline openvista

  • Forecaster
  • *****
  • Posts: 419
    • marquetteweather.com
You might consider just doubling the Davis filter to avoid any chemical fumes that can damage sensor. SF2  filters are very small so whatever putty or glue used will surely get into sensor while drying.

I don't really think doubling up that filter material will keep water off the sensor. That'd be like installing overlapping bug screens with 1-inch holes hoping that will keep the skeeters out. We're talking about tiny coagulates of water vapor that form when the air is near or at saturation.

I will investigate the possibility of mounting it with the clips when my caps arrive by drilling two very small holes in the board. That would be the ideal solution, but probably not possible. If not, then jerryg's solution might be the only one that's been field tested to work.
Davis Vantage Pro2 FARS | https://marquetteweather.com | EW7933

Offline jgentry

  • Forecaster
  • *****
  • Posts: 351
Hi guys. Don't post here a lot but have been following this thread closely the last day or two since I have been a loyal Davis customer since 1993 when I got a WMII. Still miss that thing sometimes!

I am also a Contract Weather Observer at KCLT so I have a good bit of experience with ASOS. The reason you do not see 98 or 99 percent readings with ASOS is that it directly measures the DP then calculates RH values using ambient temp where Davis directly measures RH then calculates DP.  For example if ASOS measures 70/69 it calculates a RH of 97%. IF the DP moves up to 70 then of course the calculated RH is 100%. You've got to take the differences in the method of measurement into account when comparing your readings to airports.  As for the sensors reading up to 100% even the expensive ones like the HMP155 can have issues. Back a few years ago our ASOS would rarely hit 100% even in fog with visibilities of 1/4mi. Then they changed the sensor and it hits 100% with relative ease.  I used to obsess about getting to 100% too but don't anymore especially when I can adjust the higher readings using Weather Display. 

As for the SHT31 is was initially excited about its improved accuracy but quickly learned it suffered from the same amount if not more drift towards a wet bias as its predecessors.  I think its more of a moisture problem than a dust problem.  I bought several sensors and when I first installed them I would track  about 1-2 degrees higher than surrounding ASOS stations two if which are within 5-7 miles of me (KCLT, KAKH). I do have a lot of vegetation around my house so I take that into account when comparing DP readings to other places. I also have a Viasala HM46 probe I do spot checks with. RH values would be generally the same < 3% difference.  However I would notice after the first few very humid spells where we would have a day or two of RH values remain 90% or above the sensor would never quite come back to its reading prior to that. Differences in RH between the SHT31 and the Vaisala would be anywhere from 4-8% with the SHT31 higher.  I would buy another sensor and the same thing. 

Doing a little research I learned this is really a technological problem.  Electronics and weather specifically moisture do not mix. Therefore when using electronics to measure moisture there are a lot of problems to overcome to get accurate readings over the long term of say more than two years.  The DTS1 ASOS uses to measure DP comes with features to increase sensor longevity.  Its heated in high RH environments to protect it from condensation.  It also undergoes a high heat cycle I believe once a day to cleanse any chemical vapors that it may encounter. Even so, as has been mentioned prior in this thread the NWS replaces this very expensive sensor every 18 months.  I have recently come across a memo that NWS is looking to upgrade a few sensors for ASOS and the dew point sensor is on that list.

Anyway, bottom line is that if you really want to keep accurate readings in humid climates you need to replace these sensors every year to 18months.  I actually went back to the sht75 as they are very easy to switch out and actually seem to drift slightly less than the board mounted sensors. I wired in a female connector then mounted it through a slot I cut into the sensor plate in the FARS unit. Now when I change out the sensors I simply pull out the old one and replace it with the new about every 18 months.  I still use the Davis filter and replace that every year.  Senserion is phasing out the SHT75 however so I think I'm going to use Jerryg's idea and buy a 31 and replace the Davis filter with the SF2 using plumbers putty to see how that works out.

Do you know what sensors the NWS will upgrade too?
Davis Vantage Pro2. SHT-75. WU: KXALJEMI2 & KALTHORS2. CWOP/APRS: C6353 & E6358

  

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 5471
You might consider just doubling the Davis filter to avoid any chemical fumes that can damage sensor. SF2  filters are very small so whatever putty or glue used will surely get into sensor while drying.

I don't really think doubling up that filter material will keep water off the sensor. That'd be like installing overlapping bug screens with 1-inch holes hoping that will keep the skeeters out. We're talking about tiny coagulates of water vapor that form when the air is near or at saturation.

I will investigate the possibility of mounting it with the clips when my caps arrive by drilling two very small holes in the board. That would be the ideal solution, but probably not possible. If not, then jerryg's solution might be the only one that's been field tested to work.

You shouldn't be getting water on sensor anyway, jerryg was because he was sucking water in with overpowered fan. Not sure if he mentioned that on this thread.

My fars is slowed down so much I need to be within 3 feet to hear with the speed controller. Still faster than solar but not by much. I no longer seal around top of AC fan reducing air intake even more. Most of this was done to reduce the humidity swings. 
Randy

Offline openvista

  • Forecaster
  • *****
  • Posts: 419
    • marquetteweather.com
You shouldn't be getting water on sensor anyway...

No, I'm not talking about sucking rain into the shield. I'm talking about some combination of very high humidity events that seem to knock the sensor out of spec. That moist air WILL enter the shield either passively or actively. Air nearing saturation becomes like a cloud where tiny water vapor droplets collide, combine and grow in size, even inside the cap. It's got nowhere to go. The SF2 rejects anything over a certain size from getting to the sensor and building up on it. Doubled up, oversized screens aren't going to prevent that situation.
Davis Vantage Pro2 FARS | https://marquetteweather.com | EW7933

Offline kcidwx

  • Forecaster
  • *****
  • Posts: 329
You shouldn't be getting water on sensor anyway...

No, I'm not talking about sucking rain into the shield. I'm talking about some combination of very high humidity events that seem to knock the sensor out of spec. That moist air WILL enter the shield either passively or actively. Air nearing saturation becomes like a cloud where tiny water vapor droplets collide, combine and grow in size, even inside the cap. It's got nowhere to go. The SF2 rejects anything over a certain size from getting to the sensor and building up on it. Doubled up, oversized screens aren't going to prevent that situation.

Correct! The only solution I'm aware of for that is the heated humidity sensor. I'm not aware of a filter that will prevent that or how it even could.

« Last Edit: July 16, 2018, 04:27:15 PM by kcidwx »
Meteorological Technician
NWS Certified Aviation Weather Observer
B.S. Meteorology

Offline jerryg

  • Forecaster
  • *****
  • Posts: 745
  • weather is never boring
    • victoria texas weather
I think a problem we have with any kind of solution is that the temp/humidity is in one unit so no way to heat humidity without raising temp. I wish the sensors were separate so the temp could get max airflow and the sensor would be located away from the air flow. Oh well i guess it is more cost effective for Davis and the others to have it that way, cost wise that is.

 

anything