Author Topic: The Reliability of the SHT-31 Humidity Sensor & What Psychrometer Should I Buy?  (Read 107170 times)

0 Members and 1 Guest are viewing this topic.

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
Finally had an extreme drop in humidity today after weeks of high humidity. Raw humidity for both my baked Davis 31 and SHT75 reached 26%. I have never seen the Davis 31 match the SHT75 at humidity levels this low. My Cumulus correction is likely excessive at this humidity level, the total correction I was using was -5%. Normally for my sensors about -5% for humidity in the mid to low 20s is about right compared to my Belfort and to match nearby ASOS stations.

To compare with ASOS, my raw data at 555p was:
Temp: 59°F Hum: 27%  DP: 25°F

KTIW at 553p:
Temp: 59°F Hum: 28% DP: 26°F

KSEA at 553p:
Temp: 61°F Hum: 24% DP: 24°F

My raw data for these sensors never compares well to these ASOS stations at this low humidity especially after long periods near saturation.

Here is the calibrated data for my station that is maybe overcorrected:
https://mesowest.utah.edu/cgi-bin/droman/meso_base_dyn.cgi?product=&past=1&stn=C7491&unit=0&time=LOCAL&day1=3&month1=01&year1=2020&hour1=18

To get 100% after the bake and now this kind of performance at the low end from a Davis 31 is really welcome news to me. I hope it continues and isn't an anomaly.
« Last Edit: January 03, 2020, 11:54:45 PM by CW7491 »

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Finally had an extreme drop in humidity today after weeks of high humidity. Raw humidity for both my baked Davis 31 and SHT75 reached 26%. I have never seen the Davis 31 match the SHT75 at humidity levels this low. My Cumulus correction is likely excessive at this humidity level, the total correction I was using was -5%. Normally for my sensors about -5% for humidity in the mid to low 20s is about right compared to my Belfort and to match nearby ASOS stations.

To compare with ASOS, my raw data at 555p was:
Temp: 59°F Hum: 27%  DP: 25°F

KTIW at 553p:
Temp: 59°F Hum: 28% DP: 26°F

KSEA at 553p:
Temp: 61°F Hum: 24% DP: 24°F

My raw data for these sensors never compares well to these ASOS stations at this low humidity especially after long periods near saturation.

Here is the calibrated data for my station that is maybe overcorrected:
https://mesowest.utah.edu/cgi-bin/droman/meso_base_dyn.cgi?product=&past=1&stn=C7491&unit=0&time=LOCAL&day1=3&month1=01&year1=2020&hour1=18

To get 100% after the bake and now this kind of performance at the low end from a Davis 31 is really welcome news to me. I hope it continues and isn't an anomaly.

The 31 you’re talking here is the one with the clear potting and not the black one right?
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
The 31 you’re talking here is the one with the clear potting and not the black one right?

That's correct. Another Davis 31 came yesterday. It also has clear potting. Manufacture date also looks like November 2019. I baked it last night wrapped in aluminum foil with just the actual sensor exposed. Worked great. No melting. I will post the pre and post bake results later. I've actually put the other clear baked Davis 31 as my main sensor now. It has a -2% offset and then a Cumulus calibration of (1.021x-2). This produces a -2% correction at the top of the scale and max -4% correction at the lower end of the scale. I may take the Cumulus calibration out completely as I see how it performs. The SHT75 I was using as my main sensor is still outside for comparison. My CWOP ID is CW7491 of course for those that want to see how this baked Davis 31 does

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
Nothing surprising here, but passing along anyway. Here are the pre bake and post bake comparisons for the latest Davis clear potting SHT31:

Pre-bake:
Davis 31
Temp: 73.6°F  Hum: 47%  DP: 52°F

Belfort Psychrometer
Temp: 73.4°F  Hum: 42%  DP: 49°F

Post Bake:
Davis 31
Temp: 67.4°F  Hum: 42%  DP: 44°F

Belfort Psychrometer
Temp: 67.4°F  Hum: 43%  DP: 44°F



Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
Good results I just wonder how long it will last?
Randy

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
Good results I just wonder how long it will last?
That is the question isn't it? I'm cautiously optimistic based on yesterday's humidity drop. The Davis 31 I am using now has been baked and outside in near saturated air for 2 weeks. I've never seen the raw data match the SHT75 or ASOS even after previous bakes, but I've been disappointed with these things before ... I would not be surprised to see it no longer hit 100% and/or for the high humidity bias to return.

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
The 31 you’re talking here is the one with the clear potting and not the black one right?

That's correct. Another Davis 31 came yesterday. It also has clear potting. Manufacture date also looks like November 2019. I baked it last night wrapped in aluminum foil with just the actual sensor exposed. Worked great. No melting. I will post the pre and post bake results later. I've actually put the other clear baked Davis 31 as my main sensor now. It has a -2% offset and then a Cumulus calibration of (1.021x-2). This produces a -2% correction at the top of the scale and max -4% correction at the lower end of the scale. I may take the Cumulus calibration out completely as I see how it performs. The SHT75 I was using as my main sensor is still outside for comparison. My CWOP ID is CW7491 of course for those that want to see how this baked Davis 31 does

I baked one with the black potting. So far, it’s performing well. But I have no filter over it and the sensor is receiving good airflow in a passive shield.
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6731
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
The 31 you’re talking here is the one with the clear potting and not the black one right?

That's correct. Another Davis 31 came yesterday. It also has clear potting. Manufacture date also looks like November 2019. I baked it last night wrapped in aluminum foil with just the actual sensor exposed. Worked great. No melting. I will post the pre and post bake results later. I've actually put the other clear baked Davis 31 as my main sensor now. It has a -2% offset and then a Cumulus calibration of (1.021x-2). This produces a -2% correction at the top of the scale and max -4% correction at the lower end of the scale. I may take the Cumulus calibration out completely as I see how it performs. The SHT75 I was using as my main sensor is still outside for comparison. My CWOP ID is CW7491 of course for those that want to see how this baked Davis 31 does
I have no filter over it and the sensor is receiving good airflow in a passive shield.
Even without a fan, I'd think particulates will eventually get into that tiny sensor hole without some kind of filter.

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
The 31 you’re talking here is the one with the clear potting and not the black one right?

That's correct. Another Davis 31 came yesterday. It also has clear potting. Manufacture date also looks like November 2019. I baked it last night wrapped in aluminum foil with just the actual sensor exposed. Worked great. No melting. I will post the pre and post bake results later. I've actually put the other clear baked Davis 31 as my main sensor now. It has a -2% offset and then a Cumulus calibration of (1.021x-2). This produces a -2% correction at the top of the scale and max -4% correction at the lower end of the scale. I may take the Cumulus calibration out completely as I see how it performs. The SHT75 I was using as my main sensor is still outside for comparison. My CWOP ID is CW7491 of course for those that want to see how this baked Davis 31 does
I have no filter over it and the sensor is receiving good airflow in a passive shield.
Even without a fan, I'd think particulates will eventually get into that tiny sensor hole without some kind of filter.

True. I “melted” my filter some when I baked the sensor. Should’ve taken the filter cap off. Lol. But I’m going to see how long it lasts without a filter.
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
So far, I’m still liking the results of the baked sensor. Anybody else still liking their results?
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
So far, I’m still liking the results of the baked sensor. Anybody else still liking their results?

Same here. My baked clear potting Davis 31 is in an AC aspirated shield. Baked SHT75 is in a passive 7714 now. They're side by side still though. I've noticed that the Davis 31 is now running a little lower as humidity drops, which is a first. At the high end, it also reads a little lower. Haven't had real heavy saturation conditions since I have had it as my primary sensor, which has been less than a week. We'll see if this drier performance will prohibit it from getting to 100%. Since it has been my primary sensor, max RH has been 98%. I'll update when we get conditions that are full saturation.

Offline fkapp

  • Forecaster
  • *****
  • Posts: 333
Hi,
Just wanted to update that still am very happy with the SHT with the clear coat that came with my VP2 that installed last week of 2019.

This is in passive 7714 shield, will get FARS going before summer, but results are very close to nearby NWS stations.
I have not gotten int baking the sensor (to be honest a bit concerned on doing this first time) but this has the SF2 Cap mod right out of box before going into outside air.

CWOP analysis has been completed and figured would post link to the results here in case anyone was curious how this compares
https://weather.gladstonefamily.net/site/F6381

Thanks for all the tips
Frank

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Anyone still having good results with the baked sensor?
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
Anyone still having good results with the baked sensor?

I am still having great results. My clear potting baked Davis 31 actually reads about 2-3% drier side by side with a baked SHT75 in the mid to upper humidity range. I have still yet to see really heavy saturation, but the 31 has gotten to 99%. The SHT75 has hit 100% multiple times. I’m still waiting to get a heavy fog to see if the 31 will hit 100%.

I have also noticed fkapp is getting great results compared to KBUF ASOS without a bake. What is the max humidity you’ve seen Frank?

As I posted earlier in the thread, I saw an improvement in the performance at room temperature post bake, but I am not running an unbaked clear Davis 31 for comparison, so I can’t say for certain how much this improved performance is attributed to the bake and how much to a potential improvement by Davis. I’ve seen enough compared to the SHT75, so I may try a baked Davis 31 with black potting for comparison.

Offline fkapp

  • Forecaster
  • *****
  • Posts: 333
Hi guys,
Max humidity so far has been 96% back on Jan 12th.

I just checked against the NWS Buffalo daily climate summary and shows the Max as the NWS BUF was 96% on this day as well.

Overall still very happy

Thanks
Frank

Offline jerryg

  • Forecaster
  • *****
  • Posts: 886
  • weather is never boring
    • victoria texas weather
Well i haven't updated in a while because the new 31 that i was testing went belly up after only two months and i was in mourning for some time because i took a chance and did a little reshaping of the board to fit my passive shield and thus no warranty. I just got a new 31 in yesterday and after wrapping it in a wet wash cloth for some time it seems to max out at only 96%. I put it outside last night so i could compare it with three other sensors i am testing and this morning the 3 sensors showed one at 92% and the other two at 91% while the new 31 was showing 86%, just lousy reading. I am now baking it to see if it will get back on track with a good reading. Will report tomorrow if it is going back to Ryan for another one.




Offline CW7491

  • Senior Contributor
  • ****
  • Posts: 202
Well i haven't updated in a while because the new 31 that i was testing went belly up after only two months and i was in mourning for some time because i took a chance and did a little reshaping of the board to fit my passive shield and thus no warranty. I just got a new 31 in yesterday and after wrapping it in a wet wash cloth for some time it seems to max out at only 96%. I put it outside last night so i could compare it with three other sensors i am testing and this morning the 3 sensors showed one at 92% and the other two at 91% while the new 31 was showing 86%, just lousy reading. I am now baking it to see if it will get back on track with a good reading. Will report tomorrow if it is going back to Ryan for another one.

Did you happen to notice the color of the potting and/or the presumed date on the sticker on the back?

Offline jerryg

  • Forecaster
  • *****
  • Posts: 886
  • weather is never boring
    • victoria texas weather
I did notice the potting stuff was clear and seems to cover less of the board than the heavy black stuff. I didn't pay any attention to the sticker.

Offline jerryg

  • Forecaster
  • *****
  • Posts: 886
  • weather is never boring
    • victoria texas weather
Ok finished the 10 hour baking and testing of the sensor and so far all is great. The max reading so far is 99% and i just finished installing it as my main sensor and it is reading the same humidity and temp. as the two other sensors co-located  on the same mount.  One is in a passive shield and the other in a Davis fars shield. The response time is really very good, went from 42% to 99% in less than an hour under test. Most sensors i have tested top out at 98% after an hour and a half to two hours. Now if it will just keep working for a long time i will be a happy camper. Next check is for any kind of wet bias to show up. =D>

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Keep in mind that these two stations are around .08 miles apart from each other. One station has the baked 31 and the other one has a 15 inside of a Barani shield. Both recorded a high of 73° and the 15 read 2% lower than the 31 and that correlated with a 1° variance in DP
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Not sure but I think my 31 has succumbed back to having a wet bias. But then again, it’s been raining a lot and when it’s dry, the sensor would only have one day to recalibrate itself.
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline andro700

  • Chuck
  • Forecaster
  • *****
  • Posts: 420
    • Gobles Weather Page
Not sure if this has been asked. When you bake your sensor, do you put the wires and all in oven or take them off.

Chuck

Offline galfert

  • Global Moderator
  • Forecaster
  • *****
  • Posts: 6822
Not sure if this has been asked. When you bake your sensor, do you put the wires and all in oven or take them off.

Chuck

Yup, the whole thing. https://www.wxforum.net/index.php?topic=34658.msg394802#msg394802
« Last Edit: February 12, 2020, 07:41:20 PM by galfert »
Ecowitt GW1000 | Meteobridge on Raspberry Pi
WU: KFLWINTE111  |  PWSweather: KFLWINTE111
CWOP: FW3708  |  AWEKAS: 14814
Windy: pws-f075acbe
Weather Underground Issue Tracking
Tele-Pole

Offline andro700

  • Chuck
  • Forecaster
  • *****
  • Posts: 420
    • Gobles Weather Page
I believe I have an SHT15. I bought my station around 2013 or 2014 I believe. Would I be better off getting an SHT31? I also live on the western Washington. A lot of humidity here.

Chuck

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6731
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
I believe I have an SHT15. I bought my station around 2013 or 2014 I believe. Would I be better off getting an SHT31?
If you bought it new and have not modified it, you have an SHT11. The 31 is a superior sensor, problem aside, which has "seemingly" been addressed. Personally, I would spring for a 31. If you do, don't forget the -0.9F offset for the temperature.