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

[dendrite]

here is a sample of reports from victoria and port lavaca metars and mine at 1 pm victoria temp 93 dp 71 h 48  port lavaca temp 92 dp 65 h 41 and mine temp 91 dp 72 h 54 That is a lot of variations to try to compare to my dp was pretty close to victoria but way off from port lavaca just too many variations to try and compare being i am half way in between the two. I just have to trust that my 29 dollar sensor is doing as good as it can without some type of calibrated instruments to compare to on site. Heck for my site i am right and they are wrong 

The 65 at PKV was very brief.

https://mesowest.utah.edu/cgi-bin/droman/meso_base_dyn.cgi?stn=KPKV&unit=0&timetype=LOCAL

If you are running 1-3F higher on dewpoint you basically just have to decide if you are happy with it. Either you think the higher values physically make sense (foliage vs a rocky/dirt ASOS) or you feel there’s an instrumentational error there that you can live with. I think the psychrometer is really the only way to tell. I would measure it at your place, take a drive to one of the ASOS stations at the top of an hour and measure there, and then head home and do another measurement. I run a little higher than ASOS too, but I’d say the majority of the difference meteorologically makes sense given my lighter winds and moist ground. Some of it probably is simply sensor error and I’m OK with that as long as my temperature readings are on par. My old SHT31 was only peaking at 93% this summer. The new SHT75 has already hit 97% in 2 nights. So the old 31 definitely saw some drift.

[dendrite]

Current dewpoints

Jerry 1944z 67.8°
PKV 1940z 66.9°
VCT 1940z 68°

[jgentry]

Jerry’s 75 has a lower DP than the RW station.

[openvista]

Current dewpoints

Jerry 1944z 67.8°
PKV 1940z 66.9°
VCT 1940z 68°

You can't just take one instantaneous reading from CWOP and compare it to airports several minutes ago. When I checked earlier, it was stagnant & sunny and his DPs were bouncing around by several degrees even within a matter of seconds. He has an SHT75 loaded into a FARS. It's like a caged monkey! You have to average several of his readings. That means sitting on his website for several minutes, recording at least a couple minutes of DP readings and then deriving an average.

Of course, comparing to reference stations is only a possible indication of a problem, not confirmation. That would require a calibrated device.

[dendrite]

Current dewpoints

Jerry 1944z 67.8°
PKV 1940z 66.9°
VCT 1940z 68°

You can't just take one instantaneous reading from CWOP and compare it to airports several minutes ago. When I checked earlier, it was stagnant & sunny and his DPs were bouncing around by several degrees even within a matter of seconds. He has an SHT75 loaded into a FARS. It's like a caged monkey! You have to average several of his readings. That means sitting on his website for several minutes, recording at least a couple minutes of DP readings and then deriving an average.

Of course, comparing to reference stations is only a possible indication of a problem, not confirmation. That would require a calibrated device.

I know. I made that same point earlier...

https://www.wxforum.net/index.php?topic=34658.msg355192#msg355192

What I was getting at is that they can all be different one hour and then lined up the next. Sorry, for not clarifying what I was getting at by posting the dewpoints. 

[openvista]

Current dewpoints

Jerry 1944z 67.8°
PKV 1940z 66.9°
VCT 1940z 68°

You can't just take one instantaneous reading from CWOP and compare it to airports several minutes ago. When I checked earlier, it was stagnant & sunny and his DPs were bouncing around by several degrees even within a matter of seconds. He has an SHT75 loaded into a FARS. It's like a caged monkey! You have to average several of his readings. That means sitting on his website for several minutes, recording at least a couple minutes of DP readings and then deriving an average.

Of course, comparing to reference stations is only a possible indication of a problem, not confirmation. That would require a calibrated device.

I know. I made that same point earlier...

https://www.wxforum.net/index.php?topic=34658.msg355192#msg355192

What I was getting at is that they can all be different one hour and then lined up the next. Sorry, for not clarifying what I was getting at by posting the dewpoints. 

Ah, somehow I missed that post. Yes, that's a great idea. Get an average of dewpoints for an afternoon from one or more reference station(s) and then an average of DPs of the station in question. I could see that, due to sensor or micro-scale weather differences, one hour's readings could align and another's be significantly different. The afternoon average better tells the story.

[jgentry]

As of 4:47p Jerry is reporting 96/59-60 and the RW station is reporting 96/65

[ValentineWeather]

You can do a station to station compare.

https://weather.gladstonefamily.net/qchart/AT358?date=20180801&addnl=KVCT&Add+to+charts=Add+to+charts&.cgifields=addnl


At the bottom is nearby sites select what you want and add to charts.

[Bobvelle]

He has an SHT75 loaded into a FARS. It's like a caged monkey!

And I have a SHT11 loaded in passive shield in the shade... It's like a Three-toed Sloth on Xanax.

[Bobvelle]

Well as of an hour ago. I have a new set-up. Yeah I got rid of the SHT11 after a couple days of data.

St.#1 --SHT31  //   24 hr FARS  //  mostly in sun  // latest ver. ISS running raw (So no -0.9 deg offset needed)

St.#2 --SHT31 // Passive shield // mostly shaded // older ver. ISS running with the suggested -0.9 deg offset.

 Also St.# 2 is an older SHT31 that I baked and re-hydrated Monday.

I will record and report what I see. Several different variables going here, but it should still be interesting data.

[jgentry]

here is a sample of reports from victoria and port lavaca metars and mine at 1 pm victoria temp 93 dp 71 h 48  port lavaca temp 92 dp 65 h 41 and mine temp 91 dp 72 h 54 That is a lot of variations to try to compare to my dp was pretty close to victoria but way off from port lavaca just too many variations to try and compare being i am half way in between the two. I just have to trust that my 29 dollar sensor is doing as good as it can without some type of calibrated instruments to compare to on site. Heck for my site i am right and they are wrong 

How did the 31 and 15 compared against the 75?   Your 75 was reporting DPs lower than the metars and the RW station at times during this afternoon

[galfert]

In reference to the SF2 mounting issues...

Why not use Closedcube Breakout SHT31 solution that includes SHT31 in small PCB with mounting holes for the SF2 filter, and includes the SF2 filter. Then all you need to do is solder the 4 leads.

Closedcube makes a very small SHT31 breakout and they have a bit larger one they call PRO. I'm not sure what the difference is. Also there is a D (digital) version and an A (Analog) version. Has anyone determined if Davis uses the Analog or the Digital?

[Bobvelle]

In reference to the SF2 mounting issues...

Why not use Closedcube Breakout SHT31 solution that includes SHT31 in small PCB with mounting holes for the SF2 filter, and includes the SF2 filter. Then all you need to do is solder the 4 leads.

Closedcube makes a very small SHT31 breakout and they have a bit larger one they call PRO. I'm not sure what the difference is. Also their is a D (digital) version and an A (Analog) version. Has anyone determined if Davis uses the Analog or the Digital?

Then you have the issue of finding a way to mount it to the FARS mounting plate or the VP2 Passive shield mount. So either way there is some customizing going on .. not to mention, now you gotta pull out the soldering iron...and then encapsulation/epoxy/silicone.
However, this may be the perfect solution for the Meteoshield, and for that matter, the 7714 since there is no specific mount for the Davis sensor in these to begin with.
I am too Interested in what version (digital or Analog) works with the VP2 protocol.

[jerryg]

You need to keep in mind that the break out is i2c formatted and will not work with Davis iss.

[galfert]

You need to keep in mind that the break out is i2c formatted and will not work with Davis iss.

Maybe the Digital version is the i2c and the Analog is the one for Davis.

[dendrite]

You need to keep in mind that the break out is i2c formatted and will not work with Davis iss.

Maybe the Digital version is the i2c and the Analog is the one for Davis.

The Davis uses the SHT31-LSS

[jas340]

FWIW my Davis VP2 is just 2 months old. It has been running about 5 degrees wet since day 1. My Vue also ran wet. Until this thread I though I was the only one. I almost got rid of my Vaisala WXT because it did not match the humidity of the VP2!!!
It's good to know I'm not alone however I feel kind of taken. For years everyone has raved about Davis quality.

The rain gauge is no where near accurate if the wind is blowing. The anemometer not reliable under 2mph. Both are these are just nature of the beast due to design limitations and are really not a problem.

There is no excuse for the SHT-31 issue. It is a $3.50 sensor. How about spending $7.00 wholesale on a much better sensor?????

 

Edit: I made a total noob mistake and disparaged the Davis brand even though my stations humidity is spot on with 2 reference stations. I am using my old Vue console with the new VP2 and decided to input a -4 degree calibration offset for humidity. Low and behold I had a +4 offset already in there. Zeroed it out and now I am spot on. I don't remember ever putting an offset in and I've had the console 7 years or so. In conclusion, my bad. I do live in a semi-arid climate etc. but am glad the sensor is accurate when new (in my case).

[jerryg]

I got my Barani pro passive shield in a while ago and i got it up and running for test and so far it is working great in a short test time. The wind has been calm to light so far and with it calm the pro has been running .5f lower than the fars and when the wind has been up around 4 mph or more it has been running about 1 degree lower. I sure hope this proves out in the long term i will use it and get rid of the fan problem and maybe help the sensor with the high humidity problem.

[jgentry]

I got my Barani pro passive shield in a while ago and i got it up and running for test and so far it is working great in a short test time. The wind has been calm to light so far and with it calm the pro has been running .5f lower than the fars and when the wind has been up around 4 mph or more it has been running about 1 degree lower. I sure hope this proves out in the long term i will use it and get rid of the fan problem and maybe help the sensor with the high humidity problem.

Please keep us up to date.

[galfert]

You need to keep in mind that the break out is i2c formatted and will not work with Davis iss.

Maybe the Digital version is the i2c and the Analog is the one for Davis.

The Davis uses the SHT31-LSS

So does that mean LSS version is not the Analog version? I read that LSS is a legacy version. So are three 3 versions of the SH31 or just 2?

[jerryg]

I just moved the pro over to the main station, did not want to hook it up until i could see that it was going to work as advertised. Now is am going to keep an eye on it and my other fars that runs even in readings with it. When i reversed the sensor cable the results where still the same so it has nothing to do with one fars verses the other. As long as the pro shield runs no worse in temp than the fars i will be happy. Just the thought of not having to mess with a powered shield makes me giddy lol.Right now i am keeping an eye on the humidity readings too. Tonight will really be interesting when the humidity starts to go up and how the fars and the pro compare.

[jerryg]

There are 3 version of the sensor one is analog the other is digital with i2c format and the third is what Davis uses the LSS chip which is also digital just a different format.

[CW2274]

Well, as much as I hate to say this, looks like I'll be jumping on the "bash the 31" wet bias bandwagon. After weeks of scrutinizing even more than usual, I've come to the conclusion that I'm suffering the same symptoms, maybe just not quite as bad as everyone else. As a generality, I'm running 3-4F in dew higher than my "official buddies". The thing is, neighboring stations tend to run wet also, so makes me dubious, but yes, most are Davis's. My 11 and 15 behaved the same way, which also makes me wonder. I guess I'll dial back the humidity 2% and see how that looks for now. Crap...

[ValentineWeather]

Well, as much as I hate to say this, looks like I'll be jumping on the "bash the 31" wet bias bandwagon. After weeks of scrutinizing even more than usual, I've come to the conclusion that I'm suffering the same symptoms, maybe just not quite as bad as everyone else. As a generality, I'm running 3-4F in dew higher than my "official buddies". The thing is, neighboring stations tend to run wet also, so makes me dubious, but yes, most are Davis's. My 11 and 15 behaved the same way, which also makes me wonder. I guess I'll dial back the humidity 2% and see how that looks for now. Crap...

Welcome aboard and about time.
I can confirm Davis has made shipping changes but I'm skeptical that's the issue. Haven't figured out yet because the Acurite SHT31 version doesn't display the same high bias.  I'm also testing on my NEW just built 2 days ago outdoor shaded porch and it's running close to 4-5% lower humidity all the time except in the morning dew when they all come together around 97% today.  It's not the LCC version however and isn't embedded into plastic like the Davis either.

[CW2274]

Well, as much as I hate to say this, looks like I'll be jumping on the "bash the 31" wet bias bandwagon. After weeks of scrutinizing even more than usual, I've come to the conclusion that I'm suffering the same symptoms, maybe just not quite as bad as everyone else. As a generality, I'm running 3-4F in dew higher than my "official buddies". The thing is, neighboring stations tend to run wet also, so makes me dubious, but yes, most are Davis's. My 11 and 15 behaved the same way, which also makes me wonder. I guess I'll dial back the humidity 2% and see how that looks for now. Crap...

Welcome aboard and about time.

Didn't want to proclaim the sky was falling until I truly gave my particular situation a thorough evaluation. Perhaps being in a drier climate does help mitigate the bias, but from my perspective, it's there. Taking out 2% has me perfectly in-line.... for now.....