Author Topic: The Reliability of the SHT-31 Humidity Sensor & What Psychrometer Should I Buy?  (Read 107454 times)

0 Members and 1 Guest are viewing this topic.

Offline drew1021

  • Senior Contributor
  • ****
  • Posts: 229
Got 2 new SHT31's in handled by a different company lets see if they do any better. I just happen to have passive shield I may try. Getting data could be an issue however they do record high and lows however.
 
Good point was brought up why rainwise sensors seem better using the SHT15 on mid to high dp temps in 60-80 range.

Something was brought to my attention by someone not on this thread the molding process Davis uses to mount sensor could be contamination source dooming sensor from getgo. Maybe that's why the reckless handling.

The sensor looks embedded in rubber or plastic which would be against all handling procedures. 

 [ You are not allowed to view attachments ]

Or perhaps RW manipulates the raw readings with software??
VP2 with 24 hour FARS. WU: KNCLEWIS2. CWOP/APRS: DW4712, COCORAHS: NC-FR-7

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6731
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Got 2 new SHT31's in handled by a different company lets see if they do any better. I just happen to have passive shield I may try. Getting data could be an issue however they do record high and lows however.
 
Good point was brought up why rainwise sensors seem better using the SHT15 on mid to high dp temps in 60-80 range.

Something was brought to my attention by someone not on this thread the molding process Davis uses to mount sensor could be contamination source dooming sensor from getgo. Maybe that's why the reckless handling.

The sensor looks embedded in rubber or plastic which would be against all handling procedures. 

 [ You are not allowed to view attachments ]

Or perhaps RW manipulates the raw readings with software??
Doubtful.
The RW shield is passive, and as has been discussed, perhaps a difference maker.

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
Okay 15 minutes in shade.
Snapshot of station at same time I took photo.  Still watching running 5% higher on Davis sensor.
 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]

Randy

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6731
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
But less than a 2F in the dew. Not perfect.......
http://www.dpcalc.org/

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
Yes thought specs were +/- 2 % humidity :?:  Test wasn't really great because with clouds now the station has cooled below patio +3° on temperature so test was only 20 minutes during that period averaged 5% difference.
Randy

Offline CW2274

  • Forecaster
  • *****
  • Posts: 6731
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Rainwise sensors ain't gonna be cheap. Look at these "Government" (GSA) prices!!!

https://www.gsaadvantage.gov/ref_text/GS07F5342R/GS07F5342R_online.htm
At least the feds got a 29.25% discount. What's there to whine about? :roll:


Offline CW2274

  • Forecaster
  • *****
  • Posts: 6731
    • Conditions @ CW2274 West Tucson-Painted Hills Ranch
Rainwise sensors ain't gonna be cheap. Look at these "Government" (GSA) prices!!!

https://www.gsaadvantage.gov/ref_text/GS07F5342R/GS07F5342R_online.htm
At least the feds got a 29.25% discount. What's there to whine about? :roll:
We ain't the Feds (nor the Federale's either).  :roll:
Even with the discount, the prices are ludicrous. 

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
Longer test this evening with the 2-Acurite SHT31's sitting on top of Davis Shield. Took snapshot of Davis running 8% higher humidity. Not exactly scientific because Davis is aspirated.

So this is 2 brand new SHT31's different company and a 5 day old Davis sensor VP2 aspirated.
 

Now down to 71° as I type still running 8% high. 74% vs 83% on Davis vs 2 different Acurite's .

Maybe this is why the Rainwise sensors look normal, non aspirated and probably don't destroy sensor mounting on board.


Randy

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Randy, my newest SHT-31 sensor is housed in a passive shield and it basically performs the same as if it’s in the FARS. Could just be the board itself. If David used a probe-type sensor, we might not have the problems we are currently dealing with.  Who knows...
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
Acurite don't use probes these are the real Mccoys.

 [ You are not allowed to view attachments ]

Randy

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
This morning with humidity up the new acurite with sht31's both peaked at 95% same as the new Davis aspirated. The 2 year old Davis peaked at 93% while the handheld Extech 96% this morning just to complete how these sensors compare.

They do appear to run substantially lower 8% daytime and much closer with airport ASOS yesterday afternoon and evening, along with reaching the upper end of humidity scale 95%, airport also was 94% today. This makes everything really frustrating knowing the 31's can perform just as data sheet says but the Davis Units don't.

Why the Davis sensors run so high I can only conjecture molding into board process contamination.

This is very aggravating after seeing the Acurite $15 sensor work so much better. This probably goes to why the rainwise sensors are much closer and don't seem to suffer like the Davis sensors. They don't destroy sensor before consumer gets them.

I'll let you guys continue but I've seen enough to know the Davis units are junk.

Randy

Offline jgentry

  • Forecaster
  • *****
  • Posts: 533
Well that ticks me off too. Either David just doesn’t realize the their board is screwing up the humidity sensor, or they do but don’t care.  Thank you for doing these tests. I think we now got into the culprit of our problems.
Davis Vantage Pro2 & WeatherFlow Tempest. WU: KXALJEMI2, KALJEMIS7, KFLPANAM363 & KALTHORS2. CWOP/APRS: C6353 & E6358

Offline Dj1225

  • Senior Member
  • **
  • Posts: 67
This morning with humidity up the new acurite with sht31's both peaked at 95% same as the new Davis aspirated. The 2 year old Davis peaked at 93% while the handheld Extech 96% this morning just to complete how these sensors compare.

They do appear to run substantially lower 8% daytime and much closer with airport ASOS yesterday afternoon and evening, along with reaching the upper end of humidity scale 95%, airport also was 94% today. This makes everything really frustrating knowing the 31's can perform just as data sheet says but the Davis Units don't.

Why the Davis sensors run so high I can only conjecture molding into board process contamination.

This is very aggravating after seeing the Acurite $15 sensor work so much better. This probably goes to why the rainwise sensors are much closer and don't seem to suffer like the Davis sensors. They don't destroy sensor before consumer gets them.

I'll let you guys continue but I've seen enough to know the Davis units are junk.

Randy would you consider switching back to the SHT15?

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
This morning with humidity up the new acurite with sht31's both peaked at 95% same as the new Davis aspirated. The 2 year old Davis peaked at 93% while the handheld Extech 96% this morning just to complete how these sensors compare.

They do appear to run substantially lower 8% daytime and much closer with airport ASOS yesterday afternoon and evening, along with reaching the upper end of humidity scale 95%, airport also was 94% today. This makes everything really frustrating knowing the 31's can perform just as data sheet says but the Davis Units don't.

Why the Davis sensors run so high I can only conjecture molding into board process contamination.

This is very aggravating after seeing the Acurite $15 sensor work so much better. This probably goes to why the rainwise sensors are much closer and don't seem to suffer like the Davis sensors. They don't destroy sensor before consumer gets them.

I'll let you guys continue but I've seen enough to know the Davis units are junk.

Randy would you consider switching back to the SHT15?

No the high humidity started about when the 15 came out, plus 15 temperature accuracy suffers below 14°f and I live way below that in winter and would rather have accurate temperature over humidity.
Randy

Offline dport

  • Senior Contributor
  • ****
  • Posts: 203
This is highly annoying as I just bought a VP2 with 24/hr FARS and the WifiLogger last month.  That's a substantial investment for me.  I was under the impression that they SHT31 was one of the best prosumer sensors out there.  My humidity has hit 99% since I have installed, but I have also noticed that humidity reads high more often than not.

So, how can we engage Davis at a higher level?  Other options available to us?

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
I still believe the 31 is a great sensor just not after Davis gets done with them.
As a consumer and spending oodles of money on Davis products I'll now look at other options. Meanwhile I'm hoping some smart electronic people will come to the rescue with a new transmitter so we can install our own sensors not pre trashed before we get them.

 
Randy

Offline ValentineWeather

  • Forecaster
  • *****
  • Posts: 6364
    • Valentine Nebraska's Real-Time Weather
dport look at Cumulus also I run the desktop version you can run some of that high bias out with the humidity multiplier and still reach 100%. It's kind of a pain however once you get below the 35% range the high bias is gone so you need to keep a watch all the time. Frustrating.. ](*,) Other softwares also may have this ability.
 
Randy

Offline dendrite

  • Forecaster
  • *****
  • Posts: 367
This morning with humidity up the new acurite with sht31's both peaked at 95% same as the new Davis aspirated. The 2 year old Davis peaked at 93% while the handheld Extech 96% this morning just to complete how these sensors compare.

They do appear to run substantially lower 8% daytime and much closer with airport ASOS yesterday afternoon and evening, along with reaching the upper end of humidity scale 95%, airport also was 94% today. This makes everything really frustrating knowing the 31's can perform just as data sheet says but the Davis Units don't.

Why the Davis sensors run so high I can only conjecture molding into board process contamination.

This is very aggravating after seeing the Acurite $15 sensor work so much better. This probably goes to why the rainwise sensors are much closer and don't seem to suffer like the Davis sensors. They don't destroy sensor before consumer gets them.

I'll let you guys continue but I've seen enough to know the Davis units are junk.
Try a 75...tiny thermal mass, preinstalled to a board by Sensirion, and plug and play after a little soldering or 1.27mm 4-pin connector.

As for your evening experiment. I would expect the Davis to run higher RH with fan aspiration and dew beginning to form at the surface. If you're seeing it in the afternoon as well then maybe it's a different issue. I don't like comparing my humidity to ASOS since I'm in an area with lots of trees and transpiration. The airports have wide open areas with great mixing which cuts down on those pooling dewpoints from the vegetation. You're in NE so maybe this doesn't apply since you may have mostly plains and fields too. But out here, it's lots of hills and trees.

I feel like we've been discussing this for years now. Personally, 98% vs 100% isn't a big deal to me. Your local ASOS miles may vary, but most of the ones out here never reach 100% when viewing the 1-min data...even with M1/4SM or 1/4SM visibility. I just logged in yesterday for the first time in awhile so maybe you've addressed this, but have you tried a NIST certified sensor?

Offline johnd

  • Forecaster
  • *****
  • Posts: 4827
    • www.weatherstations.co.uk
I just logged in yesterday for the first time in awhile so maybe you've addressed this, but have you tried a NIST certified sensor?

I hesitate to intrude again on this long-running thread, but isn't the issue here more one of drift from the initial specification over a period of time, whether shorter or longer?

All a NIST sensor from Davis is going to tell you is that when brand new (unless it's a subsequent retest) the RH is confirmed as within ±3% of the reading of some NIST-traceable reference instrument (apparently a General Eastern M4-RH Dew Point Monitor but presumably with some intrinsic error of its own) at 33%; 80% and 90% RH only. Does that really help a lot?

More details in attached PDF. This is the document currently downloadable from Davis but I can't see its original date. Conceivably the tests specs have been tightened since the introduction of SHT31 but there's nothing to indicate that.
Prodata Weather Systems
Prodata's FAQ/support site for Davis stations
Includes many details on 6313 Weatherlink console.
UK Davis Premier Dealer - All Davis stations, accessories and spares
Cambridge UK

Sorry, but I don't usually have time to help with individual issues by email unless you are a Prodata customer. Please post your issue in the relevant forum section here & I will comment there if I have anything useful to add.

Offline dendrite

  • Forecaster
  • *****
  • Posts: 367
I just logged in yesterday for the first time in awhile so maybe you've addressed this, but have you tried a NIST certified sensor?

I hesitate to intrude again on this long-running thread, but isn't the issue here more one of drift from the initial specification over a period of time, whether shorter or longer?

All a NIST sensor from Davis is going to tell you is that when brand new (unless it's a subsequent retest) the RH is confirmed as within ±3% of the reading of some NIST-traceable reference instrument (apparently a General Eastern M4-RH Dew Point Monitor but presumably with some intrinsic error of its own) at 33%; 80% and 90% RH only. Does that really help a lot?

More details in attached PDF. This is the document currently downloadable from Davis but I can't see its original date. Conceivably the tests specs have been tightened since the introduction of SHT31 but there's nothing to indicate that.
I'm pretty sure in his recent posts he was doing experiments with 2 newer acurite sensors and a 5 day old SHT31.

Offline johnd

  • Forecaster
  • *****
  • Posts: 4827
    • www.weatherstations.co.uk
I'm pretty sure in his recent posts he was doing experiments with 2 newer acurite sensors and a 5 day old SHT31.

That would be the shorter period then.  :grin: But obviously there may be issues over short-term and then long-term drift, though that's a separate issue. The main point I was aiming at was that NIST certification probably doesn't add much over a random unchecked SHT31, other than to rule out one that was way out of spec, which really shouldn't happen.
Prodata Weather Systems
Prodata's FAQ/support site for Davis stations
Includes many details on 6313 Weatherlink console.
UK Davis Premier Dealer - All Davis stations, accessories and spares
Cambridge UK

Sorry, but I don't usually have time to help with individual issues by email unless you are a Prodata customer. Please post your issue in the relevant forum section here & I will comment there if I have anything useful to add.

Offline openvista

  • Forecaster
  • *****
  • Posts: 481
    • marquetteweather.com
I'm pretty sure in his recent posts he was doing experiments with 2 newer acurite sensors and a 5 day old SHT31.

That would be the shorter period then.  :grin: But obviously there may be issues over short-term and then long-term drift, though that's a separate issue. The main point I was aiming at was that NIST certification probably doesn't add much over a random unchecked SHT31, other than to rule out one that was way out of spec, which really shouldn't happen.

The thread keeps evolving. Yes, some of us are experiencing drift over a long period. Others of us are experiencing bad sensors from the start (Randy got two bad ones in row recently). All of us reporting issues have wet dew points/humidities in common.

As for NIST certification, apparently now Davis is claiming in an email conversation their certified sensors are good to +/- 1% at 33% and 80% and +/- 2% at 90%. This was mentioned a couple pages back by jerryg. But I would again point out that they don't mention what temperature(s) they test. This is relative humidity after all. Also, there is no checkpoint in the middle range (40-70%) which is where most of the problems lie. So, presumably, the sensor could pass all those tests and someone could hook it up and see that at a temp/dew point of 90F/70F humidity is running high by 5% or more.
« Last Edit: July 19, 2018, 12:19:26 PM by openvista »
Davis Vantage Pro2 AC FARS | https://marquetteweather.com | EW7933

Offline WheatonRon

  • Forecaster
  • *****
  • Posts: 1237
    • WUnderground
What is most irritating is I, and many other folks posting in this Forum, thought by purchasing Davis equipment we were buying the best available in the consumer market. Now I have second thoughts and I am on my third VP2, the latest being the 24hour FARS I bought from Ryan about 2 months ago. The latest appears to be better rainwise than the prior two I have owned but I haven’t had a huge downpour yet to see if the tipper does better than the prior two. However, the humidity and temperature readings have been inconsistent over the past 2 months—sometimes high other times low—when compared to what I believe are good comparables nearby. I informed Ryan that I won’t be asking for a replacement of the SHT31 yet, as I am not convinced I will be gaining anything. As I posted earlier, I thought my 2 year old VP2 was reading humidity high when compared to my new one but the next day—the results flipped and continue to be variable—when compared with my other VP2s on the same fence 10-50 feet apart. And as others have noted, I don’t recall ever seeing a 100% humidity reading recorded by my stations since I bought my first one in 2005.

In short, I wish Davis would address these issues or a decent competitor would arrive on the scene. Sorry Davis, but adding bird spikes, creating an aerocone, adding the Davis logo to the anenometer, replacing the RH sensor in both the ISS and console with the SHT31 that hasn’t lived up to expectations (poor packaging and bad caps?), dropping the USB Weatherlink software that frequently drops the connection with my console (not once by the new WiFiLogger), doesn’t make me feel better.

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 johnd

  • Forecaster
  • *****
  • Posts: 4827
    • www.weatherstations.co.uk
What is most irritating is I, and many other folks posting in this Forum, thought by purchasing Davis equipment we were buying the best available in the consumer market.
....
In short, I wish Davis would address these issues or a decent competitor would arrive on the scene.

But what is the evidence that a VP2 is not 'the best available in the consumer market'? (At least in the sense that there's no better all-round station for performance and capability even if, arguably, some other device may offer better accuracy or whatever in some specific respect - though even there I think hard evidence is lacking, except possibly in respect of rainfall accuracy.)

I'm not particularly looking to defend Davis here, but more making the - fairly obvious - point that it's surely unrealistic to expect higher-end drift-free accuracy at the price-point of a VP2. If a potential competitor felt that there was money to be made from introducing a new range of stations with better all-round performance but at the VP2 price point then surely they would have done so.

But to date this simply hasn't happened, which tends to suggest that it's just not an attractive commercial proposition. VP2-type weather station volumes are in the several thousands per year, not the millions which make innovative mass production much more appealing.

Davis manufactures stations at VP2 prices because it makes it makes some initial design decisions and then sticks with them over extended production periods, with only limited design evolution along the way until a complete new model is released. Anything else would push up the price further.

Don't get me wrong. I think a strong competitor to Davis would be excellent news to drive innovative features and and overall performance of VP2-class station onwards at a faster pace, but equally I can understand the commercial reasons as to why there's no real sign of it happening at present. (Maybe there'll be an announcement of some new wonder station tomorrow even, just to prove me wrong, but somehow I doubt it.) For now IMO, the VP2 represents what is commercially possible at its price point. For users willing to pay substantially more there are other options.
« Last Edit: July 19, 2018, 01:32:18 PM by johnd »
Prodata Weather Systems
Prodata's FAQ/support site for Davis stations
Includes many details on 6313 Weatherlink console.
UK Davis Premier Dealer - All Davis stations, accessories and spares
Cambridge UK

Sorry, but I don't usually have time to help with individual issues by email unless you are a Prodata customer. Please post your issue in the relevant forum section here & I will comment there if I have anything useful to add.

Offline klschmidt

  • Member
  • *
  • Posts: 38
What is most irritating is I, and many other folks posting in this Forum, thought by purchasing Davis equipment we were buying the best available in the consumer market.
....
In short, I wish Davis would address these issues or a decent competitor would arrive on the scene.

But what is the evidence that a VP2 is not 'the best available in the consumer market'? (At least in the sense that there's no better all-round station for performance and capability even if, arguably, some other device may offer better accuracy or whatever in some specific respect - though even there I think hard evidence is lacking, except possibly in respect of rainfall accuracy.)

I'm not particularly looking to defend Davis here, but more making the - fairly obvious - point that it's surely unrealistic to expect higher-end drift-free accuracy at the price-point of a VP2. If a potential competitor felt that there was money to be made from introducing a new range of stations with better all-round performance but at the VP2 price point then surely they would have done so.

But to date this simply hasn't happened, which tends to suggest that it's just not an attractive commercial proposition. VP2-type weather station volumes are in the several thousands per year, not the millions which make innovative mass production much more appealing.

Davis manufactures stations at VP2 prices because it makes it makes some initial design decisions and then sticks with them over extended production periods, with only limited design evolution along the way until a complete new model is released. Anything else would push up the price further.

Don't get me wrong. I think a strong competitor to Davis would be excellent news to drive innovative features and and overall performance of VP2-class station onwards at a faster pace, but equally I can understand the commercial reasons as to why there's no real sign of it happening at present. (Maybe there'll be an announcement of some new wonder station tomorrow even, but I somehow doubt it.) For now, the VP2 IMO represents what is commercially possible at its price point. For users willing to pay substantially more there are other options.
Thank-you for the well written post johnd, I agree with you. In most manufactured products, the price increase vs quality increase is non-linear. (A station that is .5% more accurate may cost 20 times as much.) When providing feedback to Davis I think we should convey one of the following points: I as a customer/potential customer
(A) would pay more for a higher quality station.
(B) am happy with the price and quality of the current stations.
(C) would accept lower quality if the price could be lowered.
Complaining about the quality AND high price probably won't accomplish much.

I would like to provide highly accurate data to NWS and others. However, after spending a lot of time and effort trying to tweak my station to perfection, will forecasters be aware of the fact that my station is the best in the county? I stopped by the NWS office one day to mention my really good station siting. I got the impression that they would rather use data from a station that had been online longer than mine, and therefore had more history. They were not aware of the fact that this was a Vantage Vue mounted about two feet above a metal roof and surrounded by grain bins. After this visit it was very clear to me that any efforts that I put into high accuracy are for my own benefit only; other's don't know or care. If I'm the only one who cares about my station, I find it easier to say "hmmm... dense fog is more humid than the 96% my Davis is showing".