[qldbureau]

There was some talk about sht10 sensors failing this week. Another one of mine is failing now, and I am getting less time out of each replacement. Usually the humidity sensor fails by rapidly starting to under read until it stops registering moisture altogether.

I am going for an sht31 to try this time and have ordered this to put in my radiation shield:

https://www.aliexpress.com/item/TR0231-T...4c4dZn2Nuw

There are plastic shell versions available as well. I hope the screen does not mask the parameters.

[41south]

-40 deg C temp yesterday, those nights are cold over there! 

Another failure symptom or rogue data?

[qldbureau]

-40 deg C temp yesterday, those nights are cold over there! 

Another failure symptom or rogue data?

Another failure symptom. I've edited the data files again to remove. This time it is randomly dropping down. In my experience if the connection was bad and no sensor present everything would read "0", so I do not think the connection has been lost.

Now I have not climbed up the mast yet to investigate, so it is possible ants might have nested around the sensor; which is why I like the idea of the replacement I have linked. The other thing is I am putting together a new radiation shield to go with it.

[engolling]

Hello,

yesterday I had the third SHT31 dying and bringing the Arduino to halt. It is fact that the Arduino gets stuck because there is a very strange behavior on the I2C bus. I had not the chance to check with an oszi until now what is happening.
First I thought is was a fault of the library (see here: https://www.meteocercal.info/forum/Threa...y-blocking) but it is not.
So I must confess it is possible to buy very bad quality SHT 31, maybe containing some counterfeit SHTs … and I did it  .

At the moment I'm thinking of making my own breakout PCBs and getting the SHT sensors for example from Farnell.
Does anybody else has similar problems.

Best regards,
engolling

[41south]

I've used Adafruit supplied breakout boards and have just ordered 3 from an Aliexpress supplier so will see how they go. It seems that the more expensive ones have a protective membrane over the top to help keep out unwanted moisture etc

As per this one ...  https://www.adafruit.com/product/2857

[JT118]

These for the SHT11

https://www.ebay.co.uk/itm/SENSIRION-SF1...2828379749?

and these for the SHT2XX and 3XX

https://www.ebay.co.uk/itm/SENSIRION-SF2...1948246757?

and this for the board;

https://www.ebay.co.uk/itm/Electrolube-W...4036552581?

Have helped even in a Scottish climate! Though my SHT31 is beginning to fail.

[uncle_bob]

I've had another SHT11 fail. Got two months out of the last one. Not sure why these have become so unreliable. Maybe it's an issue with Chinese manufacturers? I used to get a least a year out of them.

Funny thing, I've got two previously failed SHT11 now working in another project. One of them gives a bad humidity reading occasionally but has been reading good for weeks. This behavior starts me wondering if the WeatherDuino code is maybe pushing them too hard or something.

Anyhow, I've ordered some SHT31's but my weather station is offline till they arrive in about 3 weeks. Fingers X'ed these last longer.

Can someone remind me which sensor the Davis stations uses? I thought it was the SHT 11? Whatever they are they seem to last quite a while.

[engolling]

My sensors not even has been outside yet… so I think I bought bad quality stuff. 

So that's the overview of the sensor manufacturer:
https://www.sensirion.com/en/environment...lications/
According to the manufacturers info the sensor should be able to handle outside conditions, you only might get an offset if humidity is very high for a long time, which is self healing.


I do not think the communication with the Arduino has any influence of the lifetime, since the sensor is powered up all the time and suitable for a lot of applications having at lest the same demands.

That's why I'm up to sourcing from a more expensive supplier or trying to get some breakout boards on my own and buy the sensor with protective cover from a big named supplier.

Regards,
engolling

[werk_ag]

...

Funny thing, I've got two previously failed SHT11 now working in another project. One of them gives a bad humidity reading occasionally but has been reading good for weeks. This behavior starts me wondering if the WeatherDuino code is maybe pushing them too hard or something.

Hello uncle_bob

Poor WeatherDuino...  
The STH31 sensors can be read several times a second, however WeatherDuino only reads the T/H sensor (no matter which sensor is used) at 17 seconds intervals. Doesn't seem that the WeatherDuino code is pushing the sensor too hard.

Regards

[qldbureau]

Hello everyone. I have commissioned my new and first SHT31, the one as shown in the link at the first post. I built a new radiation shield to house it in as I said I would. I am sorry I did not stop to take construction photographs and I should have.

I opened the casing and found that the sensor is very well wrapped in a membrane which is apparently permeable yet water resistant. The mesh arrangement is air permeable though I can see very little light through it. I was skeptical, however the response to temperature changes both up and down is almost instant, likewise humidity.

Shield consists of:
9 manufactured white pot plant saucers - 250mm.
90mm diameter PVC storm water pipe - 200mm long.
3 x 6mm stainless steel threaded rod stems.
Spacers between saucers.
Fan
All nuts and bolts stainless steel.
4 days solid work! I was on leave.

My previous shield, although worked well for a couple of years was nowhere near as robust. To make servicing easier I put plugs and sockets at the top, but out of the weather. The failure this time around was not the sensor as I expected; it was corrosion in the cable plugs! Regardless the rebuild was on the cards.
The sensor casing is inside the pipe halfway up. A hole in the side lets the cable out, and I made cable holes through the lower half plates to the bottom. Strain relief by cable ties, one at entry to the pipe back to one of the threaded rods. The other at the bottom to one of the fan bolts.

The SHT31 comes as illustrated, and its signal lead is 500mm in length. Slightly annoying and limiting where to put it and the transmitter, granted the length limitation for I2C, the cable could have been a little more generous. Because separable cables everywhere means more failure points, I was determined this time to eliminate them.

Thus my version 2.50 TX is up the mast now in the PVC box, I changed the connector headers to PCB screw headers and the sensor is hardwired with no joints. The fan has received the same treatment; I soldered an extension close to the motor and used heat shrink.
If I need to maintain anything, I have found it is no effort getting my mast down and up. I can reach the transmitter with a three step ladder, so I can still easily pull the Nano out for programming.

It has been online roughly two weeks, and is inline with nearby stations. The official station about 5km away I read typically up to 2 degrees higher, which I consider very reasonable. Overnight lows seem to be about 1 degree lower. I hope this SHT31 will last! It should be very well protected from the elements.