Humidity sensor installed outside

[mackowiakp]

I come back to the old topic of the humidity sensor installed outside the house.
The problem is that I live in an area with very high humidity. 80-90% is "normal" and in the morning it is also 100%.
There is no problem with temperature measurement, but with humidity. I used DTH22, BME280, BME680 and none lasted more than 9-10 months.
There were no signs of corrosion or anything like that on them.
I would like to point out that the sensor is installed under the roof at a height of 2 m, on the northern side, it does not rain on it. The sensors were placed in the "cup" with the opening facing down.
Now I'm going to use something that has the markings (simultaneously) SI7021, SHT21, HTU21.
Now I want to install something like this from the photo as a sensor housing.
This is called the "Stevenson Weather Cage" (3D printout).
I would like to ask my colleagues what experience they have with outdoor humidity sensors? Is there any better method to protect the sensor?

Obudowa.jpg (67.36 KiB) Viewed 4007 times

[TD-er]

Just a simple thought...
Why don't you place some "heater" inside the enclosure?
OK, waiting till you stop laughing

...
...

The idea is that with a constant amount of heat, you simply create a constant temperature offset.
A somewhat closed confinement just acts as a heat spreader or like an aluminium cooling block.
It has a constant coefficient expressed in K/W (Kelvin per Watt)

Thus a block with such a coefficient of 1K/Watt connected to something generating 10 Watt of heat, will heat up the block by 10K compared to the ambient temperature.

Given this constant temperature offset, you can measure the temperature and (relative)humidity and correct both.
First you compute the dew point temperature based on these measurements.
Then subtract the temperature offset from the measured temperature and use that with the dew point temperature to compute the corrected relative humidity.
See the standard conversions at the bottom of the sysvars page in ESPEasy

N.B. the BME280 plugin already has this built in for ages when you set a temperature offset.

Having a higher internal temperature, prevents condensation. Or at least as long as the internal temperature exceeds the dew point temperature (as by the definition of "dew point temperature")

However, the "constant temperature offset" may not always be as constant as hoped.
For example wind will change the coefficient and I can imagine ice may also have some effect.
So you can also have another temperature sensor mounted outside the enclosure to properly compute the temperature difference.

[mackowiakp]

As for the temperature shift, I actually have no problem. I use DS18b20 for temperature measurement.
It is hermetic and can be placed somewhere next to the BMExxx. I only use BMExxx for humidity measurement, ignoring other data from it.
I even have a second sensor inside the winter garden to read the pressure, because the temperature there is always above zero and the wind has no influence on the readings.
So I have OK readings of temperature and pressure for years.
The problem is with the notorious breakdown of the humidity sensor.
The probable cause is - as I wrote - a very high and constant level, reaching even 100%.
The question is whether there is any sensor that can withstand it for a long time.

[TD-er]

Like I said, add a small "heater" in the same enclosure as the BME280 and you should be good to go.
The "heater" can also be an ESP

[mackowiakp]

Will, try. But potential problem could be influence of higher temperature to humidity readouts.

[TD-er]

Will, try. But potential problem could be influence of higher temperature to humidity readouts.

That's exactly what is intended here, and thus you need to correct those.

The whole idea of this is to affect humidity inside the enclosure so you will have a lower relative humidity by increasing the temperature.
And thus you need to correct the measurements in software.

[TD-er]

Just to explain my idea in other words:

A humidity sensor measures the relative (!!!) humidity.
The amount of water in air highly depends on the temperature.
Higher temperature => air can hold more water.
Thus: if you heat up air with some relative (!!!) humidity value of X percent, the relative humidity will drop since the capacity to hold water will increase.

When cooling the same amount of air with X percent of humidity will cause the relative (!!!) humidity to rise.
This cannot rise forever as at some temperature, the amount of water in the air is "full" (100% humidity)
The water will then condensate from gas to liquid and the temperature when this will happen is called "dew point temperature".
This dew point temperature is very specific to the temperature and relative humidity (and air pressure, but that will not change enough around your house to have a significant effect)

Thus when you experience "100% humidity", this simply means the temperature is at or below the dew point temperature.
By increasing the temperature in your sensor enclosure, you essentially lift the internal temperature above the dew point temperature (not always, but often enough) to protect the sensor from condensation.

A typical "Stevenson Screen" setup only protects the sensors from rain, not from condensation.
Increasing the temperature does protect from condensation (to some level).

Since you will change the relative humidity around the sensor by heating it, you have to compensate for this.
Thus you need to know what temperature difference there is => see my earlier reply on how to compensate.

[chromo23]

A simple fix you could try is to put some kind of sponge around or at least in front of the sensor (i use usually dish sponges).
I had similar issues in my (very small) bathroom, especially when showering. There i used the sensor (bme280) for a fan.
Worked also for my more delicate dht22 which had even more issues with high humidity.

[chromo23]

I don't know how reliable this is, but it has a similar approach:
https://www.adafruit.com/product/4099

[TD-er]

A sponge may also "filter" the air.
Not only will it keep out some dirt, but also comparable with a mathematical filter where it adds a delay in response.

Not sure how it may behave when it freezes.

[ThomasB]

The whole idea of this is to affect humidity inside the enclosure so you will have a lower relative humidity by increasing the temperature.
And thus you need to correct the measurements in software.

+1. This would be a painless / simple solution.

Perhaps use a 12V 20 Watt yellow or red tinted automobile bulb; Operate it at at lower voltage (~10V) for long life and to reduce wattage. As a plus, you can turn your 3D printed housing into a cool looking outdoor light fixture.

https://www.sylvania-automotive.com/syl ... L.BP2.html

The sponge idea sounds good too. But with periods of 100% humidity I suspect it would eventually become saturated, which would affect readings. And a constantly wet sponge may promote mold growth on it, which could also affect readings.

- Thomas

[chromo23]

The sponge idea sounds good too. But with periods of 100% humidity I suspect it would eventually become saturated, which would affect readings. And a constantly wet sponge may promote mold growth on it, which could also affect readings.

It´s heavily tested already in my bathroom for two years. And i am very good at long and hot showers until i am unable to find my way out because of all the steam. (i can think best when showering)
So far no mold and no funny readings.
Having it outside with temperatures below zero is something to test (gets more difficult every year ). I therefore would maybe increase the surface area of the "sponge-cage" and give it some space to the sensor surface...

[TD-er]

[...]
Having it outside with temperatures below zero is something to test (gets more difficult every year ). [...]

What a remarkable level of sacrifice, to test showering with temperatures below zero, just to see how long the sensor will last

[chromo23]

Everything for science!

[mackowiakp]

A humidity sensor measures the relative (!!!) humidity.
The amount of water in air highly depends on the temperature.
Higher temperature => air can hold more water.
Thus: if you heat up air with some relative (!!!) humidity value of X percent, the relative humidity will drop since the capacity to hold water will increase.

The phenomenon is known to me. I have a gas stove in the basement.
In accordance with the regulations in force in my country, the outside air is supplied to the basement through a pipe with a diameter of 20 cm.
The fumes from the stove are discharged from the basement through a single channel in the chimney. Additionally, there is a second channel in the chimney for exhausting air from the room with the stove in the basement.
This results in very intensive airing of the room.
The temperature in the room is constant throughout the year 19-21C.
As soon as the temperature outside is above 20C, I have moisture in the basement. Hence, I put in there an air dehumidifier that works on the principle of an "inverted refrigerator".
And so in summer I have 90% humidity there. Probably without a dehumidifier, it would rain constantly in the basement in summer
And yet DTH22 has been working there for about 4 years without any problems.
Thus, the working conditions of the sensor in the basement are almost identical to those outside.
So why do outdoor sensors get damaged on a regular basis?
It is not entirely clear to me.

[ThomasB]

And yet DTH22 has been working there for about 4 years without any problems.
Thus, the working conditions of the sensor in the basement are almost identical to those outside.
So why do outdoor sensors get damaged on a regular basis?

I would suspect the stable temperatures and humidity in the indoor installation are partly responsible for longer life. Outdoor environments vary much more and that may have something to do with the shorter life.

Regardless, these cheap hobby sensors are not known for having long life. For example:
https://arduino.stackexchange.com/quest ... -some-time

Perhaps the most practical solution is to keep an inventory of spares and design the enclosure for easy sensor swaps. And have your ESPEasy device email you when the sensor fails.

- Thomas

[TD-er]

Yep, those DHTxx units vary in quality from "moah" to "absolute garbage"
I would consider it just luck when they last this long as they are known for getting "stuck" at some humidity level after which you need to extensively dry them (like in an oven) to get them to work again.

Don't know how they are mounted? Like is there an ESP mounted close to them?
Is there a near constant draft of air moving around the sensor?

I don't know what actually kills these sensors. Is it some kind of delamination of materials? mold? Chemical reaction (e.g. oxidation)?

[mackowiakp]

Both in the case of DTH mounted in the basement and sensors mounted outdoor, they are mounted outside the housing with ESP. So the temperature of ESP has no influence on them.
There is intense airflow in the basement for the reasons I wrote earlier.
The sensor installed outside is - also as I wrote - covered and located under the roof. So there is not direct rain on it, direct sunlight and the air flow is as it results from the current strength of the wind.
Of course, all sensors are CNN production (China No-Name).
I think that @ThomasB is right. I have to buy several sensors for outdoor installation and replace them if they are damaged, hoping that one will survive longer. Such "Near to hot-plugable". That is, connected, for example, with gel-coated screw joints.
There are miniature "boxes" containing gel that snap onto these types of junctions. This, of course, protects against corrosion and is completely removable when you need to access the screws.