[Off Topic] Silicon Labs SI7021 and Humidity

[hmmpic]

I have user the Silicon Labs SI7021 i some project, over 100 pcs. One general problem i face is when using the sensor outdoor.
When the Humidity >85%Rh and the temperature <8dC, close to the dew point the sensor came into a unstable state over time.
The Humidity raise to max and it stay there! The only way to bring it back to normal is dry it (4.6 in the datasheet)

Is this a normal behavior for sensors made with "polymer sensor film"

I think the sensor are not useful for outdoor use in low temperature and high humidity.
Is my conclusion right?

From the datasheet:
4.3. Prolonged Exposure to High Humidity
Prolonged exposure to high humidity will result in a gradual upward drift of the RH reading. The shift in sensor
reading resulting from this drift will generally disappear slowly under normal ambient conditions. The amount of
shift is proportional to the magnitude of relative humidity and the length of exposure. In the case of lengthy
exposure to high humidity, some of the resulting shift may persist indefinitely under typical conditions. It is generally
possible to substantially reverse this affect by baking the device (see Section “4.6. Bake/Hydrate Procedure” ).


4.6. Bake/Hydrate Procedure
After exposure to extremes of temperature and/or humidity for prolonged periods, the polymer sensor film can
become either very dry or very wet, in each case the result is either high or low relative humidity readings. Under
normal operating conditions, the induced error will diminish over time. From a very dry condition, such as after
shipment and soldering, the error will diminish over a few days at typical controlled ambient conditions, e.g.,
48 hours of 45 ≤ %RH ≤ 55. However, from a very wet condition, recovery may take significantly longer. To
accelerate recovery from a wet condition, a bake and hydrate cycle can be implemented. This operation consists of
the following steps:

[temtronic]

Yes, normal (nature of the beast) and the documentation tells how to 'recover' from it.
I assume they will need to be recalibrated as well after the 'recovery' process is completed.
Perhaps design/build a PIC based testing/recalibration unit....that'd keep this 'thread' relevent

I can't offer any alternatives as I was only interested in temperture sensors when I did remote SCADA 3 decades ago.

[Ttelmah]

Polyimide film humidity sensors are a common technology.
Basically different types of film, and you measure the change of
capacitance as these absorb water. They dominate the market, but there
are other technologies. Also variations in the actual structure of the
film, and different methods to actually measure the capacitance. All
affect the usable range. Ceramic sensors have issues with dirt in
the air for example.
Failures at different parts of the range, like you are experiencing, is
fairly common. This paper describes the three commonest types:
<https://www.fierceelectronics.com/components/choosing-a-humidity-sensor-a-review-three-technologies>
Ceramic coated resistive sensors are the most reliable type, but will
almost certainly cost more than the devices you have.

[newguy]

I have experience with a different sensor (TI HDC2080) but it's basically the same way. It has a built-in heater which can be used to "burn off" condensation....I suspect that a lot of similar sensors likewise have such heaters.

To test the toughness of the LoRa sensor we created using the HDC2080 we actually threw it into a river with a weight to sink it. Even submerged it still merrily worked - sending packets to our gateway we had locked in our vehicle about 400m away. The RH reading of course registered 100% but over perhaps a weeks' worth of drying, the RH readings returned to normal.

[hmmpic]

Thanks:-)
I already use the internal heater. It is working solution to bring the sensor back to operating at Rh again. Downside is it drain power from the battery. The sensor is a LORA sensor and are expected to be running for years, therefore using the heater inside the SI7021 is only used when it is really needed. Some dew point calculation and some lookup list for temp and a counter for to know how long the Rh was calculated to be out of spec...

And it is all controlled by a PIC16LF pic:-)

[temtronic]

hmm.... battery powered equipment,in Denmark, and you need YEARS of operational life ? Must be BIG batteries !! I hear you have a looooong Winter !!
Can I assume you've done everything to minimize power ?

[newguy]

Battery life: our prototype achieved 700,000 wireless transmissions on a pair of AA alkaline batteries. This was with one transmission every 20s. Ratchet that back to a more realistic 20 minutes per, and you hit the shelf life of the batteries.

LoRa is pretty cool in a bunch of different ways.

[hmmpic]

Maybe not enough, but a lot, 16uA in sleep about 1,4C, this is the biggest drainer.
We only tx a package every 30 minutes, and do a sample every 5 min.
Complete system use about 4C/day and the bat is a CR123 lithium(5400C) normally it run at 15-22 degree celsius.
When outdoor or in cool place i have maybe 40-60% of the bat power so still years.
We have some sensors placed at -28degree Celsius for about 8 month now without any problem, so it run.

tx every 20 sec and 700000 cycle about 160 days on your AA, you have a nice system with real low power, or maybe low power Lora TX. We use 17dbm for to get ok range.

[Ttelmah]

I have to say that 'honestly' the best thing to do is to work with a bigger
battery, and you can then use the heater more often. Remember you
can reduce the current you have to use to keep the chip dry, by
improving the insulation round the chip. It only needs a clear air path to
the sensor. The rest of the casing can be 'lagged'.

[newguy]

Max output is 20dBm, but in north america we're limited to max SF10 as opposed to the SF12 allowed in Europe. Testing was mainly on SF7, but even with that spread factor we found we could penetrate many meters of reinforced concrete.

The batteries aren't actually completely drained, however. The LoRa module is only rated down to 2.3V and I started having issues with it at around 2.4V so that's where I ended the battery drain experiment. The PIC can easily run lower than that, but without an RF link to the outside world, that isn't much use at all, is it?