Odd Si7021 humidity readings over 6 nodes in same jar

[WhiteHare]

I will continue looking into this and try to keep anyone else interest updated on my findings about this humidity reading creep I seem to have.

Mark me as interested.  I look forward to hearing what you find out.

[WhiteHare]

The last time I went looking for a really high accuracy humidity sensor (around 3 years ago), the one I settled on was from this company:  http://www.ist-usadivision.com/sensors/humidity/

They're not cheap though.  They also have some really high accuracy temperature sensors too that, at the time anyway, seemed excellent.

I mention here just in case someone wanted a reference for testing their other sensors against. 

[TomWS]

TomWS, what is the average humidity where you have your sensors deployed? I have a sinking feeling that these sensors work perfectly if the humidity stays in the 45-60% range and anything above that will cause them to creep up.  These sensors were used to help avoid mold from extra moister by my fish tanks, but how fast are they creeping up?  I will continue looking into this and try to keep anyone else interest updated on my findings about this humidity reading creep I seem to have.

The space where the most recent units were deployed were two rooms in my workshop.  The workshop had been idle all summer long and the humidity was between 75-78% when I arrived back from vacation.  This measured by one of my WiFi sensors with the HIH-6120 TH sensor.

The Si7021s were the last two of a batch I purchased over 6 months ago and had spent the summer in a garage that was even more damp than 75% (although I didn't have a sensor in there).  The units were housed in an MS3 foil envelope, but the desiccant had long been past its useful life and the envelope was only taped shut with masking tape.

When I deployed the two motes in the rooms adjoining the workshop they immediately registered around 78%, a bit higher than the workshop, but not extremely so.  I then installed a dehumidifier in the workshop to get the humidity down to a proper level and, as the readings dropped in the workshop, the readings tracked in the two rooms with the motes.  The readings are now:
Workshop 76.3°F 50%
Toms Office 78.08°F 48% RH
Finishing Room 74.66°F 51% RH

In another location, where I have a mix of WiFi HIH-6120, Tiny TH Motes with Si7021, and one Mote with DHT-22 & DS18B20 (for outdoor temp), all the indoor rooms are within +/-2% of each other, the Motes with Si7021 the highest of 4 rooms.   This location has very little in the way of conditioning and the indoor rooms in the spring and summer will easily reach into the mid to high 70s RH.  In the winter, however, the temps will drop into the teens (degrees F) 

Net: These sensors work for me even when stored in less than ideal conditions.

[TomWS]

They're not cheap though. 

Much cheaper by the pack of 50...

[joelucid]

Do you happen to know what the humidity level at your lab was in the 48 hours after your si7021 fabrication?

It was in the 50-60% range I'd say.

Reading the si7021 datasheet is a little unsettling: after reflow rehydrate, then whenever it's too dry or too wet for longer periods of time bake and rehydrate again. Finally if it's really humid remember to turn on the embedded heater to avoid over-hydration. And it really helps to say you love them and bring flowers from time to time ...

I don't know about you, but to me these chips seem a bit high maintenance 

[CallOne_Phil]

I am probably just blind, but is there example code for the new sensor on the R2 boards?  I saw there is some bme related code inside the motionmote example, but nothing on the weather shield stuff that jumped out at me.  It's not the end of the world if there isn't, but I figured I would try whatever I can to limit user error in my testing.

[sparky]

I am probably just blind, but is there example code for the new sensor on the R2 boards?  I saw there is some bme related code inside the motionmote example, but nothing on the weather shield stuff that jumped out at me.  It's not the end of the world if there isn't, but I figured I would try whatever I can to limit user error in my testing.

Did you look at these sketch's that Felix pointed to in the blog;
https://github.com/LowPowerLab/WeatherShield
https://github.com/LowPowerLab/RFM69/blob/master/Examples/WeatherNode/WeatherNode.ino

[CallOne_Phil]

It seems the new bme280 R2 code never made it there and the newest code is WeatherShield_Si7021_only.ino from 6 months ago.

[WhiteHare]

It turns out I have the same sensor boards as in the article that Joe referred to.  So, for grins, I hooked up both the si7021 and the BME280 to see how they compare.

For the Si7021, using the Adafruit library, I'm getting:
Humidity:    48.93   Temperature: 21.72

For the BME280, using the Sparkfun library, I'm getting:

Temperature: 22.77 degrees C
Temperature: 73.00 degrees F
Pressure: 99002.00 Pa
Altitude: 202.13m
Altitude: 663.15ft
%RH: 47.00 %

So, actually, for humidity, they're fairly close for an off-the-cuff, rough-and-ready test.  Maybe next weekend or thereabouts  I'll try a more careful comparison.

Conveniently enough, the boards have different I2C addresses, so I suppose one could run both from the same sensor node and then average their results, if one were so inclined.

[Felix]

@CallOne_Phil, My apologies, I thought I uploaded the sketch to github. Let me get back to this asap, I will update the blog entry also.

There's the MotionMote example that reads the BME280 sensor.

Otherwise any other BME280 library would work.

[joelucid]

The main observation from this is that the sensors reading the highest humidity and the slowest to recover all came from higher humidity sections of my home.  I had the the 2 worst sensors in the basement near my fish tanks which can drive the humidity up to 70ish% on bad summer days.

That is interesting. I use these sensors to determine by measurement of absolute humidity inside/outside whether to run a fan to exchange air as dehumidifier. Outside it's often close to 100% humid (e.g. during the night) - so this problem could easily spoil my approach, letting me miss good opportunities because the outside sensor lags.

I've thrown together a little project which uses the heater to keep the sensor humidity always below 80%. Obviously for once not a coin cell project ... I should still be able to measure absolute humidity provided both sensors heat up to the same degree.

Joe

[WhiteHare]

The main observation from this is that the sensors reading the highest humidity and the slowest to recover all came from higher humidity sections of my home.  I had the the 2 worst sensors in the basement near my fish tanks which can drive the humidity up to 70ish% on bad summer days.

That is interesting. I use these sensors to determine by measurement of absolute humidity inside/outside whether to run a fan to exchange air as dehumidifier. Outside it's often close to 100% humid (e.g. during the night) - so this problem could easily spoil my approach, letting me miss good opportunities because the outside sensor lags.

I've thrown together a little project which uses the heater to keep the sensor humidity always below 80%. Obviously for once not a coin cell project ... I should still be able to measure absolute humidity provided both sensors heat up to the same degree.

Joe

At least in the places I've lived, usually absolute humidity doesn't change all that rapidly.  One of the exceptions is a storm front coming through, but you could handle that as a separate case (usually identified by rapid changes in air pressure). So, although not a perfect solution, perhaps you could grab an ourdoor sensor reading late in the day when the RH is lower to compute the dew point and then make your air exchange decisions based on that during the night when the RH is no longer in the goldilocks zone. 

Ahem...  Well, to quote Simon Pegg from Mission Impossible, "I'm spitballing this.  It's not all going to be gold."   

[joelucid]

t least in the places I've lived, usually absolute humidity doesn't change all that rapidly.  One of the exceptions is a storm front coming through, but you could handle that as a separate case (usually identified by rapid changes in air pressure). So, although not a perfect solution, perhaps you could grab an ourdoor sensor reading late in the day when the RH is lower to compute the dew point and then make your air exchange decisions based on that during the night when the RH is no longer in the goldilocks zone. 

Oh it changes quite rapidly here. But I think the problem with long term exposure to high humidity environments is not the single wrong measurement when it's wet. It's the long term drift to higher humidity readings that happens with such exposure. The datasheet says it can take substantially longer than the standard "couple of days" to return to normal afterwards and that indeed the error can become permanent.

That's why I think one should just not expose the sensor to >80% humidity by switching on the heater when it gets that wet. The datasheet sort of guides that direction, too, although staying clear of a general recommendation.

I built two weather motes yesterday and they still measure 10% too dry after reflow. Being the impatient type they are now rehydrating on the balcony at a heater controlled 70-75% humidity.

Joe

[WhiteHare]

t least in the places I've lived, usually absolute humidity doesn't change all that rapidly.  One of the exceptions is a storm front coming through, but you could handle that as a separate case (usually identified by rapid changes in air pressure). So, although not a perfect solution, perhaps you could grab an ourdoor sensor reading late in the day when the RH is lower to compute the dew point and then make your air exchange decisions based on that during the night when the RH is no longer in the goldilocks zone. 

Oh it changes quite rapidly here. But I think the problem with long term exposure to high humidity environments is not the single wrong measurement when it's wet. It's the long term drift to higher humidity readings that happens with such exposure. The datasheet says it can take substantially longer than the standard "couple of days" to return to normal afterwards and that indeed the error can become permanent.

That's why I think one should just not expose the sensor to >80% humidity by switching on the heater when it gets that wet. The datasheet sort of guides that direction, too, although staying clear of a general recommendation.

I built two weather motes yesterday and they still measure 10% too dry after reflow. Being the impatient type they are now rehydrating on the balcony at a heater controlled 70-75% humidity.

Joe

Does the BME280 have the same drawbacks?  If not, it sounds better to just use the BME280 instead for outdoors.  Indoors, the si7120 might be fine, because if your house is >80% humidity, then  it's probably a mold forest and you've got much bigger problems than just inaccurate sensors.

[joelucid]

Does the BME280 have the same drawbacks?  If not, it sounds better to just use the BME280 instead for outdoors.  Indoors, the si7120 might be fine, because if your house is >80% humidity, then  it's probably a mold forest and you've got much bigger problems than just inaccurate sensors.

They have the same reconditioning stuff in their datasheet. But other than the si7021 they claim it's only needed if operated out of fairly broad ranges e.g. after soldering. I've ordered a couple to try out. Plus some si7021 with cover in case I'm damaging the sensors during reflow.

Joe