Relative Humidity (& subsequently Dew Point Temp) reading seems to be incorrect/too high

Hi All,

First time posting in this community, but I’m proud to be a member & have been lurking around for some time. I first got into this whole weather data thing while I was making my original plans to ‘smart-ify’ my home - plan was to use the weather data to assist with irrigation and to a lesser degree the HVAC system/anything else that might benefit from it.

I purchased an irrigation controller early on since they are better for the environment (I hate people who have their lawns set to water right after/during storms), they save money - mine paid for itself in 1 season, and if it’s a fairly dry season parts of the lawn start to get burn while others are quite lush.
I got a RainMachine since it’s supposed to be open source (been having issues with the WF plugin for it, but that’s for another time) & I liked the level of customizability/it could use WU data & DarkSky data. It paid for itself within the first season (water is $$ here) and early on I figured I’d get a cheap acurite 5-1 Sensor to see how it’s data compares with what that sensor sees.

I mounted the sensor right above the top corner of the roof (my sky sensor is still there actually, I’ll be moving it so that it isn’t partially obstructed by a tree some 30-40 feet above it. But it is unobstructed in the horizontal axis and given the number of trees here the sun is going to be blocked by them at some point during the day and it’s right at the end of the tree canopy ‘line’ in regards to rain detection… for now). Given that it was just a couple of feet above the roof it got very high temp readings during the day in the summer, even with the solar powered aspirating fan (which wouldn’t really help much in that situation). But still, the sensor had a gold medal 85% of the year on WU.

One day, while it was still beta testing, I came across the WeatherFlow and I loved what I saw. Based on the #s the unit should be as actuate as at least your entry level Davis system while costing less & having more (except for the physical readouts, but I plan to remedy that with a smart home display). I’ve seen that a number of people on the forums often use the price difference/age of the product as an excuse for slight irregularities in the data, but as far as I can tell if the product is placed correctly and working correctly it puts out data that is in line with its competitors (but there lies the rub).

So, I have a decent place to put the sky for now (have a mast that I will mount under an eave so it can get clear wind & rain readings, not that the rain readings are bad now. Sometimes if the rainfall is super feint the sensor will either not register it because a couple leaves will catch it or it will register, but it will be below the measurement threshold - but this is in the sub 0.01” range).

Believe it or not, the Air was the tricky sensor for me to place. Now, given my experience with the Acurite I know a sensor near dense objects exposed to sunlight will yield higher than actual temps, so placing the sensor in the shade makes sense. But, when you are surrounded by as many trees as I am (and other various vegetation) an area that is always shady will probably be a bit moist. I originally placed the sensor in my screened porch, in line with the top edge of the screening so that it should have gotten some decent airflow. But the humidity numbers were high as far as I could tell (no gold badge on WU & compared to other sensors the humidity did look high).
Since the porch is near the deck, which would obstruct air flow to a degree, I decided to move the sensor.

This is where the matter got a bit confusing - I wanted actual temp readings, but there is slightly conflicting information out there. And then some of it isn’t possible for me (like put it X number of feet away from vegetation or buildings or asphalt - not possible! There are either the road/driveway, the house, or the yard with plenty of trees in it. I’m not going to put up a 50’ mast for the sensors & even the areas exposed to a lot of sun aren’t always in it & they are covered in grass). So I did some research on radiation shields, ended up buying the acurite one recommended, but the gist I got was it the radiation shield works correctly it shouldn’t matter if the device is 100% in the sun (I actually came away thinking it would be a good thing as things would dry up). It just became a matter of the type of shield & where to place it.

I saw lots of people complaining about how the acurite shield is a tight fit (specifically 1 or 2 of the top plates). This made sense to me so I put a layer of aluminized fiberglass insulating tape around the top 1” of the air that’s made to withstand temps or 400 degrees and reflect up to 1200 degrees (I’m a big ‘car guy’ who works on & tracks his cars. I use tapes/heat shields like this on lots of little things & have used it in the past on a radar detector that was mounted near the radiator. The engineers at Bell said the tape wouldn’t effect the detector so I figured an inch wouldn’t hurt the lightning detector in the air - so far it seems fine). I also covered the top most plate of the radiation shield in aluminum tape, the kind used in HVAC systems/repairs. I was reading a few white papers on radiation shields and they had favorable results when they covered a standard Gill-type shield with the stuff - nothing crazy, but it lowered the temp enough when exposed to the sun that it seemed worth the 30 min of work.

Placement became the last issue. I still wasn’t/am not 100% on where it should be mounted in regard to sun exposure, but based on instructions from Davis equipment they seem to recommend at least 5’ off the ground, especially if grass is involved, good airflow, and obviously away from objects that can store latent heat. I also wanted it to be exposed to the sun at least part of the time, so ideally on the south side of the house, but I ended up mounting it on the corner of the deck, furthest from any buildings.

The airflow is fairly decent, it gets exposed to the sun for at least a portion of the day (even though it kind of is under a tree, the path the sun takes gets around this issue), and it’s 5 feet above the ground - which is also bare soil. At one point this was a more active flow bed, now it only has a few bushes here and there. There are some sprinklers, but the shield is above and behind them. Plus they are set to water 10 min ever other day ‘during a regular hot summer day’, aka when the ET value is 0.23 and field capacity is low. Given the general location of the flower beds (it wraps around the house, so the zone does as well, and the southern facing portion which is more crowded is covered by the eave & thus doesn’t need to be watered much) the controller doesn’t water the flower bed often. When it does it’s for a minute or two. Once or twice a month maybe the full 10.

So I thought this would be a good place for the sensor and mounted it there. I knew that I would have to deal with the system recalibrating for the new location & even though it’s not above grass, it’s near it, so that could effect humidity. But given it’s location when the sun rises it hits that part of the lawn first so I hoped it would cancel out.

However, after some weeks I’m still getting fairly high humidity levels. Especially at night (and yes, still no gold badge!). I’ve compared my sensor with other ones in the area that do get gold badges (which seems harder now that the PWS has been overhauled) and I don’t think I would get one even if the sensor was 100% accurate given where I live, but I want the humidity value to be as close to correct as possible. 2% or so off is tolerable and within spec, and that’s where I thought I was, except there had been a few instances where the Air was reporting much larger deviations.

So, being the nerd that I am, I ordered a WirelessTag Pro ALS. For those that don’t know these are some of the cheapest ‘smart’ tags I have found (ranging from mid $20 to mid $30 for the ‘regular’ temp/humidity/motion tags, be it motion using accelerometers, magnetic sensor, or PIR. With newer tags also having the option of an Ambient Light Sensor & can take measurements, store them in memory, and upload when they re-establish connection). To boot, these are also the most accurate smart sensors I have come across in the last 3-4 years. When I say accurate, I mean they are accurate. They are used by a number of large companies such as LG etc (the whole list is on their site). They are a pain to integrate into over-the-counter smart home environments like Vera as they aren’t popular enough, but people have done the work to integrate them (and you can program the tags to send out REST calls). With OpenHAB and other such products the community has developed full integration for them.

Unfortunately, the water resistant Pro ALS Tag (Pro meaning the number of data points it collects, the bit size of its memory/how granular the data points are, and the ability to log the data to memory when out of range. ALS for Ambient Light Sensor - this rage collects Temp, Relative Humidity & Lux values. It can determine motion based on quick changes in light detected, but I’m not sure if it includes an accelerometer or 3D compass like the others - they are used by people to get pinged when a fridge/exterior/garage/mailbox door is opened etc) didn’t work out as planned, but they still make the regular Pro ALS so I got one of those.
When I got home a couple of days ago I paired the tag, let it acclimate, and once it was done raining I placed it right next to the solar shield housing the air. The idea is that it will record temp, humidity, and Lux data that can be compared with the data from the air. The Lux data could be used as a proxy to see if there is enough sun hitting the shield that would result in a significant temp/humidity reading.

Don’t worry, pics off all this are included as I know I can only explain the setup so well with words.

I put the sensor outside around 11:40 am on June 11th. I haven’t had many of the new tags that can take measurements every X period and send them out in batches to save battery so I had the update settings wrong (I had it set to take measurements every 5 min for temp & light & to update every hour. But when you check the ‘batch update’ feature it automatically determines when to send out the data so the ‘overall update’ value should not be changed. Thus the data points are an hour apart until I set the update to every 5 min as well).

The data from both the wireless tag & the WeatherFlow can be accessed from the links below. Over the course of the first day or so everything seemed fine. The humidity values were within 1% of each other & those were slightly over what other PWS on WU were showing (the ones with gold badges).
However, this morning (the 12th, day after I placed the tag outside) I looked back through the data and noticed that very early in the day the humidity value reported by the Air is much much higher than the one reported by the WirelessTag. I know it didn’t rain last night & for those that are curious the sprinkler/irrigation system was not active at all either (it has been cool & there was a good amount of rain a couple days ago).

I haven’t looked at the amount of wind that we had, but based on this information it seems as though either the Air is sending bad data, the CL algo is having some issue (the Sensor has been in the shield for over a couple weeks now & the setup itself is not even 2 months old so there wasn’t much prior historic data), or something of that nature. I don’t know if it’s the sensor itself that is bad (note, the tape is on the upper inch of the sensor, the drainage hole is unblocked), the algo that isn’t adapting correctly or what, but it’s clear there seems to be some kind of issue here.

As for humidity caused by morning dew - I can see how a Gill-type shield would cause some humidity to get somewhat trapped, especially if there is zero airflow), but the tag is right next to the shield so I would expect it to be effected by any changes in humidity caused by morning dew.

Hopefully the WeatherFlow people can help me out with this (the humidity/dew point temp readings are quite important for calculating the amount of irrigation required so this is pretty important to me) whether it be updates to the SW/FW or a bad sensor. I will leave the tag where it is & will update this post with the time & date of when the tag is moved if rain is to be expected and when it is returned back to its original location. Hopefully this helps the developers out (and I can adjust it’s default location if that is requested & increase sample rate to less than 5 min).
I believe the link from WirelessTag should update with new information, but if it does not I would be happy to re-create the shared links & update the post.

Note - June 12, 7:30 AM:
WF Air RH = 79%
WirelessTag = 62% ~16100 lux
June 12, 1:00PM:
WF Air RH = 57%
WirelessTag = 48% ~7300 lux

WirelessTag Sensor Data (CSV format):
NEW LINK (apologies, had some issues with the tag manager while updating the frequency. Got it sorted out now, and the tags are running at a greatly reduced signal strength, but I lost some data in the process):

My WeatherFlow Station Link:

Location: Rye, NY. Fairly close to the sound, total elevation about 55ft.

1 Like

I’ll post the dates/times when I move the wireless tag below. Maybe in the future I’ll create some sort of enclosure that is transparent & has great airflow to keep the rain off of the sensor. Wireless tag sells a clear pouch for such purposes, but it’s fully sealed so no RH data is available with it. I will look around for parts with which I can make an acrylic canopy or something.
Based on data I’ve gotta today the RH delta is still about 10%.

WirelessTag Updates:
June 13th, 2019 - tag has been under the screened porch since it started re-recording

**** Below updates no longer apply to the new WirelessTag CSV Link****
June 13th, 2019 7:37AM - tag moved to screened porch due to imminent rain
June 13th, 2019 2:27PM - tag moved back to WF Air