Here’s the humidity sensor the NWS uses. (they call it a temp/dewpoint sensor) The film capacitor sensor is inside a fine mesh screen, the unit itself is housed under a protective cover, but is otherwise exposed to the elements. I’m thinking if the wire mesh screen were available, it might make a good sub for the PTFE tape.
Here’s what the cover/enclosure looks like:
The probe is mounted verticaly under the “mushroom cap.” (outlined in red)
The white box to the left houses the sensor electronics.
this site has all the details. I tried it myself, plugged 2 brand new DHT22’s and an older DHT11 into a breadboard with a DIY emonTH node and monitored it for a couple of months. Mine were 15-20% out on average. The 11 was even worse.
Complete waste of time. There is a routine to attempt to calibrate them, but it made it it even worse (though that may have been my fault with condensing moisture).
I’ve now got a Bosch BME280 in to replace them, but not got around to fitting it yet…
In order to fit the emonTH board into the enclosure, it was necessary make some modifications. The onboard AA cell holder was removed to lower the profile and replaced with two single AA cell holders, secured either side of the board. In order to fit the PCB, existing standoffs in the case were cut off and a small groove cut which allows the board to fit snugly and be retained without fixings. The hole for the gland was drilled and then tapped with an M8 tap - the plastic is strong enough to nip up the gland and compress the seal without stripping the threads providing you are careful. Try as I might (even with a vacuum desoldering station) I could not remove the turned pin socket for the DHT22, so simply soldered a piece of 0.1" pin header to the DHT44 cable once it had been passed through the gland, and plugged it in. It’s all a bit of a squeeze, but the lid does go on
I also have the Ambient Weather Solar Shield as mentioned above, Ambient Weather SRS100LX Temperature and Humidity Solar Radiation Shield. After trimming off unused mountings on one of the plates, it is possible to fit the entire enclosure between the two bottom plates of the solar shield - you might be just able to make it out in the photo below. The DHT44 sensor is secured to a rod which runs up the centre of the solar shield.
Wow that’s quite a nice job you did. Am also keen to know how it behaves. Do you have a nearby weather station to compare (Metar or a Davis VP2 for ex ?)
As the weather geeks will tell you, it is very easy to have bad readings …
Bill - You never know how these things will turn out until the hardware arrives, but it’s a nice little enclosure and very solid once the lid is on. I’m not sure I would trust it at the bottom of a swimming pool like one of the previous posters, but that’s more about the threaded gland rather than the enclosure itself.
Eric - I never intended to compete with the ‘weather geeks’ just build something to survive the northern weather. There is a Davis VP2 a few miles from me, so should be able to get an idea of the accuracy. I am very limited as to placement of the solar shield, so this is likely to be the main factor.
I wrapped the sensor with thread seal tape called PTFE (teflon tape). So far so good since I half expected to destroy the sensor with the high humidity.
I had a few Do’h moments. My backyard has lots of bushes and flowers and trees. Not the best environment to get good readings. From what I have read since the sensor should be far away from plants, grass, & trees. And it should be mounted two meters above the ground. Mine was mounted on a 4 foot (1.2 meter) fence post in our garden - not good!
Here is a comparison between the emonTH in my backyard and the local airport readings. The temperature readings are fairly good (blue is the emonTH & yellow is the local airport). The humid readings are not so good (green is the emonTH and red is the airport).
www.aviationweather.gov
Click and drag the map to the area you want the ob data for, then click the METAR button on the toolbar near the top of the screen. Select the ob data you want to see by checking / unchecking the boxes just below the map. Zoom controls are in the upper left corner of the map.
http://mesowest.utah.edu/cgi-bin/droman/mesomap.cgi?state=IL&rawsflag=3
Hovering on a station - the black dots - pops up the ob data. Clicking and draging works the same as on the Aviaton Weather page. Zoom controls are in the lower right corner of the map. Ob data displayed on the map, and the source of that data, can be selected from the pull-down menus on the left side of the screen.
OEM node in yellow and blue, Wunderground data in red and green. Overall it compares pretty well, certainly close enough for me given the constrained mounting location.
Bill - forgot to tell you “thank you” for the two links. The mesowest link was the one that had the best info. I asked for the local NWS site because the data I get from Weather Underground (WU) is different. I started staring at the data when comparing WU data to my outside emonTH and noticed something odd. The WU temperature always increments/decrements by 1.8 degrees. It almost seems like WU is grabbing the NN:52 (8 minutes before the hour) METAR data and converting the temp from ºC to ºF. Not quite what I wanted!
Thanks Glyn Good tip there from @johncantor. I did chuck in a silica gel bag before final assembly to try and keep any moisture at bay, but some of that semi-setting potting compound is probably a good idea long term.
Whichever enclosure you use, do remember to drill a drain hole to the underside of the box. Unless hermetically sealed, condensation will surely accumulate.