WiFi Temperature & Humidity Sensors

[robputt796]

Evening All,

Just wanted to show off some little WiFi Temperature and Humidity sensors I have been making as part of my DIY Hive / Nest equivalent. The idea of these are to act as room sensors and then plan on having WiFi enabled radiator valves. Currently I only have the sensors up and running, they connect to my home automation server and post their readings periodically and the results are graphed.

Now for the exciting bit, the cost price of these bad boys...

- ESP8266-1 - £3.70
- PCB - £2
- DHT22 Sensor - £4.50
- USB Micro Breakout - £0.26
- Capacitors - £0.20
- Voltage Regulator - £0.40

Total cost to build: £11.06

Let me know if anyone is interested in further details about how to make these. The PCB is optional and the previous version of the sensor was built on prototyping / veroboard.

Regards,

Rob

 

[MeldrewsMate]

Fantastic.
It looks like you're trending RELATIVE humidity.

How do you connect these items to your HA server? What is their power source (batteries/mains adapter/CAT5?)
What is your home server made of, and what can it do? What do you propose for it to do in future?

MM

 

[robputt796]

Morning,

- Yep it is relative humidity, the sensor comes pre calibrated and seems reasonably accurate, the spec sheet seems to think ~1% I have no way to validate it but all of these sensors left in a box with some silica gel all seem to concur with each other so I would say thats good enough for "finger in the air rough estimate" home use.
- Power is supplied by the USB micro port, it isn't used for programming the device (that is done with one of these... https://www.amazon.co.uk/Pinzhi-ESP8266-mod-prog-Programmer-Uploader-Black/dp/B078MC52NB), I power mine from either a spare phone charger and cable or I have a few sockets around the house with built in switching USB power supplies on the faceplate. You could in theory run them from a USB battery bank, but the circuit design does not allow for a super low power mode so I expect you'd only get a couple days use out of 1 charge on a small power bank. (maybe we can change this in a v2 of the board if people have a battery powered use case)
- My home server is complex, it's a Dell R210ii (a rack mount server) and it runs ESXi. On top of ESXi it runs my home router, Unifi for my WiFi access points, Shinobi for my CCTV, Emby for streaming my media to my smart TVs and my home automation server which is running Influx DB, Grafana and some custom Python stuff. Yes this is totally overkill and not the norm, so please ignore it. In reality your home automation server for something as basic as a bunch of temperature sensors could just be a Raspberry Pi.
- The home automation server in my case receives HTTP requests from the sensor and stores the results in InfluxDB via a simple Python Flask app, the firmware could be edited for MQTT, HTTP, HTTPS etc... it depends on how you want to set it up really. If you have no clue what any of this means thats fine, I did consider doing a kit version of this if there is enough interest which required 0 configuration out of the box other than editing your WiFi settings and then going to a website to view your sensors. Of course this would need to pay me a little bit of labour for pre-flashing the ESP8266s and taking the bag of bits to the post office.
- Whilst all I am doing at the moment is graphing my relative humidity and temperature as I said the end goal is to make a WiFi radiator valve (probably by cannibalising an existing valve and hacking on a stepper motor to ensure the valve end is still very much out of the box to prevent leaks) so I can replicate my own Hive setup with multi room control when the heating is running.

Regards,

Rob

 

[SpecialK]

Just ordered some of these DHT22 sensors to connect to my tasmota flashed sonoff basic .... and hopefully feedback to HA.
Any ideas how to make them slightly more attractive to mount on the ceiling?

 

[SimonH2]

Any ideas how to make them slightly more attractive to mount on the ceiling?

Put them in an enclosure. Where I used to work, we had some (bought many years ago) where they'd mounted 1-Wire sensors inside a surface mounted phone socket (dual RJ11 & beige box) with a few holes drilled in the box to allow airflow. You could do much the same with (say) a blank faceplate and drill a few holes in it, or perhaps cut a hole in it and cover with a permeable material (perhaps even wall/ceiling paper to make it disappear).

For radiator valve heads, perhaps take a look at the OpenTRV/Radbot project where they have an open design for the hardware - rather than re-invent this bit of the wheel

Personally, I'm leaning towards starting to put in some Cat5e cables in to every radiator and a backbox for a wall mounted "stat". That way everything could be PoE powered and run over ethernet - no mass battery changing every so often and no power budget to worry about. For example, the OpenTRV project doesn't yet support 2 way communications, partly due to the power required to run the receiver and process the packets. if nothing else, I could simply provide power down the cable and do away with the batteries.

Whether I'll ever have the funds or the drive to get such a project working is another matter.

 

[SpecialK]

I'm thinking of having temp sensors in all rooms - Run via ceiling rose power and poking out the ceiling away from the lamp. These will then call the boiler until the temp is at the required level. I think the TRV's I have work fairly accurately at the moment - its just when the stat in the hallway gets warm enough it kills the boiler and some rooms are left cold! Electronic TRV's sound like a nightmare waiting to happen?

 

[SimonH2]

Hmm, ceiling mounted temperature sensors will leave you cold - think where the warm air goes to in a room Where sensors are remote from the fluid control elements, then they tend to be around half way up a wall.
The benefits of electronic TRV heads are (IMO) :

• Potentially closer control of room temp since they can have an integral element to the control. Standard heads are essentially proportional only (with a hint of differential due to their thermal mass) - so will always have a different offset between setpoint and actual room temp depending on power required. Eg, if there is (say) 5˚ between fully closed and wide open, then in warm weather the room would need to be 5˚ warmer (to fully shut off the heating) in warm weather than it would need to be in cold weather (needing lots of heat).
• Remote control and/or programmable profiles and/or remote sensing. The simplest variant of this means you can program each room for a different temperature profile.
• Integration with the boiler control - so the boiler runs when a valve says it needs heat and shuts down otherwise. Avoids the "one warm room shuts down the system" problem you've mentioned.
• Some have a periodic valve cycle function. This simply winds the actuator open and closed again ever so often. In summer, valves can stick because they've been held shut for ages - so in autumn you find a radiator that won't turn on.

As you hint at, it's not all positive.

• All the ones I know of are battery powered. This means routine battery changes, and restricts some of the features to control power used. Some systems give you warnings when the batteries are getting low.
• They can be in some "rather awkward" places when it comes to programming those that need local programming.
• All the ones I know of are wireless - some via "WiFi", some via Bluetooth, some via proprietary systems, ...
• Most are "rather expensive"