The current outdoor temperature–measured live directly at the HaWo
Graph of the last 24h
Graph of the last 7 days
Graph of the last 30 days
Graph of the last 360 days
We use two sensors because unfortunately we couldn’t find a location that is guaranteed not to be affected by sunlight all year round at any time of the day. That’s why there is a pair of sensors (temperature + humidity) on the west side of the house, and one on the east side.
Technology has gone through a whole series of evolutionary stages. Generally speaking, if a sensor breaks down or can no longer be operated with reasonable effort, it is usually replaced by a new and better sensor.
For the measurement we currently use (as of November 2018) radio-connected SHT31 sensors powered by solar cells. The diagrams of the sensor on the east side can be found at https://gitlab. cs. fau. de/PoempelFox/hawotempdev2018/, while the sensor on the west side is a slightly modified Foxtemp2016 device that has been fitted together without large diagrams.
The SHT31 sensors deliver the temperature values with a resolution of less than 0. 01 degrees. The manufacturer promises an accuracy of ±0. 2°C for most of the temperature range relevant in Germany. The humidity values come with an accuracy of ±2% and a resolution of less than 0. 1%.
Access on the command line / by script
There is a small daemon that accumulates the values of the East and West sensors. To access it, simply type the following commands:
telnet ente. hawo. net 7337 (for temperature values)
telnet ente. hawo. net 7338 (for humidity)
Direct access to the sensors is currently not possible due to our involuntary network conversion.
Of course you can telnet also netcat o. e. use. The server spits out only the temperature or humidity and then terminates the connection again.
Previously used technology
Until 2013, different technology was used, the section from that time with the old blueprints is left here for documentation purposes – maybe someone wants to recreate it again (we don’t have to deliberately destroy useful information by de-publication like public broadcasting). At that time, the following was true:
For the measurement we now use two chips: both are DS1620 chips from Dallas/Maxim, one an old revision (allows more accurate reading), the other a new revision.
These allow the temperature to be determined with a resolution of 0. 01 or 0. 06°C, the manufacturer guarantees an accuracy of at least ±0. 5°C, but most of the ICs are of course much more accurate.
To read out the sensors we use a simple circuit electronics (designed according to the circuit diagrams on a page that unfortunately no longer exists for the programming language Euphoria) which is connected to the parallel port of a Linux-PC and was executed by us in RJ45 patch cabling. to be able to use.
For the first sensor, the circuit was still based on hole grids. The second version was made on an etched board – not because it was necessary, but simply because it looked better.
[Circuit diagram] [PCB layout, the assembly is printed on the bottom] [Connection of the connector for the parallel port]
By the way, the colors in the drawings refer to the connection cables of the network sockets, these are available under the order number “MEB 8-8” at Reichelt (as of May 06).
The PC holds the current temperature data via a small server on port 7337; from there it also reads the rrd script, which generates the above graphs.
This makes us (were) with great probability the first dormitory with its own temperature measurement accessible via IPv6.
(sponsored by Fox_Muld and Nico)