I’ve been using RTL_433 with a SDR for a few years now with modest success. This setup will detect and identify a variety of 433Mhz sensor readings. The output is piped to a MQTT broker and from there the readings are presented graphically.<\/p>\n
At first this was done on a Raspberry Pi running a bash script on startup, with the following one liner command.<\/p>\n
rtl_433 -F json | mosquitto_pub -h <MQTT-Broker> -i RTL_433 -l -t RTL_433\/SDR_FEED<\/pre>\nUnpredictably, problems would arise when the process would hang and readings no longer were being received. This was addressed by placing the process under a supervisor task, details can be found in this post titled “Scheduled vs Supervised Tasks” https:\/\/www.cloudacm.com\/?p=3956<\/a>. Another issue discovered was that some sensors failed to be detected by the SDR. The standard antenna included with the SDR was replaced with a 433 Mhz tuned antenna, https:\/\/www.amazon.com\/Zerone-433MHZ-Antenna-Signal-Amplifier\/dp\/B07DMXSX45<\/a>. This improved coverage, but there were still some PIR sensors that were in dead spots. It was becoming clearer that the SDR had a hard limit and to get these dead spot sensors online, I would have to place some kind of bridging device closer to them.<\/p>\n
I had seen demonstrations of 433 Mhz RF modules being used with microcontrollers in this post, https:\/\/randomnerdtutorials.com\/decode-and-send-433-mhz-rf-signals-with-arduino\/<\/a>. However, many of the demonstrations I found online used the RF modules as a wireless link between two microcontrollers. I had a Sonoff RF bridge available and looked into using it for the purpose of receiving and processing sensors. However, that would require hardware modifications and following a rather lengthy setup process with a narrow range of supported sensors.<\/p>\n
Along the way there were GitHub repos that caught my attention, https:\/\/github.com\/sui77\/rc-switch\/<\/a> and https:\/\/github.com\/ninjablocks\/433Utils<\/a>. It was this blog by Ray Wang that spelled out clearly what I was intending to do, https:\/\/rayshobby.net\/reverse-engineer-wireless-temperature-humidity-rain-sensors-part-1\/<\/a>. This was further elaborated by Brad Hunting’s GitHub repo, https:\/\/github.com\/bhunting\/Acurite_00592TX_sniffer<\/a>. From here out the pieces started to fall into place as I found more details on Brad’s blog, https:\/\/www.techspin.info\/<\/a>.<\/p>\n
One of the key steps in getting the microcontroller to detect and decode the RF sensors was identifying the pattern. I used a logic analyzer to get the timings and sequences. From there I was able to display the sensor data with the microcontroller. This worked for my PIR motion sensors, temp\/humidity sensors, door reed switch sensors, and key fob controllers.<\/p>\n
I had some problems porting the code to work on the Wemos ESP8266. The code worked fine on the Arduino Uno, but the interrupts are handled differently on the ESP8266 microcontroller. The function called by attachInterrupt had to have ICACHE_RAM_ATTR defined. This video briefly explains some of the interrupt dependencies of the platform.<\/p>\n