Recreating Unobtainium Weather Station Sensors

Imagine you own a weather station. Then imagine that after some years have passed, you’ve had to replace one of the sensors multiple times. Your new problem is that the sensor is no longer available. What does a hacker like [Luca] do? Build a custom solution, of course!

[Luca]’s work concerns the La Crosse WS-9257F-IT weather station, and the repeat failures of the TX44DTH-IT external sensor. Thankfully, [Luca] found that the weather station’s communication protocol had been thoroughly reverse-engineered by [Fred], among others. He then set about creating a bridge to take humidity and temperature data from Zigbee sensors hooked up to his Home Assistant hub, and send it to the La Crosse weather station. This was achieved with the aid of a SX1276 LoRa module on a TTGO LoRa board. Details are on GitHub for the curious.

Luca didn’t just work on the Home Assistant integration, though. A standalone sensor was also developed, based on the Xiao SAMD21 microcontroller board and a BME280 temperature, pressure, and humidity sensor. It too can integrate with the Lacrosse weather station, and proved useful for one of [Luca’s] friends who was in the same boat.

Ultimately, it sucks when a manufacturer no longer supports hardware that you love and use every day. However, the hacking community has a way of working around such trifling limitations. It’s something to be proud of—as the corporate world leaves hardware behind, the hackers pick up the slack!

19 thoughts on “Recreating Unobtainium Weather Station Sensors

    1. I’ve got 15 year old, bottom of the line Acu-Rite and LaCrosse stations, each with an external temp/humidity/barometer sensor. The Acu-Rite has never needed more than a battery replacement. The LaCrosse’s base station and (replaced) sensor have never managed to go for more than a week or so before losing the plot.

  1. La Crosse is around 70 miles from here.
    In the recent past, I would stop at La Crosse Scientific’s outlet store there.
    IOW, I have bought a couple of wx stations from them, as they seem to work for a few years and then quit. Knowing that their transmission protocol has been deciphered, I might be able to get those operational again.
    NB: I found my current wx station on Banggood, it was the same as one of LCS’s products, but without their branding, and cost about 25% less.

  2. I have a weather station, forget the brand. It worked good for a few years until one day it seems to have just stopped transmitting (or receiving, not sure which yet). It was just stuck on a temperature, wind speed, humidity and pressure for like a week.

  3. same same, I have a whole box full of dead weather stations. They can’t seem to make it past 2 years.

    When my current one dies (which it already did once – I’m on the second one) I’ll do this instead.

    1. I’ve ended up working on my own home grown solution using ESP8266s and tinkering with a variety of sensors. It’s a pretty standard hobbyist IOT setup using the Arduino framework for the ESP8266 and they communicate over WiFi to an MQTT broker. I had been playing with trying to create a solid batter powered device. Taking and reporting readings every minute I Was able to get about a month of run time from 4 AA batteries. But I tried some 18650s which got me to ‘good enough’ operating times so I stopped trying to optimize.

      I’ve been iterating as I get time. Realizing a bunch of the places I’m taking measurements from are indoors and near outlets, I’m working on something that is a suite of sensors including air quality and not focusing on power optimizations.

      I’d like to eventually migrate to ESP32s since they have some features that address some of the hacks I’m having to make for the ESP8266s and I’d like to make a dedicated base station even if what I end up with is a Pi Zero tied to a basic LCD that just shows a single dashboard page from Node-Red.

  4. Another dissatisfied LaCrosse customer here. My latest replacement I decided to buy a $35 no-name off Amazon. It’s at least as good as the LaCrosse whose sensor died. I just junked it…their sensors are ridiculously expensive, for the price of a replacement LaCrosse sensor, I could buy two of the Amazon no-names.

    1. LaCrosse are total garbage. TX141B temperature transmitter was inaccurate and worse at low temps. Warranty claim and they sent me another, but the PCB inside was much older. Dig out the scope and find the units have bad firmware – not enough delay for RC to settle. In the garbage that system went. So frustrating.

      Oregon Scientific was excellent but now they are a zombie company. Their temperature transmitter (IC) died after many years and you can’t buy a replacement. I loved that system.

      Now, I’ve given up and using ESP32/DS18B20 this year to see if I can monitor outdoor temps, requires mains power. Old systems were 1 or 2 AA lithium batteries and worked find at -40C.

      1. I’ve just been slapping Govee bluetooth temp and humidity sensors all over the place. I haven’t put the e-ink variant outdoors, but I find myself getting roughly 6 months to a year with cheap alkalines.

        They pair nicely with HA over bluetooth, and don’t need the Govee app to configure. Plus, I get battery data and they continue to transmit even when the voltage is too low to drive the LCD.

        1. I have an Ambient 2902C running for over four years… no trouble. Integrates with Home Assistant. Solar powered and I used backup lithium AA batteries… haven’t had to change them yet. Nice Display

  5. Temperature and cloud conditions are pretty easy to observe without a weather station. It’s the rainfall amount that can vary from mile to mile. It can rain 1 inch on the other side of town and I might only get .1 inch. So I only concentrate on collecting rainfall data by using a good quality “tipper bucket” gauge with an ESP32 and a website (or online cloud data collection service). Snowfall is a different story. Accurate snow fall sensors (ultrasonic for example) are expensive because they need to also be temperature compensated. So I just measure rainfall and use my bald head to determine the temperature and cloud cover. A very good quality rain gauge, stainless (or metal) tipper bucket with .01 resolution costs around $100. The ESP32 with a LIPO solar/charger and a plastic weatherproof box are the only other items needed. The ESP32 deep sleeps until the “tipper” interrupts it, wakes up and counts the tips for 5 minutes, reports the data and resets the count. Once it is quiet for 30 minutes it falls back asleep.

    1. This is the actual rainfall data I obtain for my home rain gauge, in Minnesota.
      https://www.catpin.com/rain/

      At the bottom is a link to show all raw data and a link to Google Charts that compiles the raw data into a comparison between days and years. September was a draught month this year. The ESP32 sends data to my website, processed by PHP and simple text files.

    2. I’ve always thought an infrared thermometer (you know the non contact ones) would be good at measuring cloud cover. Reason being if you point them at clouds they measure a lot warmer than the clear sky (presumably due to space being cold)

  6. I had that exact weather station in the top banner! 😀
    I loved it, it worked flawlessly for years but then the base station was what failed, not the sensor. However the sensor did live in my bike shed under cover.
    I miss it terribly, I got a replacement but it just wasn’t the same, it didn’t get the rugby time signal.

  7. Interesting thought, you could also use this to have a weather station display data on the weather anywhere in the world. Why you would want to do it I don’t know but it should be easy.

  8. I gave up on proprietary weather stations a long time ago and gradually built out my dream open-source alternative using ESP8266s and a LAMP-stack backend to plot the data. Then I added a way for my ESP8266s to poll to see which of their pins to switch to a different state to control things. This allowed me to create an all-inclusive front-to-back IoT and environment monitoring solution. I’m in the process of added infrared diode support to my ESP8266s so they can turn on proprietary equipment like minisplits using copies of their controller data streams. https://github.com/judasgutenberg/Esp8266_RemoteControl

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.