Best Zigbee Temperature and Humidity Sensors with Display

25 thoughts on “Best Zigbee Temperature and Humidity Sensors with Display”

1. Great article! Thank you for comparing so many sensors.

   ___
   I would change this part:
   […]
   Time (AM/PM, 24H), Date (YY-MM-DD), Day (Mon-Sun), Built-in Alarm Clock, Battery indicator, Connectivity indicator

   to:
   […]
   Time (AM/PM, 24H), Date (YYYY-MM-DD)[…]

   because, the device shows the date as 2023-07-08 so with 4 digits for the year (YYYY) instead of 2 (YY)

   • Thanks, edited!

2. Fine description, but a lot of the senmsors does not work in a fridge and freezer. specially those with small battery as CR2032 etc.
   For that purpuce a device from INKBRID or SWITCHBOX was tested and work fine, they are BLE and for that a M5Stack ESP32 devicxe was used as ESPhome BLE proxy and it works fine.

   • Thanks for the suggestion.
     It was never my intention to test sensors in a fridge.

3. Hi, great article.
   I am looking for this kind of devices with buttons to use them as a thermostat. Use case is : walk in a room, see that it’s 19°, press a button to ask for 20 without having to get your phone, launch HA, go to correct tab to adjust thermostat. (Climate is handling by HA that will start/stop heating).
   Do you know if such a device exist (cheap ZigBee one) ? (like a Nest thermostat but dumb, the smart part will be handled by HA).
   Thank you.

   • In my experience, Zigbee thermostats all come with built in temperature sensors so you do not need an external one.
     Here’s one example from Moes: https://smarthomescene.com/reviews/moes-zigbee-smart-thermostat-bht-002/

4. Nice writeup!
   But unfortunately, none of these seem to be able to attach an outdoor unit, and be able to read and display 2 temperatures at once.
   Have any suggestions for models like that?
   Thanks

   • No, you are going to have to go the DIY route for that and use Dallas probes.
     https://smarthomescene.com/diy/diy-waterproof-double-probe-temperature-sensor-with-an-esp32/

5. HI! Great work! Anyone of this temperature/humidity sensor are compatible with Philips hue hub? Thanks!

   • No, they aren’t.

6. One question. How did you know that your reference device Xiaomi LYWSD02 was showing the correct temperature and humodity?

   On my desk there are following devices:

   – ThermoPro TX2 (433Mhz)
   – ThermoPro TP357 (BLE)
   – Aqara Xiaomi WSDCGQ11LM (Zigbee)
   – Ecobee Smart Sensor (AFAIK, 433Mhz)

   All of them show difference 0.5-2.5C from each other. Humidity difference is 10-25%.

   Go figure which one is correct…

   • I calibrate the LYWSD02 against an industrial thermometer used in the meat industry from time to time.
     These are calibrated once every 3 months, a friend has a meat store 🙂

     Cheers

     • Do you know where to get that thermometer for cheap? 🙂

       Jokes aside, 1C difference will have serious financial implication if sensor based threshold is used to start heating/aircon.

       In my case, it takes 1h burning of propane for a home furance to rise temp by 2C. If the sensor is off by 2C and every day it burns 1h longer than needed to be, at the end of the month I am spending $50 more on propane or $250 more for the season…

7. Are you able to confirm if any of the displays can change to °F? All of them are listed on the chart as using both scales, but I have not been able to convert the displays to °F, specifically with the Tuya ZL02-ZX.

   Overall, great job with the standardized product comparison.

8. Hi, and thank you for this great write up and set of tests.
   I am wondering if you also at any point tested zigbee network range for these units? I have two Tuyos and they connect and report back correctly only within eight feet of my zigbee device. For me, the range would be a critical consideration given that a user might want to put these devices in a basement, or a garage. They are *most* useful in inaccessible areas, and those are the areas where the range becomes critical.

   Do you have plans to do another survey and if so, would you consider including range as part of the testing?

   Thank you again.

   • If you add enough network routers (smart plugs, presence sensors, switches, mains connected devices) range will not be an issue.
     Zigbee traffic can be relayed at long distances this way.

   • Tuya has various RF modules. They have the same power, but different antennas. Some are better, others are worse. It is important to know what module is used in the sensor and a photo showing whether the module is installed correctly to avoid interference from metallic surfaces.

9. Bought the Tuya SZ-T04 from the link included in this post. It looks exacly the same but in z2mqtt it identifies as the TH01Z. After opening it, it has a Sensylink sensylink CHT8310 temperature/humidity chip in it.

10. Thanks for this extensive review! Staring with the ambition for devices with screen made me realize I am looking for a Zigbee temp and humidity densor, but without a display.

    Are you planning to do some testing for small ‘invisible’ devices I scan place in my rooms? Or do you have a tip for a reliable and accurate affordable Zigbee device?

    Many thanks!

    • Hey Mark,

      Even without doing research on devices without screens, I know the Aqara Temp & Hum devices will come out on top:
      https://s.click.aliexpress.com/e/_DEiBK7r

11. The test did not include information about the sensor model Xiaomi has. We don’t know its accuracy. Saying that it is calibrated means nothing, because all sensors are factory calibrated. The reference sensor should be a class more accurate.

    The specification table shows different accuracies for AHT20 and SHT40 sensors, why? For example, AHT20 sometimes has ±0.3°C and other times ±1°C. This appears to be a copy of the Chinese specification and not the actual sensor specification.

    The most important data from the test is the sensor model, reporting interval, and the rest is useless in my opinion.

    • Hey Lukasz,

      First off, I calibrate the LYWSD02 against an industrial thermometer used in the meat industry from time to time.
      These are calibrated once every 3 months by a governing body. A friend has a meat store, and I just use that for reference.
      It’s 99.9% correct.

      Second, the accuracy for the devices is different because it doesn’t correspond to the sensors manufacturers specification for a couple of reasons:
      First, the precision step has to be coded in by the device manufacturer according to its factory accuracy. Even if the sensor is specced for ±0.1°C accuracy, most brands round it out to either ±0.2°C or ±0.3°C (with exceptions).
      Hence a slightly different reading on the same sensor.

      Second, different devices generate a different amount of heat which needs to be accounted for and calibrated against. This again falls as a task to the device manufacturer.

      Third, I have no doubt Chinese manufacturers skip over these calibration steps, which is why I test them myself.

      If you didn’t find anything useful in this article, not a problem, just carry on.

      Cheers

      • Thank you for your work and sharing information about smart products.

        The methodology used to evaluate the sensors is incorrect.

        First of all, the measurement result is not one value. The result is a range of values determined by the accuracy of the instrument. Even if you calibrated the LYWSD02 with a sensor that is a class more accurate, the LYWSD02 still has an accuracy of ±X.XX°C. This sensor does not suddenly become a class better, it still makes the same error.

        Sensors, e.g. AHT20, are calibrated by the factory. The manufacturer of the smart sensor is not able to read the values with greater or lesser accuracy (I checked the datasheet). He will always get the result with accuracy ±0.3°C.

        The values given on the packaging are random and are probably copied from other products or the manufacturer uses the same packaging for different versions of the sensor. Overall, they are worthless.

        You’re right about generating extra heat. However, I don’t think a device powered by a coin cell battery would waste energy on heat. The size and location of the holes in the housing through which air flows will have a greater impact on the measurement result.

        As I mentioned, sharing information about the sensors used and how they are placed on the PCB is invaluable and I thank you very much for that.

        • Hey Lukasz,

          Thanks for prompting this interesting discussion 🙂

          I feel like we are in agreement fundamentally and talking about the same thing, however there’s a miscommunication somewhere in here.

          First, the LYWSD02 will keep it’s accuracy (which is 0.1°C), no matter what. This is not the point.
          The point is, the device is calibrated against an industrial grade T&H sensor which makes the base reading of the LYWSD02 correct.
          Whether or not it creates errors going up and down is impossible to verify, since I’m not a test lab, and naturally I cannot guarantee this kind of precision.
          I believe this is good enough for this use case (Smart Home), however if you find it critically flawed, please don’t rely on it.

          Second, the AHT20 will always measure with accuracy of ±0.3°C and the device will get that reading. What it decided to do with it (e.g. round to 0.5°C) is up to the manufacturer of the device.
          This is the case with the ZTH08. It will display measurements at 0.1°C increments, while the sensor has an accuracy of 0.3°C? Makes no sense whatsoever.
          You are right about the 1°C degree stat though, it’s a mistake and I’ve corrected it. The device has other fatal flaws too, which is why I recommend skipping it completely.

          After all, these are Chinese devices that cost a few bucks and it’s unreasonable to think you will be getting a completely perfect and capable temperature and humidity sensor.
          Other things are also considered here, like display quality, time/date, additional features, size, backlight, Zigbee performance that make the overall rank of the device.
          My effort is geared towards eliminating the total trash from the solid devices – at least in my own, personal tests.

          I’m glad at least some of it was helpful.

          Cheers

          • I agree, you provide a lot of information about devices and we can draw our own conclusions from it. I am also looking for a very cheap and at the same time very good sensor.

            Regarding numbers and accuracy.

            AHT20 gives measurement results with a resolution of 0.01°C and an error of ±0.3°C. For example, the device may receive two results: 22.34°C and 22.36°C. If the device has a display with a resolution of 0.1°C, it will display these results as 22.3°C and 22.4°C. The difference in the measurements was 0.02°C and in the presentation of the result itself there are differences of 0.1°C. In addition, there is measurement accuracy. It does not affect the way the result is displayed. Measurement accuracy is used to interpret the measurement result. This means that the real temperature is somewhere in the range of 22.04°C – 22.64°C. I would draw thick lines on the charts and as long as the lines from different sensors overlap at least a little, I would assume that these sensors are of the same quality.

            Of course, the manufacturer can use software to worsen the measurement result and, for example, round the result in a different way.

Comments are closed.