Question about 3D printed case and BME280 reliability

[ollybeck]

First off thank you for writing this guide! I made one of these per your instructions and it works beautifully. I'm a newb to ESPHome and this guide really connected the dots for me.

The ESP and mmWave sensor seem to give off too much heat for the BME280 temp/humidity readings to be useful, which is a shame. I wonder if a different placement on the board or some kind of heat shielding could help with that.

Anyways, I wanted to ask if you had any pictures of the 3D printed case that you provided in this repo? I'm curious what it looks like in real life!

Cheers :)

P.S. Did you ever test out the new tweaks from crlogic for your device? Looks pretty cool, I just can't wrap my head around how to patch it on to this project.

[Schwayla]

Hi Olly, thank you so much for the feedback! I'll see what I can do to redesign the 3D-printed case, I ran into some issues with it before and forgot to remove it from the site.

I'll do some testing on the sensors to try to figure out the heat/humidity readings and what I can do about that.

I have looked into the CRLogic tweaks and have wanted to see how to implement them into the project, the good news is the TX and RX wires connected to the DFRobot Radar should allow for the addition of the features CRLogic pointed out. I'll see what I can do to implement but will most likely be a little down the line.

[ollybeck]

I'll look forward to the new 3D-printed case whenever you get a chance to finish it! Then I can finally start putting these awesome sensors around the house :)

For temp sensors on other devices I've been testing out 90º header pins to some success, just still needs to have some distance between it and the warmer components.

By the way, I added this tweak to your code which deactivates the SEN0395 LED:

button:
# Turn off mmWave LED
  - platform: template
    name: "mmWave_LED_toggle"
    id: turn_off_led
    on_press:
      - uart.write: "sensorStop"
      - delay: 1s
      - uart.write: "setLedMode 1 1"
      - delay: 1s
      - delay: 1s
      - uart.write: "setUartOutput 1 0"
      - delay: 1s
      - uart.write: "setUartOutput 2 1 1 2"
      - delay: 1s
      - uart.write: "saveConfig"
      - delay: 3s
      - uart.write: "sensorStart"
      - delay: 6s # the sum of mmwave_sensor delays   

[Schwayla]

Love this! I was wondering if it could be done, saw CRLogic did it so I'm glad you were able to figure it out!

How are the 90-degree pins working? I've been trying to design a case to mitigate the heat issue however it seems to be making it hotter. I'm thinking of just adding an area where you can + or - units to the sensor readings to tweak the output.

Here is the case so far

Ignore the orange, it's all I had.

[ollybeck]

I haven't tried the 90º pins on this sensor yet, only on other projects w/ different positioning on the board. When I get a chance to test it I'll share here.

The model looks great!! I bet that separator there helps ease the temp interference a bit. One thing I read about someone having similar issues w/ the BME280 on a PCB is that much of the heat is coming from PCB itself through the pins, which appears to be true. Right now at 22ºC room temp, using an infrared thermometer, the SEN0395 is 34º, the D1 Mini is 33º, the BME280 is 31º, and the PCB is 29º. I saw one recommend thinner header pins or wires to minimize heat conductivity.

For tweaking the temp with an offset I plotted readings from other accurate sensors I have and noticed it's never quite consistent. Sometimes the board is hotter and sometimes cooler, not sure why. When I get a chance to test again I'll share here and show you what I mean.

Would you mind sharing the STL? Would love to test this out in that pretty new case!

[Schwayla]

New case STLs posted