4 ms·

An ESP32/ESP8266 with ESPHome's built-in Thermostat controller [0] and a relay (or solid state switch) board with optoisolation. There are a few other options

An ESP32/ESP8266 with ESPHome's built-in Thermostat controller [0] and a relay (or solid state switch) board with optoisolation. There are a few other options that are pre-made, like this [1], or that offer guidelines for DIYing it [2] [0]: https://esphome.io/components/climate/thermostat.html https://esphome.io/components/climate/thermostat.html [1]: https://joestump.github.io/hass/esp8266-thermostat/ https://joestump.github.io/hass/esp8266-thermostat/ [2]: https://youtu.be/WgDV-PlD3AE?si=74iQeFaVBgZcmTVp https://youtu.be/WgDV-PlD3AE?si=74iQeFaVBgZcmTVp

ssl-3 2y ago
The relay doesn't really need to be opto-isolated. It can be, and optos are cheap enough that many common pre-assembled relay boards include them, but it's not really necessary. If a person has enough stuff in front of them to make a relay reliably click on and off, then they also have enough stuff in front of them to make a conventional forced air furnace turn on and off just as a thermostat would.
- spdustin 2y ago
  I wouldn’t want the load current flowing through anything it shouldn’t. Maybe I’m too cautious, but it’s a binary rule that I’ve always followed when the load current is enough to cause harm to anything, or release the magic smoke inside the controller.
  - wwmiller3 2y ago
    To your point, I have had the high voltage side of a high-efficiency furnace (e.g., inducer fan) fail and take out the entire string of low voltage components (thermostat, humidifier controls) connected to the furnace through an unfortunate coupling on the primary control board.
  - ssl-3 2y ago
    I wouldn't want that, either. But a relay is already an isolator: There's no path between the switched contacts and the coil of the solenoid that switches those contacts. The coil doesn't know or care how much current or voltage the contacts are switching: It's just a coil that moves a little bit of metal using electromagnetism. And, sure: It takes more power to drive a relay's coil than it does to drive an LED inside of an optocoupler, but at some point that coil still needs to be driven by something that can accomplish that small amount of work. Adding an optocoupler just seems like adding extra parts to a circuit that would be simpler without. KISS?

ssl-3 2y ago

The relay doesn't really need to be opto-isolated. It can be, and optos are cheap enough that many common pre-assembled relay boards include them, but it's not really necessary. If a person has enough stuff in front of them to make a relay reliably click on and off, then they also have enough stuff in front of them to make a conventional forced air furnace turn on and off just as a thermostat would.

spdustin 2y ago
I wouldn’t want the load current flowing through anything it shouldn’t. Maybe I’m too cautious, but it’s a binary rule that I’ve always followed when the load current is enough to cause harm to anything, or release the magic smoke inside the controller.
- wwmiller3 2y ago
  To your point, I have had the high voltage side of a high-efficiency furnace (e.g., inducer fan) fail and take out the entire string of low voltage components (thermostat, humidifier controls) connected to the furnace through an unfortunate coupling on the primary control board.
- ssl-3 2y ago
  I wouldn't want that, either. But a relay is already an isolator: There's no path between the switched contacts and the coil of the solenoid that switches those contacts. The coil doesn't know or care how much current or voltage the contacts are switching: It's just a coil that moves a little bit of metal using electromagnetism. And, sure: It takes more power to drive a relay's coil than it does to drive an LED inside of an optocoupler, but at some point that coil still needs to be driven by something that can accomplish that small amount of work. Adding an optocoupler just seems like adding extra parts to a circuit that would be simpler without. KISS?