REST was a bad idea, It was not low power enough, and after playing with MQTT, i made the obvious choice.

New project: https://github.com/leon-v/MQTT-ESP8266-nRF24L01-Solar-Battery-I-O-Network

Low power real time ESP8266 with REST interface.

THIS PROJEST IS STILL COMING OUT OF PROOF-OF-CONCEPT AND IS NOT STABLE

I will be keeping an eye on this thread on EEVBLOG

The Environment:

So you understand this build environment, here is its description.

I am using Windows with the Linux bash shell. crosstoon-NG is installed in /opt/Espresif. There is a script in /src/misc/getcompiler.sh to install dependencies and install required tools.

I am editing with Sublime text 3 and have a build script to simply execute gen_misc.sh in the bash shell.

You may well need to edit gen_misc.sh to suit your device. I have an ESP12E using the generic FTDI RS232 to USB module.

There is a test REST script ESP_Test.html is an example of how to make a REST AJAX request from a browser. This is the script i use while doing the power analysis to give it some real world conditions.

The Software:

This firmware will eventually be able to take and send REST requests when events are triggered. So far i have only thought about hot the ADC will be configured.

It will respond to request to /adc/0 thought to /adc/7
The ADC will be attached to an 74HC4052 which will be used to switch between channels.
Each channel will have:
- A Maximum - Will not fire events over this value, needs updating by remote host to prevent getting re-triggered
- A Minimum - Will not fire events under this value, needs updating by remote host to prevent getting re-triggered
- An Interval - If an event is not fired within this time, an event will be triggered
- Endpoint - The URL to the endpoint that this ADC will make a request to when firing an event.

I might have some modes where there is no 74HC4052, maybe /adc/gpio or something. The 74HC4052 will run from a power circuit that will be switch on and off, and I will be trying to charge a large capacitor, and disconnecting the power source at a predictable time before taking a reading so there is a stable / low noise power source to the analog circuits. The ESP8266 can make a hell of a lot of noise with its radio.

The Hardware.

I am using an ESP12E soldered onto a DIP breakout which has requisite pull down/ups.

The ESP8266 will go into Auto Light sleep, and needs to be woken via an external interrupt. The interrupt is triggered by a discharging capacitor on GPIO14. After a lot of testing with various ways to trigger the wake, including using my scope to test many different signals. The best results by far were by using a capacitor on GPIO14. You can't just put the cap directly on GPIO14, there must be a resistor in between.