With clock firmware 46 and WiFi module v1.0.4 I have noticed that the connection between the two is lost after a couple of days. Power off and on reconnects it.
I can connect to the module and see the web GUI but can't make any changes.
The line "Found I2C slave at 104" is gone.
The clock then runs of the 16 MHz crystal.
Any recent code changes that may cause this?
I really like the addition of showing the IP adress of the module. Very helpful.
I have the "classic" version.
The module is stacked directly on the bord via the level shifter with no wires.
Nothing is changed except the firmware (Atmega and ESP8266).
This never happened with previous versions so I thought it has to do with firmware.
I'll keep an eye on it.
Don't know yet if it's gone after a specific time, a little hard to tell, can't log it.
It's not very obvious when it happens, only after some time the clock starts to drift.
Last time it was between 1 and 17 hours since start. (I checked the uptime but don't know when it got disconnected).
I could try "downgrading" the firmware for the board and module, one at a time to see if it persists.
I'll first try reverting from feb 28 to feb 21 (main board).
You mean the logic level shifter board?
I can't find any faulty components on it. But to be sure I've built another adapter to rule that out. Testing it now.
If you mean the ESP8266 module; I have tried a couple of them, no difference.
Thanks for the offer though!
My observations so far.
I've changed the logic level shifter - no change.
Replaced ATMEGA chip and tried different firmwares - no change.
Checked continuity on PCB for I2C connection - nothing wrong.
Factory reset after all changes.
Still the module loses connection after some time (hours or more than a day).
Flashed firmware from 1.0.4 to 1.0.3 - The clock has since worked (for some days - can't say for sure yet if this really is the problem).
So are you sure nothing in that firmware is the problem?
It could very well be something else, beyond my understanding, just my observations so far.
I've tried those changes on two different boards, on both of them WiFi version 1.0.4 fails within 24 hours while 1.0.3 works fine.
Of course there can be other reasons but this is as close as I can get in narrow it down.
I wonder if I need to put a "watchdog" in the ESP code to confirm that the connection is good. At the moment it is fire and forget, and if there is a problem in the communication, we won't know that. We will just see the clock drift, and becuse the crystal oscillator can be good or bad, it might be a long time before it becomes obvious.
If I add a "ping" to the ESP, we could indicate that there is trouble with the communication.