It is gratifying to see something you built work as designed, even though you may have tested it many times before. With day light saving time coming up last week, I took that to spur me to finish off the design of my WiFi web connected clock. During that time I changed the program to go out and check the time about 3 AM each day. That handles calibrating the clock crystal as well as the bi-annual time changes.
Among other things I also fixed a nagging issue with serial commands between the ARM Stamp and the ESP8266. I often saw that communication garbled, and I tried a number of things, though I am not sure which one actually cured it.
First added a bigger bulk tantalum cap on the 5V from the USB power coming in. And changed the wiring to larger wires and rerouted power to the ESP8266 first. During that may have also fixed a cold solder joint.
I also used an old trick of sticking my finger on the serial lines to see if that helped. It seemed to, but I may have been fooling myself on that one. Sometimes the added capacitance of a finger can cure a noisy line.
Also went through some examples of code at AdaFruit for the ESP8266. Which showed additional wait times between connecting to a WiFi router. And also a configuration command that was missing, that was causing some of the ESP8266 parts to seem to not connect.
My house has 2 WiFi routers at each end of the house, and changed the code to alternately try each router, and that seemed to help the clock sync up faster.
Now I have to run around the house and change all those other clocks. Maybe its time to abandon day light saving time. We here in California approved that in a non-binding initiative last year.
More details at the forum and earlier posts here in the blog.