We have decided to start the production of 50 devices for more thorough testing. A few wattmon boxes are already running in the field and I am actively monitoring and fine tuning the firmware and web interface but so far the feedback has been positive.
Installing The main issue so far during installation was getting Internet connectivity to the battery room. This proved to be challenging using wifi even though I have a dongle specifically designed to bridge Ethernet and wifi. In the end I had to run Ethernet cable from the main adsl router.
Connecting up the wattmon current sensors is easy - but requires everything to be switched off beforehand as the battery wires need to be pulled through the hall sensor. Power is taken for the battery terminals directly allowing for the voltage to be sensed as well. I am thinking of adding an external sense input for high voltage sensing.
So far the major problem has been erratic behaviour on the login, session handling had some major bug that made it seem to behave very unpredictably. I squashed a bunch of bugs and traced the problem finally to http headers. Large headers with lots of accept tags would cause the http server to miss the cookie tag and hence the session would remain uninitialised. In firmware 1.39 this issue has been solved. There was also a problem with the way post requests were parsed. I have now tested it with over 50 parameters and it works smoothly, which should be sufficient for most embedded developers.