I recently installed an emonTx and am using a Raspberry Pi B+ for the base station. The setup went pretty well using the pre-made SD card image and everything was pretty easy to set up. I'm using the AC-AC adapter with the emonTx.
We have an induction cooker-top. For some of the lower heat settings the cooker seems to pulse the ring on and off over a period of around 1-3 seconds. This shows up as a high power usage occasionally in the graph, but since the emonTx is configured to collect data every 10 seconds (I think?) I wonder whether the power data collected is inaccurate, because it might hit or miss the pulse (and worse, the timing might 'beat' with the pulsing of the cooker so that it mostly misses or mostly sees the power from the cooker).
I wondered if there is already any mitigation for this in emonTx or if others have noticed this problem? Presumably there are other devices that would have a similar behaviour.
Obviously I could increase the frequency that the input is checked, but I also don't want to fill the emon database with many values that are the same most of the time (i.e. when the cooker isn't used).
I was thinking of modifying the emonTx sketch to have two configurable parameters: the sampling period and the transmission period. The sampling period would always be <= the transmission period. A reading would be taken from the CT input every sampling period. Each time a reading is taken, we'd increment a counter and accumulate the reading in a variable, and then compute the average of the CT inputs when we reach the transmission period, send them, and reset them to zero to start the next transmission period. This would allow a much higher sampling frequency (e.g. 10 per second or more) to get more accurate data, but keep the data storage requirements and write counts the same. We could potentially make the configuration dependent on whether the AC adapter is present, in order not to ruin the behaviour when running in the lower-power scenario.
Any thoughts welcome.
Re: Energy monitoring in home with induction cooker
Check http://openenergymonitor.org/emon/node/2406
Re: Energy monitoring in home with induction cooker
Or any of the "continuous" sketches on Github.
You are by no means the first to come across this. The problem, as you rightly say, is that the original emonTx "discrete sample" sketches assume that the load is relatively constant, therefore they report the instantaneous power (or very nearly). The continuous sketches actually total joules and divide by the time to report current and power, hence average the consumption over the reporting period.
Obviously, the continuous sketches (including Chaveiro's) are not suitable for battery operation.