The Monitoring Shields look like a good & economical solution.
It appears this is a new product offering – maybe not a new development but the ‘guts’ of the old EmonPi unit?? – so a proven component. And no doubt there’ll be a subsequent add to this forum thread which will clarify.
You will have noted the Optical sensor vs CT discussion above. The Monitoring Shield accepts an Optical sensor. Using this at the PV end would cost £19 and save the £21 for a CT and AC/AC voltage sensor.
Pulse counting is accurate. As dawn breaks and the first few generated watts start to flow, the CUM pulse count recorded by emon every few seconds does not change at every sampling. It changes every nth sample but that quickly changes becoming continuous as the sun rises. And if you are visualising data at say 5 or 10 min intervals then you won’t even see this. But the pulse count will have accurately recorded that tiny first flush of generated power.
Whereas an inaccurate CT which measures instantaneous current (from which instantaneous power is calculated) will be even more inaccurate at the low generation levels at dawn & dusk. All CT’s are inaccurate – it’s just a matter of degree and cost. Furthermore power flow is calculated from instantaneous current & voltage values sampled every circa 10 secs – this of itself is an approximation. And, to cap it all, the AC/AC voltage sensor is inaccurate to some extent.
As an aside – I recently became aware that smart meter clocks drift (my son’s is 6 mins slow) and so that can complicate precise comparisons between downloaded smart meter 30 min data and emon recorded data. If connected to the internet, emon time is accurate internet time. Also smart meters use GMT throughout the year whereas emon visually presented data switches to BST when appropriate – but this is selectable.