ProDave and Marvin, thanks so much for that input, it's useful and interesting.
Yes I can see the advantages of missing out the DNO in the issue, but I'm wondering whether I would miss the flexibility. I need to think about whether I could justify the expense of a MyEddi.
Marvin, my back of the envelope calculations were:
2400 W rated system, so 2000w good case scenario output for 8 hours max per day = 16KW. Just to get an attainable upper benchmark.
A 300 Litre tank, with ideally temperature increasing from 5C to 65C (an increase of 60C), would require this much power:
Pt = (4.2 × L × T ) ÷ 3600
L = litres = 300
T = temperature = 60
Pt = (4.2 X 300 liters x 60) ÷ 3600
= 21KWH
If we only wanted to heat it to say 45C, and only needed 200L, then it's:
Pt = (4.2 x 200 x 40) ÷ 3600
Pt = 9 KWH
So, basically the same number you got, which is reassuring!
So this system would, on a good day produce enough to sort us out for hot water. Obviously, as with any scenario with PV, when there's less sun, less output, but even so, it would be a helpful addition for us at what might be quite low cost. Hopefully I can get an old inverter (of either type) for say £300 and 6x 400w panels for say £115 each, so about a grand. I have a roll of 6mm cable. No idea what the ROI would be, but probably around 10 years (10% ROI), plus reduced CO2 emissions.
But still not sure what the best set up would be and how to automatically prioritize (or not) where power gets used if it's via the domestic AC system, or else how to set up a direct system.