Thanks guys!
Here's my generating tally. The first one shows total energy use.
This one focuses in on solar as a proportion of total electric use.
This last one shows actual solar generation on its own plus the lifetime offset from the grid.
The really big 4kWh spike a couple of days ago was me running the fridge-freezer for 24 hours on battery / solar only. Drained the bank to about 80% DoD so won't be doing that often...
In my case, the system evolved and grew out of a simple 4A on-off controller to a 10A on-off controller and then one 15A Morningstar at 12V and then another Morningstar at 12V and then both converted to 24V to allow more PV power to be used as they are limited to 15A at 12V or 24V. And yes, the main reason now for two controllers is both capacity and mis-matched arrays with very different Vmp ratings that could not be optimised on a single controller.
The MPPT-15 can do load diversion to the load terminals but only up to 15A current limited and has to be programmed by the serial port to do this non-standard behaviour.
I'm waiting to see what the new Morningstar Tristar MPPT is like. This will do 60A at 48V and can take arrays of up to 150Voc. But I'd rather keep the DC side at under 60V as it makes DC wiring less problematic (cheaper) and I tried running the amorphous array in triplets for a max Voc of 75V (capped to about 68V with some bodging to protect the controllers that only go up to 75Voc). It didn't really work out so I went back to 35V.
Various bits of the arrays get shaded by the roof eves, trees, and even the overhead wire to the garage at times of the day and I found overall yield was better at 35V because when you have a triplet part shaded it effectively takes the whole triplet out of the array. If you have one of a pair part shaded then you only lose a pair in the array. My amorphous panels are quite shading resistant as the cells are long strips that run the length of the panel so a spot shadow (like a leaf) doesn't shade a whole cell (like it does on the traditional panels).
Apart from buying a more capacious charge controller (or a SunnyBoy 1100LV) I might redeploy the excess amorphous panels so that some are partly oriented for early morning light and others are oriented for late afternoon light. That way, I can create a passive tracked system. I can't use the peak output of all the panels but I can extend the generating day... The big Sharp panels probably will obscure the direct winter sun on the little Kyocera panels but I only installed them this summer so it hasn't been a problem yet. The frames that they are on point them almost straight up anyway so they mostly work off of either direct overhead summer sun or diffused light from clouds. This gives me some variability in collection as the big Sharps point at the the February solar altitude for maximum output on rare clear days but the others are all at a lower angle for either summer sun or diffused light collection. Playing with a few panels in different weather it soon became clear that on a cloudy day the best power is from the panel facing the bulk of the sky rather than the invisible sun. The wall mounted ones can be adjusted as they are mounted using brass hinges so by changing the stays (screwing the stay hinges to different positions on the wooden stay) I can change the panel angles. But I liked the idea of the big BP one being a sort of roof for my patio door.
I don't expect any real output between November and January as the sky is mostly very dark and the direct sunlight at 15 degrees elevation in mid December is so weak as to be mostly useless (over 65% atmospheric absorption). The graphs are a little mis-leading though as during the time from December to July, progressively more and more PV came on-line. So this winter will be marginally better than last.
I'm not really that bothered about return on investment... I'm just doing this for fun. In my view it actually has cost me nothing because I don't have to drive 50 miles a day to my old office. That saved me about £200 a month in petrol so all I did was invest my petro-dollars in renewables rather than blow it on booze and holidays
I manually load optimise at the moment... I let the bank charge on priority in the morning and when one or more of the chargers starts blinking that it's in absorption mode, I go and switch the computer room over from grid to solar power by swapping the input plug on the UPS. I toyed with doing this automatically (I might still try) as the Morningstars do have a data logging and realtime output that the PC could read and with a bit of Visual Basic and a USB relay driver from Maplins I could get the PC to manage its own load by selecting the UPS input. As Sharkey pointed out to me though, you need to be careful switching the AC to big inductive loads (the UPS is line interactive and has a big transformer) as a mis-matched phase on a change-over relay will cause a kick-back that can blow up stuff. You need to let the transformer rest for a second (to let its magnetic field decay) before connecting the mis-matched phase.
If you use two relays programmed with a delay you then also need to consider relay failure modes (welded contacts) that could allow both AC sources to meet and BANG!..
The new Tristar is especially interesting as it will have a direct Ethernet interface (rather than serial) so with a wi-fi transmitter, I can have the PC talk to it without running any more wires around the house. Wi-fi transmitters usually run off of 12V DC so it's easy to power them from the solar system directly.
I'm toying with the idea of using a 3kVA site tool transformer to step the 230VAC to 110VAC and feed that by a change-over socket to the immersion heater upstairs. Again, it's because I've only got a 1kW inverter. The immersion heater is a regular 3kW type so it's too big as a load for my system but if you run it at 110V it only consumes about 600W so would make a good diversion load on sunny days. That would pre-heat the water before the main heating cycle at night, saving some of the 5kWh grid power it takes to heat 100L of water.
There's plenty more toys in the box...
The Japanese have been going solar for decades. My wife's nan's house has the bones of a solar water heater on the roof and I've even seen a monks house at a temple with one on it. PV is taking off now as well but you see loads of water systems all over the place when you travel by train. I even saw a PV bus shelter that powers a hotel hot spa complex and astro observatory.
