Hi all,
I posted details of my solar tracker under solar.
This is an overview of the rest of the system. Fuses, circuit breakers and switched have been omitted for clarity but res assured they are there; It probably makes most sense if I explain how it developed (and not much, even then):-
Stage1
This shows my 2 unisolar ES64's feeding power to regulator reg1. which works as a shunt regulator, maintaining 27.6v on battery B1 and sending the surplus to B2.
B1 is a pair of ex-UPS yuasa AGM 12v 100Ah (parallel connected to the GEL batteries in my small powerware truesine UPS).
At this stage B2 is 3x 100Ah FLA traction batteries (I got from the scrapheap in the boatyard) wired in parallel. Actual Ah measured as: 50, 35 and 25 using a 48A test load (4x 12v 12a lamps), but with my usual loads of 1 to 4 amps seem to be good for ~200Ah together.
Reg1 is based on a 34167 5A switching reg. A 741 opamp is used to 'reverse' the logic of the regulator IC feedback input (NB if you try this, don't forget to limit the volts going into the feedback input to 7volts max). A simple FET & zenner circuit operates the inhibit pin to shut the IC down if the battery (B1) is <26.5v to conserve power.
I thought that the surplus amps would convert down to 12v OK, but I had forgotten that the maximum 5A of the 34167 switch would limit the OUTPUT to 5 amp (and therefore the input at 28v to 2.5A (panels can output up to 5.5A on a very bright day)). Oops!
luckily I'd acquired some more boatyard junk (12v boat fridge and an optima yellowtop 75Ah & a yuasa gelcell 24Ah) so the system developed to stage 2:-
I reconfigured B2 to a series/parallel setup (i.e. the two weaker batts parallel,in series with the strong one - I can practically hear the 'puritans' of the battery-charging world turning purple in the face already
). This means that reg1 can handle the amps from the panels.
A new (temperature compensated (LM335) 27.6v reg was built using an LM324 to PWM a fet at a few Hz (Reg2). This is not a shunt regulator but goes open circuit instead. Reg1 was adjusted up to 32v Vpp-ish for these panels.
I set up the 'new' batteries in parallel as a 12v (nominal) 100Ah (B3) to run the fridge and using another 34167 made a v/2 reg so that B3 remains at B1/2 volts. Not ideal, but 24v fridges don't grow on trees.
If your wondering what stops B2 (the junk traction batteries) overcharging, the answer is: nothing - except that the max output from the panels is too small to seriously overcharge these batteries even if I wasn't using the power almost as soon as it turned up - the biggest problem was that B2 wasn't getting enough charging to keep it equalised.
This worked quite well (apart from the obvious equalization issues of B2) for a while as I converted more and more things to run off 24 or 12v instead of 240v ac (PC/TV, radio, lights, the above mentioned fridge, and the electric fence energiser).
Last winter I decided I needed more power, and as monocrystalline seems to have got a lot cheaper over the last 2 years I got two 80w panels and added them (along with another junkyard battery) to my system:
As reg1 is still limited to ~5A I connected these to B2 (B2 now has the 3 traction batts in parallel again, but now in series with two 6v batts (I don't know much about these - the only writing in their cases says "genuine hard rubber" - they are quite old and have high internal resistance but are good for ~200Ah at the drain currents B2 sees (<5amps )).
Now that B2 can see 10A+ it was time to add some dump load regulators. Regs 4 and 5 work in the same way as Ghurd's controllers but use op-amps (If I'd had the components to make ghurd controllers I would have, but as I did have salvaged opamps lying around it seemed sensible to use them). Due to the mismatch of the two halves of B2 it seemed sensible to use independent controllers.
The traction batteries have lower voltages than the other batteries - don't know why - but their rest voltage is around 12.0v and anything above 13.3 is an equalization voltage for them - hence the low setting for reg4.
The dumps themselves are peltier heat pumps (12v, 4.5a ish) that I salvaged from some junk at work. they were attached to heatsinks with fans to keep them cool. I've reversed the peltier connections so that they work as air-source heat pumps. This pic's not very good, but the two peltiers are mounted to a piece of thick aluminium angle which is thermally bonded to the 28mm copper pipe that convects up to the hot water tank. Don't really know how effective they are vs plain resistors but the pipe does get hot above them.
The auto switch simply switches the small loads (including the 12v reg) onto B2 preferentially (i.e it used the junk batteries up first, then when they're down to 23.5v switches to B1 and goes back to B2 when it rises past 25.5v.
This is a general view of the battery room (actually the airing cupboard):
Yes it is a mess, but there is method in the madness. the blue topped batts are the traction batts. the black ones are the 'genuine hard rubber' ones. they are connected together with 2.5mm wire as they don't see high currents (and are protected accordingly).
on the right are the yuasa AGMs with the 750va truesine UPS on top. In the foreground is the 1400va MSW UPS - also wired to the AGMs.
I want to put the new panels on to B1 (via reg2) but need to make a new 'mppt' shunt reg that can handle the amps.
I desperately need more batteries, but the ex UPS / traction that are coming available are all a long way away or expensive (or both).
And I'd like to automate the inverter to use it as another dump load (work in progress).
I will get around to it - eventually.
If you would like to see cuicuit drawings of an of the circuits I can probably post them - trouble is I tend to draw the circuits, try them on breadboard, make corrections 'til they work, then build them on veroboard without correcting my original drawings (or taking pics).
I'll post again showing what I do with the power (and how).
if you've read through all that - well done
.
I'm now bracing myself for the Opinions....
mab
