Size and wire a 12V solar system

The order you have to do this in

Almost every undersized system was designed backwards. Someone decided on a panel, bought a battery to go with it, then found out what it would run.

Do it in this order and the sizing falls out of the arithmetic:

1. What do I need to run, and for how long? Watt hours per day.
2. How many days must it survive with no sun? This gives the battery.
3. How much do I need to put back in each day? This gives the panel.
4. What sits between them? The controller.
5. What carries the current? Wire and fuses, sized to the current, not to what is in the shed.

Get step 1 wrong and everything after it is wrong by the same factor.

Step 1: measure the load, do not guess it

Write down every item, its wattage, and the hours per day it genuinely runs. Multiply. Add.

The duty cycle trap. A 50 W fridge is not a 50 W load. It is 50 W while the compressor runs, and the compressor might run a third of the time in a cool room and two thirds of the time in a warm one. Fifty watts at a one third duty cycle is 400 Wh a day. At two thirds it is 800 Wh. Same fridge, double the system.

If you can, measure rather than calculate. A cheap inline watt meter between the battery and the load, left running for 24 hours, gives you a real number including everything you forgot.

Add 20% to whatever you arrive at. You will find loads you did not think of.

Step 2: size the battery

Two numbers decide this: your daily watt hours, and how many days of no sun you need to survive.

Usable capacity is not rated capacity.

A 100 Ah battery at 12V holds 1,280 Wh nominal. In lead acid you may use roughly half of that before you start destroying it. Discharging lead acid deeply and repeatedly is the single most common way people kill an expensive battery in one winter.

The sum:

daily watt hours, times days of autonomy, divided by usable fraction = battery watt hours needed

Then divide by 12 to get amp hours.

How many days of autonomy? In Britain, three days is a sensible floor and a week is comfortable. A December week with heavy cloud will produce close to nothing, and that is a normal British week, not a disaster scenario.

See charging and storing a LiFePO4 bank before you buy, because the cold weather rules change what you should install and where.

Step 3: size the panel, and the December problem

Panel output is quoted at 1,000 W per square metre of sunlight. You will rarely see that. What you need is peak sun hours, which is the day's total energy expressed as the number of hours of full sun that would produce it.

The estimate:

panel watts, times peak sun hours, times about 0.75 for real world losses = watt hours per day

The 0.75 covers wiring loss, controller loss, heat, dirt and the fact that panels do not sit at their rated point.

So a 100 W panel gives you roughly:

• June: 100 x 5.5 x 0.75 = about 410 Wh a day
• Annual average: 100 x 3 x 0.75 = about 225 Wh a day
• December: 100 x 0.8 x 0.75 = about 60 Wh a day

Size for the season you need it in. A system sized on the annual average will disappoint you every December, which is precisely when a power cut is most likely and the nights are longest.

The two honest options are to fit far more panel than the summer needs, so that December still delivers something useful, or to accept that winter runs on stored energy and a second source. Both are legitimate. Pretending December looks like the annual average is not.

Panels are the cheap part now. Doubling panel area costs far less than doubling battery capacity, and it is usually the better answer for a British winter.

Tilt matters more in winter than summer, because the sun is low. See angling a panel through the seasons for the tilt arithmetic, which is the same for a panel as for a dryer.

Step 4: the charge controller

The controller sits between panel and battery, and stops the panel overcharging the battery.

Size it by current, not by watts: panel watts divided by battery voltage, then add at least 25%.

A 200 W array on a 12V battery is roughly 200 / 12 = 17 A, so a 30 A controller. Controllers come in standard sizes; going one size up is cheap insurance.

MPPT, not PWM, in Britain. MPPT converts surplus panel voltage into extra charging current instead of throwing it away as heat. Reported gains run up to about 30%, and the gain is largest in cold weather and where panel voltage sits well above battery voltage. Both of those describe a British winter, which is exactly when you need every watt.

PWM is cheaper and is fine in hot climates with a voltage-matched panel, where the advantage largely disappears. That is not here.

Check the controller's maximum panel voltage before buying panels, and remember panel voltage rises as temperature falls. A cold, bright morning is when an array exceeds a controller's input limit and destroys it.

Step 5: wire and fuses

This is the part that starts fires, and it is the part most often done by eye.

The voltage drop point deserves repeating, because it catches people who did the ampacity sum correctly. At 12V, everything is high current and low voltage, so resistance in the cable matters enormously. Doubling the run length doubles the drop. A cable that is fine for two metres may be useless at ten, not because it will overheat, but because too little arrives at the other end. When in doubt, go one size heavier. Cable is cheap compared with what it protects.

Fuse positions, all of them:

• Battery positive to everything else, as close to the battery terminal as physically possible. This is the most important fuse in the system and the one most often missing. Its job is to protect the main cable if it chafes through anywhere along its length.
• Panel to controller.
• Controller to battery.
• Each individual circuit at the distribution point.

On connections: a proper crimp with the right tool beats a twisted joint and beats solder alone in anything that vibrates. A bad connection is a resistance, a resistance is heat, and heat is how this goes wrong slowly rather than suddenly. The detail is in 12V wiring safety.

Assembly order

The sequence is not arbitrary. Getting it wrong destroys controllers.

1. Mount everything and run the cable before anything is connected.
2. Battery to controller first. Most controllers detect the battery voltage to decide whether they are on a 12V or 24V system. Powering one from the panel first can leave it confused or damaged.
3. Panel to controller last, and ideally with the panel covered so it is not producing.
4. Fuses and the main isolator in last of all, once you have checked polarity twice.
5. Loads only after the controller shows it is charging normally.

Reversed polarity kills controllers instantly and often silently. Check with a meter, not by wire colour, because someone before you may have used whatever cable was to hand.

A worked example

A modest but genuinely useful household setup.

Loads:

• 6 LED lights at 5 W for 5 hours = 150 Wh
• Phone and tablet charging = 60 Wh
• Radio, 3 W for 4 hours = 12 Wh
• Raspberry Pi running the household reference library, 3 W continuous = 72 Wh
• Subtotal 294 Wh, plus 20% margin = about 350 Wh a day

Battery, LiFePO4, four days of autonomy:

• 350 x 4 = 1,400 Wh, at 80% usable = 1,750 Wh needed
• 1,750 / 12 = about 145 Ah, so a 150 Ah bank

Panel, to replace 350 Wh a day:

• In June at 5.5 PSH: 350 / (5.5 x 0.75) = about 85 W
• On the annual average of 3 PSH: about 155 W
• In December at 0.8 PSH: about 580 W

Which tells you the real answer. Around 200 W keeps this running comfortably from spring to autumn and tops up slowly in winter. To run it through December on solar alone you would need three times that. Most people fit 200 to 300 W, accept a winter deficit, and cover the gap with a second source and a bigger battery.

That is not a failure of the design. It is what British December actually is, and a plan that admits it is worth more than one that does not.

What actually goes wrong

• The load was guessed, not measured. Everything downstream is wrong by the same factor.
• No main battery fuse. The most dangerous single omission.
• Fuse sized for the appliance rather than the wire. See the danger box above.
• Cable too thin for the run, so the voltage arriving is too low even though nothing overheats.
• Lead acid run flat repeatedly, ruined inside a season.
• Controller destroyed by connecting the panel before the battery.
• Corroded or loose terminals. Resistance, then heat. Check and retighten annually.
• The system was sized on the summer figure and quietly fails in the winter it was bought for.