Basics — Atechnical Man

Tim's sculptures could be installed anywhere, which includes off-grid. For myself, I've been frustrated by cheap solar powered landscape lighting so everything is plugged into 110V AC. So this is an opportunity for me to green up.

There are lots of tutorials on DIY solar projects, so I'll keep this at a high level. The basic components are panel, battery, charge controller and cabling. Worst-case environment runs from Texas toad-chokers to Boston Nor'easters.

I started last summer, so here's a catch-up on the progress to date.

Load

First step was estimating the load to be supported. My current LED landscape lighting is 12V AC, I did a quick test and it also runs 12V DC (paying attention to polarity). All together they drew about 45 Watts, so I rounded that up to 100 Watts for expansion. It was a place to start.

Since these are for lighting, they can run for a few hours after dusk. In Boston mid-winter, that starts at 4:10pm (Atlantic Standard Time, anyone?). Some people stay up 'til midnight, but I'm off the streets by 9:00pm. So 5 hours should do it.

100 Watts for 5 hours requires 500 Watt hours of electricity. We'll see how that goes.

Battery

Assuming a 12V battery, 500 Watt hours would require 500/12, approximately 42 Amp-hours. Battery performance is affected by temperature and other factors, see this FAQ. From the FAQ, here is the capcity vs termperature curve:

Good news for Texas, capacity increases in hot weather (bad news, life decreases). Boston mid-winter gets below zero, dropping us to 80% of rated capacity, so a 50Ah rating should do it. Fortunately low temperature also improves battery life.

I chose a 12 Volt 50 Amp Deep Cycle Sealed Lead Acid battery. It's not clear from the description if it's outdoor rated, but I've got a dry location.

Solar Panel Capacity

Simplifying, if we've got a 100W panel exposed to Boston mid-winter sun from 7am to 4pm (9 hours) we'd collect 900Wh of electricity. You'd need a motorized mount that tracks the sun, and have no blockages to do that. But how much can I expect? I'm going to optimize for best performance in mid-winter.

A two axis tracker would keep the panel directly facing the sun, so it would output at its rated performance of 100W. A fixed mount will receive less energy and must be derated accordingly. I looked up Boston in the NREL Solar Radiation Data Manual (TODO: draw a figure to show the tilt angles relative to horizon). For winter optimized fixed tilt (latitude + 15 degrees), the solar energy falling on a 1m square panel in December is min/avg/max 2.0/3.1/2.9 kWh. For a two axis tracker, it's 2.1/3.1/4.4 kWh. So I'm losing about 34% of max. I'll derate the panel to 65% of rated output. Now I'll expect .65 * 900 Wh = 585 Wh.

I'll have to derate further for blockages. For that, I'll need to know the Sun's path. This calculator gave me:

On the 21st December, the sun will rise 68° east of due south and set 68° west of due South. Elevation is 24 degrees max.

I've got trees to the East and West, see picture composite below taken in early January. I get close to full sun from 10:30am to 1:30pm, and let's say another hour equivalent for the afternoon through trees. Something like 5 hours of exposure. That's another 65% derating, so a 100W solar panel's down to .65 * 585 Wh = 380 Wh per sunny day. With other losses in the system not yet accounted for, looks like we'll need two 100Wh panels to fully charge the battery on a sunny day.

Jumping ahead a bit, the charge controller I selected adapts for multiple series connected panels, so I started with one panel. Now I'm at two panels. That's max for my charge controller ...

Charge Controller

With all the other losses, I decided to go with an MPPT charger controller. I wanted rainproof, I found one with IP67 rating (Immersion, up to 1 m depth). So I can't let it swim in the deep end. It's limited to a charging current of 15A, and max 60V input voltage (from the panels). It just barely misses support for a third panel. Looking back I could have put it in an enclosure to protect against rain, and not worried about floods.

Panel Wiring

The best ground level site for the panels is about 180' from where I wanted my battery. I used this calculator to determine voltage drop vs wire size. I went with 2 panels for my calculation, knowing I'd get there eventually. I entered: Copper, 10AWG, 40V, DC, Single set of conductors, 180', 5Amps and that gave me:

Voltage drop: 1.80
Voltage drop percentage: 4.50%
Voltage at the end: 38.2

Compared to the other derating, that's not too bad. With only one panel, that would be 9%, which starts to hurt.

Load Wiring

Current on the load side is higher, since the voltage is lower (12V vs 40V), and the time spent discharging is shorter then the time spent charging. While the distance is shorter (60' max), I'm only using 12AWG landscape wire. I've got a few separate runs, but still a 60W run will draw 5A. The calculator says I'll lose another 8% on high loads with a long run.

You'll see that I ended up adding an inverter to run 110V AC to my holiday lights, which lets me use even longer runs. They don't spec efficiency on the cheap inverters, and I suspect that's another 10%-15% loss. I'll be able to measure that loss once I get instrumented for current and voltage on the load side. I've also got a kill-o-watt so I can measure the AC load too. TBD.