Northern Sun Power Ed LaChapelle and Meg Hunt We were a long way out in the Alaskan bush, over 100 miles from the nearest power grid, and spending more & more time in a two-room log cabin while planning our bigger homestead. The little cabin worked on dry batteries for a radio and kerosene lamps. The latter were a fire hazard and would never do on a larger scale. Photovoltaics were the obvious way to go, but we had to start from scratch on the design. Seasonal Swings At 61¡28' N., the seasonal swings in power demand just for lighting would be huge. No solar insolation data for the area were available. Climate data and our own experience told us that prolonged periods of completely clear weather were limited, occurring mostly in the spring and fall. "Partly cloudy" was the most common sky description, often meaning cloudy part of the day, broken clouds, thin clouds, or clouds over part of the sky. The conditions hardest to use for predicting power output from solar panels. Our power requirements were also fuzzy, except that we knew they would probably increase as our bush lifestyle developed. All of these factors combined to make us go light on theory and heavy on empirical observations and hard experience. Gathering Data So we started out small. For the little cabin we had one Kyocera J-48 PV panel, one 200 Ampere-hour battery (used, from a fishing boat), one PL-13 lamp, a radio and a homebrew manual controller with a good ammeter. A car stereo outfit was added later. We started a regular program of logging panel output throughout the day in a variety of sky conditions. We experimented with panel location, angle and effect of tracking (by hand). When we were away for a couple of months our neighbors down valley, Kirk and Lisa Olsen-Gordon, also solar energy enthusiasts, took over the panel, controller and observations. They gathered many additional numbers for our growing tabulation of available sun power here in the mountains of south-central Alaska. In the meantime we acquired and remodeled a much larger log cabin nearby. This was going to be our solar- powered homestead. By this time, we had accumulated enough of that hard earned experience to start projecting our power needs and figuring out what would be required to meet them. We took to heart a guiding principle of home power and started first on power conservation and load management. Conservation The best way to practice conservation is to unplug it altogether. Among other things, Meg found a good hand- powered coffee mill and an iron that could be heated on the stove top. She also retrofitted her sewing machine with a treadle, finding it more fun & powerful; besides it doesn't generate radio interference. We determined to use PL lamps, which combine good light qualities of incandescents with the power savings of fluorescents., We also knew that we needed to get more natural light into the typically dark log cabin so that we wouldn't use the lamps in the first place. Skylights The obvious way to get more light into a log cabin is through skylights. This can be a problem in snow country. The problem is not so much the considerable weight of the snow but what it does when it starts creeping and sliding. A skylight that sticks up from the surface of the roof is in for trouble. Fortunately, I devised a way to build skylights flush with the surface of the steel roofing. We have two 2 ft. x 2 ft. and one 1 ft. x 2 ft. skylights. These, along with existing windows, white panels in the ceiling and a pine floor give us enough light to go lampless from wakeup to after supper from March to October. Constraints Our site included a couple of constraints on our solar energy use. One was a surrounding small forest of poplars, the ubiquitous Alaskan weed tree. Fortunately our plans included adding a second storey library space to a separate shop building. A platform on the roof offered an ideal solar panel location, although it meant running about 75 ft. of cable to the batteries. Good solar access more than compensated for cable losses, figured to average around 5-7%. The other constraint was cold batteries. The logical battery location was in an existing cellar underneath the cabin, but in these latitudes not far south of the permafrost zone, the mean annual ground temperature is not far above freezing. The cellar gets well below freezing in winter and creeps up to about 45¡F. by late summer. It makes a great refrigerator, but is not a happy place for lead-acid batteries. But it was the only place for the batteries to protect them from outside temperatures down to 60¡F. below if the cabin is unheated in the winter. Our system design had to allow for loss of battery capacity, plus run a resistance heater in the battery compartment to compensate for temperature by using surplus power diverted from the PV array. Final Design Our final design included eight Kyocera J-48 photovoltaic panels, four Trojan L-16 batteries, a Trace 1512 inverter/charger and a pair of Trace C-30 controllers. This is pretty much a conventional package, but the whole system is hooked up in an unconventional fashion to solve some anticipated problems. A pair of controllers was the result. Help Early in the planning stage, I visited the Trace Engineering factory in Arlington, WA. I garnered much useful information from the helpful folks there. Mike Frost, Trace's design engineer, pointed out that when large loads are switched on and off the inverter, there are wide swings in the battery's voltage that could cause trouble for 12 volt electronics on the same battery. This set in motion the design of a split system to operate 12 volt circuits and the inverter from separate battery banks., This system has several advantages. For one thing, it is redundant, offering built-in back-up power in case something breaks down. It also becomes very flexible if provision is made to switch solar panels between the two parallel controller-battery circuits. The Split-System Control The heart of our photovoltaic system is a dual-channel controller built around two Trace C-30 PC boards. These boards have been modified by replacing the SPST relay with a physically identical relay with SPDT contact configuration. This allows the use of diversion power from the array. The A-channel (12 volt circuits) switches diversion power through a separate controller (Trace C-30A) to charge auxiliary batteries, including a neighbor's hooked up through a set of jumper cables on the cabin's outside. The B-channel (inverter) feeds diversion power to the battery heater. Two solar panels are permanently connected to the A-channel, two to the B-channel. The remaining four panels can be assigned to either channel through bistable impulse relays located in a junction box next to the solar panel array. These relays are controlled by momentary-close switches located on the controller along with LED status indicators for the switchable panels. Engine/Generator Backup When the sun appears only an hour or two a day around the winter solstice and cloudy weather is common, recharging the batteries with an auxiliary generator is essential. In choosing a generator we took careful heed of the fine print in the inverter manual warning about the importance of keeping up ac peak voltage to insure high-rate battery charging. The Onan 3. ORV has proven very satisfactory in this respect, for it delivers full power even under heavy loads. It is a compact, 3600 rpm model with high-volume axial flow cooling. Being located inside the shop space, its blast of hot air can serve as a useful auxiliary heat source. So far under our present energy demands, we have not needed to use the generator at all from late February until mid-October; the solar panels do it all. Further, owing to the wonderful performance of our Trace 1512, we never have to use the generator to run power tools. Array Angles The solar panel array faces due south and has provision for seasonal adjustment of tilt angle. In winter mode, the array is tilted up 72¡ from horizontal, in summer mode, 48¡. The angle is changed around the spring and fall equinoxes. Although some solar energy users at these latitudes simply hang their PV panels on the vertical south wall of buildings, we found that this is quite inefficient on cloudy days. Overall power production is optimum when the panels are tilted back far enough to allow exposure to bright clouds instead of dull trees, eventhough the panel angle may at times be a bit off from perpendicular to direct sun. Luxury The installation was completed in the summer of 1987. Since then we have enjoyed the luxury of all the power we need, not only the practical benefits but also the sense of satisfaction from generating silent, pollution-free electricity. In 1989 we added a 12 volt freezer, the only major increment so far in our power consumption. As received from the manufacturer, this freezer was woefully under insulated and inefficient. We covered the body with an extra two inches of blue foam insulation. Then we installed it on the north side of the cabin, where it is well shaded and the condenser coil can draw cold air by convection from a crawl space underneath the cabin. This brought about a notable improvement in efficiency, with the duty cycle now ranging from around 35% down to 15% as mean daily temperature drops from 60¡'s down to the 30¡'s. Problems Our only problems have been cold batteries and radio interference. Even with the battery heater and insulation keeping the batteries about 10¡ above cellar temperature, the derated capacity still leaves little margin for extra power storage. This problem is compounded by the lack of provision for a finishing charge in the C-30 controllers. Thanks to the dual channel system with switchable panels, we can compensate in part my manually reducing charge rates to top off the batteries. In fact, we have come to believe that the ideal controller would achieve a tapering charge by successively disconnecting panels from the array, rather than trying to taper the full array current by pulse-width modulation. Planned Improvement Our next system improvement, scheduled for the summer of 1990, is to put in pocket-plate ni-cad batteries on the A-channel (12 volt circuits) and place all four L-16's on the B-channel (inverter and freezer). Again, the flexibility of the dual system comes in handy, for we can add ni-cads for part of the power storage without having to dump the lead-acid batteries and replace the whole works. Radio Interference Fighting radio interference from a PV system with an inverter is a whole story in itself. The problem is a critical one for us here. Our main radio reception depends on weak, remote fringe area signals from AM stations on the other side of some very big mountain ranges. Extensive shielding and filtering help. Isolating the radio power supplies to separate, auxiliary batteries helps even more. The inverter generates interference even in standby mode, so this is never used. A pushbutton and solenoid allow remote control of the inverter in the cellar, so we can turn it on only when ac power is actually required. We're still working on these RFI problems and are keen to exchange information with other Home Power readers. Happy to Report We're happy to report that home power is very much alive and well in our part of the world. Most households we know have or plan to acquire at least one or two panels. Some have systems as large as ours, some even larger. The National Park Service is presently installing a full PV system to power a ranger station across the mountains to the north of us. Owing to the low solar power available in mid-winter in these latitudes, gasoline or diesel auxiliary generators are common, as well as reliance on propane for lighting. Micro-hydro is getting some local attention these days and we know of one case in which full-time diesel generation has been passed up in favor of part-time generation to charge a battery-inverter system. In the next few years expect to see Alaska become a leader in modern alternative energy systems. Ed LaChapelle and Meg Hunt POB 92723 Anchorage, AK 99509 (Printed by electronic typewriter, solar powered, of course!)