We just spent $8000 to install a new set of batteries in our solar system. We installed our off-grid solar system in 2001. We have replaced the original batteries twice since then. As I’ve worked with a solar electrician and a solar engineer for the past two months to figure out what was going wrong with the old batteries and research alternatives going forward, we’ve learned a painful lesson. There are no good alternatives. Battery technology is not ready for prime time. Batteries are the Achilles heel of off-grid systems. We knew we were going to be on the leading edge when we built an off-grid cabin; instead, we found ourselves on the bleeding edge.
Most solar systems are grid-tied, meaning that they are tied into the local utility. In these cases, if you produce more electricity than you consume, your power meter runs backward, and the excess is put into the grid. In effect, you sell the extra power back to the power company. At night, when you aren’t producing power, you just draw it from the grid. These systems are fairly simple and comparatively inexpensive when you factor in the rebates.
We are not tied into the grid--we have an off-grid system. We have batteries that store extra power when we generate more than we need during the day. Then, we draw on that stored power at night. If the system falls short, we have a generator that turns on when the batteries run out of charge. The theory is wonderful. What I’ve experienced is that the theory does not hold up in the real world. From design to installation to maintenance to battery capacity and life, we've encountered problems that even the experts don’t know how to answer.
Because we bought land with foundations in place and building permits to go ahead, we could not optimize our site for solar panels. You want your panels to face south to maximize the solar power coming from the sun. Our main house has a ridge line going north to south, so we could not put panels on that roof. Our garage building is turned to face southeast, so we designed the roof to put the panels at the best angle to catch the sun, even though it is not optimal.
One of the first design questions we faced was how high to build the roof. Since the main house sits to the south of the garage, we needed to figure out how high the garage roof needed to be to minimize shading of the solar panels by the main house. In this picture, you can see the shade creeping across the corner of the panels. When the panels begin to enter the shade, they lose power and make the whole array less efficient.
These days, they use this amazing little device to figure the azimuth of the sun and calculate the answer to this question. I never saw anyone use such a device when we were building in 2001. The designer raised the height of the walls of the garage to elevate the panels. Still, in the winter, shade begins to creep across the array at 2:30 PM.
My contractor had his regular electricians install the solar array and wire up the whole system. The solar system was designed by an engineering company that lasted about a year after we installed the system. The contractor did not use their preferred installer, because he was a little high. Big mistake. Not only did they mess up the wiring to the batteries in the utility room, but they made a big mistake in the installation of the array on the roof.
In this picture, you see snow covering the solar panels. The three rows of panels are installed so there is about a six inch horizontal gap between the rows of panels. Consequently, when the snow melts, it drops into the gap instead of sliding off the whole array. When that happens, the snow stacks up and covers the bottom quarter of the panels for another couple of days, hindering solar collection. Nowadays, it is pretty common knowledge that you need to minimize this gap in cold climates so the snow will slide all the way down the array at once, freeing up the collectors. I’ll probably have someone go up and correct this problem, but it will cost $1500 or more in labor to do that. Poor design and installation there.
Solar power comes off the roof and is fed to the batteries as direct current, 48 volts DC. We have two microprocessor powered inverters that control the system. They sense the power demands of the house and convert power to 120 volts for use in the house. Excess power is sent to the batteries. Here’s what the inverters look like.
My first battery array consisted of 32 six volt batteries. Because power is coming in at 48 volts, we had to wire eight batteries in a series (8x6=48). So, we ended up with four strings of batteries in the battery room. Now the experts tell me they would never design a system with four strings of batteries run in parallel, because it is about impossible to evenly charge four strings and get the full capacity of the batteries.
The batteries we used were the same ones used in golf carts and electric manlifts. I had hundreds of manlifts in our rental fleet during my business days, so I was fairly familiar with this old technology. Because the batteries are not sealed, you must check the water level once a month. If the water level gets down so the battery plates are exposed to air, the batteries will quickly be ruined. So, the need for regular maintenance means I had to arrange a caretaker to come check on the place and fill the batteries while I’m away.
Our first caretaker did not understand the sense of urgency around this task, so the batteries were not maintained properly in their first few years. I’m sure the wiring problems to the batteries also contributed to our problems, but we did not see optimal service out of the solar system for the first few years.
With the original designer gone, we had to find someone to service the system. I hate it when the equipment outlasts the supplier who is supposed to service it. So, we brought in one of the most experienced solar people in the state to correct our installation problems. Ken Thames quickly found the problems in the battery wiring and suggested I install some additional boxes that would maximize the power coming off the roof. He also recommended that we equalize the batteries once a month. This is an hours-long process that we never really followed through with on this first set of batteries.
The system performance improved, but still never optimized. The lack of proper maintenance early on had damaged the batteries. Our batteries started failing after about four years. They failed a few at a time over the next year. We eliminated one string of eight batteries after several had failed. I did not want to replace the whole set and was buying time. Over the next year, I put several new batteries in to replace bad ones in the system. Bad idea.
In the fall of 2006, when the batteries were five years old, I realized they were all failing. When batteries start to go bad a strange thing happens. When you start charging them with the generator, the voltage immediately goes up in an artificial reading. The generator will then cut itself off, thinking the batteries are fully charged. As soon as the generator quits and the batteries go from being charged to being discharged by the house loads, the voltage immediately drops. The telltale sign is that the generator runs way too much. By Christmas, the generator was coming on for an hour, turning off, and coming back on an hour later. Normally, it does not run more than once a day, if at all.
In January 2007, I bought new batteries, keeping only the six that I had bought over the last three or four months as the original batteries failed. So, now I had 26 new batteries and six that were a few months old. Things ran fine for a while, but still not to expectation. The experts told me that 32 batteries should hold about 48 KWH in storage capacity. Most houses in Colorado consume about 25 KWH a day, so the batteries should hold about two days, even with snow on the solar panels. We never once saw them last more than 24 hours before the generator came on to charge them.
The six batteries that I did not replace with the new set I bought in 2007 failed in 2009. Even though they had been in the string with the bad batteries for only 2-3 months, it was enough to damage them. So, I bought and installed six new batteries in 2009, along with the 26 I bought in 2007. In the fall of 2010, problems came along again. This time, I had made sure the batteries were watered regularly, but still did not equalize them regularly. By early January, it was evident by the hours the generator was running that I had a problem. I spent quite a bit of time with Ken Thames and a solar engineer I met trying to troubleshoot what was happening.
By February, the generator was running several hours a day, and I faced the fact that I was going to have to buy another set of batteries. I asked two different suppliers to recommend different options. By then, Ken had made it clear that having four strings of six volt batteries was a bad design and would never work well. I looked at 12V sealed batteries, but found I would have the same problem. I would need to arrange four of these in a series, and would need at least six strings of four batteries. That would eliminate watering batteries but would increase the number of strings, but it would be impossible to charge them equally. I also found that if you let them overcharge, sealed batteries would destroy themselves. And, the cost was over $12,000.
Ken recommended looking at heavy duty two volt railroad batteries, and installing one string of 24 in series to get my 48 volt storage. Problem is, those batteries were about $15,000 and would not fit in my storage racks, adding even more cost. Then, he found that Trojan had built a two volt battery that was the same size and storage capacity as my six volt batteries. These batteries only cost $25 more each than the six volt batteries. Even though there is tremendous work going on with lithium Ion batteries for hybrid cars, there is no cost effective battery for off-grid use. So, we decided to install 24 of the new design two volt batteries in a single string. They fit in the same storage racks. In this picture, we are removing the old batteries.
Here are the new two volt batteries in the racks.
When, after the installation, the generator kept coming on anyway, I went back and started running tests to see if the batteries were operating up to specifications. I had been told that they should hold 36 kilowatt hours of storage. When that did not happen, we began to investigate. I then realized that the storage was theoretical, and that the system starts the generator when the batteries get to about 40% full, because drawing them down too much is bad for the batteries. So, the reality is I only have about 28 KWH of storage.
Our Solar Engineer, Zeke Yewdall, hooked up a gauge to measure our consumption. With his gauge in place, I find that we are consuming around 28 KWH a day, here in the dead of winter with short days and more lights and heat being used. With some clouds and snow affecting our solar panel production on the roof, we are generating around 22KVH a day. On a sunny day, we generate almost 30KWH even in winter. In the summer, days are much longer and sunnier, and we can get well over 30 KWH production in a day. But in the winter, we are not producing as much power as we consume, so the generator must make up the difference.
When the electrical engineer was designing our system, he was afraid to run our well pump off the solar system. The well is 500 feet deep, and the pump has a two horsepower electric motor running it. The engineer specified a generator based on what was needed to start this motor. Electric motors take a spike of energy to get the motor turning, then the consumption falls to maybe a quarter of the starting power. We also have to derate engines at our elevation, 9000 feet, because the air is thinner. So, he specified a 25KW generator, which was derated for being powered by LP, and further derated for the elevation. So, whenever the generator runs, it takes about three gallons a minute of fuel.
Ken Thames thought the engineer was full of it, and promptly proved that the solar system could run the well pump fine, thank you very much. So, I probably needed a 15KW generator, and got one nearly twice that big, wasting fuel every hour it runs. It has over 6000 hours on it since it was new, but it is another several thousand dollars to replace it with the right size. Here is a picture of the generator in its shed.
So, to sum it all up, I do not enjoy being my own power company. When you are on the grid and lose power, you know the power company will have it fixed pretty soon. When an off-grid system goes down, you have to send for help, and it usually costs $500 to get someone up here for a repair. For this reason, I have an alarm system which is tied into the solar system to alert me when I’m away. When we lose power, everything goes down. No heat, no water, no lights, no nothing. It gets serious in a hurry in the winter, when the plumbing will freeze if the heat quits.
A few years ago, we were on an anniversary trip to the Caribbean. I kept my phone turned off to save the steep roaming charges. I missed a call from the alarm company. A few days later, I saw the missed call and found out my son and my caretaker had gotten calls, but had not come up here. Luckily, it was not cold enough to freeze pipes, but we had to through out all the food in the refrigerator/freezer because it all got warm.
So, although I love being off the grid and living in the wilderness, there are trade-offs that get old. I hope by the time these batteries die there will be completely new technology available. Until then, I am going to make sure that these new batteries are watered and equalized regularly and keep my fingers crossed. I’d be quite pleased to see the six years of life I have been told to expect in such a system. Haven’t seen it yet, maybe the third time is the charm. If you are interested in more about off-grid living, see my first article here.