Solar Powered Rainwater Project – Final Assembly

With all the parts in stock, it was time to assemble the whole Solar Powered Rainwater Project. I cleaned out a space in the garden shed, and added plywood to the wall for attaching the various components. We would be mounting the inverter/charge controller, solar panel combiner/breakers, A/C transfer switch and related wiring.

Mount up!

photo of Everything mounted!

Everything mounted!

The inverter/charge controller was first. At about 100 pounds, it was too much to handle alone, so my friend Norman helped. The good news was that the inverter mounting holes were 15 1/2-inches apart — perfect for mounting to a wall with 16-inch-on-center studs. That done, we mounted the solar combiner box and transfer switch.

Connecting the Solar Panels

photo of the Panel combiner wiring

Panel combiner wiring

Wiring the solar panels was next. The two, 24-volt panels are connected in series to produce 48 volts, so only one 2-conductor wire was needed. I used 12-gauge landscape wire because it is relatively inexpensive and is intended for outdoor service. Assembling the DC breaker into the solar combiner box was simple, and the box has space for two more breakers to connect additional 48-volt panel strings. Another pair of heavy wires connects the combiner output to the DC input on the inverter/charge controller.

Wiring the A/C Mains

photo of the A/C and PV panels connected

A/C and PV panels connected


The next connections to the inverter were for 240 VAC power. Commercial power is connected to the input, and the output connects to the transfer switch in my system. This way, we can select either solar generator or commercial power for each load; water well, greenhouse (with the rainwater system), lighting and outlets. It’s an added complication, but otherwise I would have to power everything with either solar or commercial power at the same time.

The A/C Transfer Switch

photo of the Transfer switch wired

Transfer switch wired

Now, the transfer switch. These are commonly used with generators so that each load can be selectively powered by either the generator or commercial power. It has the added advantage of preventing generated power from connecting to the commercial power grid. This is very important because linemen working on the commercial grid during a power outage are not expecting supposedly “dead” power lines to have power, and might be injured. This keeps everyone safe.

Configuring the Inverter and Connecting the Batteries

After mounting the 200-amp DC breaker and connecting the batteries, we were ready for testing. I configured the inverter/charge controller according to the manual and DIP switches were set for Deep-cycle battery, utility power mode, 60 Hz line frequency and battery priority. Even though I had checked each connection made over the last few days, there was still a little thrill of anxiety — did I get it all right? Also, would the inverter be stout enough to power the well pump? If it worked, I could be sure that the rainwater system would work since it uses only half as much power.

But will it work?

photo of the Inverter control panel

Inverter control panel

Time to power the inverter. The lights on the front flashed with their self-test and the “inverter” LED stayed lit. Everything was powered. As I threw one of the transfer switches for the garden shed lights to the “GEN” position, they didn’t even blink. They were running on the inverter. Now, the outlets that power the ultraviolet filter. Again, not even a blink — still running. Moment of truth time. A sprinkler was running in the garden and soon it would start the well pump. I flipped the switches for the well pump to “GEN” and waited. The pressure switch clicked on, and the well pump started humming. The inverter had started a 1 1/2 horsepower well pump. Success. We would be celebrating tonight!

Measuring the results

photo of a Digital power meter

Digital power meter

Finally, I installed the last element of the system — a digital meter that measures voltage, current and wattage drawn from the battery bank. With it, I can see the battery voltage change as sunlight and power loads change — confirming that the solar panels and MPPT charge controller are working as expected. What I know already is that we’ll need two more panels. My calculations predicted this, but I didn’t realize that the inverter draws a constant 70 watts when powered. That, combined with the  power used by the ultraviolet filter (about 30 watts) mean that most of the power generated by two panels in 24 hours is used each day. 100 watts times 24 hours is 2400 watt/hours used. 500 watts of solar panel times an insolation of 5 is about 2500 watt/hours of power generated. Most of the power generated is being used by small, but constant power drains that cannot be avoided. Still, we have a complete, working system to learn on. Very exciting! Now — about those panels . . .

photo of the Transfer switch with power meter

Transfer switch with power meter

Solar Powered Rainwater Project – Batteries

You may recall from an earlier post that the Solar Powered Rainwater Project requires substantial battery storage. In our original calculations, we decided we would need 15,796 watt/hours of storage to provide three days of autonomy (or days without sun). We would need 16, 190-amp hour golf cart batteries to reach this goal. This becomes not only a cost issue (about $1800) but also a logistical and space issue. Each battery weighs around 60 pounds. That’s nearly a half ton of lead!

To test my ideas and get the project running more easily, I decided to cut the number back to eight batteries. It turned out that Interstate 6-volt golf cart batteries were available from Costco for $83 each, and were rated at 205 ampere hours. Eight batteries would yield 9840 watt hours of storage — nearly 2/3rds of our original goal. Plenty to get the project started.

Storing a 1/2 ton of lead

graphic of Battery cart plan

Battery cart plan

The next problem was, where would they go? The garden shed has a concrete floor, and I realized they would partially fit under shelving located inside. If I had a way to roll them out for easy service, they wouldn’t take too much space. My friends at Vintage Auto agreed to weld up the cart I designed. Storage problem solved for the time being.

300 Amp Wire

Now on to cabling. The inverter is rated at 4000 watts, and accounting for inverter efficiency and the 48-volt DC supply, we would expect a 95-ampere current drain. The inverter can supply more power  though — up to 12,000 watts peak. Brief surges of up to 285 amps could be expected. Just to be safe, I decided on 0/1 rubber-covered welding cable. Perhaps overkill, but I like that. The cables weren’t hard to assemble with the copper terminals and hammer-actuated crimper, and I had the them done in no time. I also added heat-shrink tubing with internal adhesive to prevent corrosion inside the crimp.

With a shunt for metering installed on the negative lead, and a 200-amp breaker, we were ready to power the inverter!

Weber Piano Action Rebuild – New Hammers!

We got back from the Red Power Roundup and it was time to catch up on the Weber project. Bernard texted me (with pictures!) that the hammers had arrived from Renner, and I was eager to see them. I went to his shop a few days later.

The New Hammers

Hammers get progressively smaller from bass to treble, but telling them apart can be tricky since the size difference between nearby hammers is small. A few minutes numbering them with a pencil solves the problem. Next, we wanted to mock-up a few hammers to confirm operation and clearances, so I installed eight of them at various positions in the action. With the wippens reconnected on the same keys, we were ready to test operation.

An Initial Test

photo of Testing hammer actuation

Testing hammer actuation

Bernard explained we were striving for a consistent response to a 56-gram weight placed at the end of each key. He produced an elegant wooden box containing progressively-sized brass weights and proceeded to test each of the installed hammers and keys. They were a little stiff, requiring as much as 60 grams to actuate, but certainly within the expected range of reconditioned keys with new bushings. Some “easing” would be required, the next task in my home shop.

Easy now . . .

photo of Mortises with bushing and cauls

Mortises with bushings and cauls

The easing process involves a product called “Profelt,” applied a drop at a time to each pair of bushings and then “ironed” with a plastic caul. The process is similar to installing the bushings in the first place, with the goal of forming the felt to a specific thickness that will glide on the key pins with as little friction as possible. With a supply of Profelt and a box of cauls, I could look forward to a quiet afternoon of easing the bushings.

A Bit of a Sticky Wippen

diagram of a grand piano action

Grand Piano Action showing the back and balance rails

The wippens could also be a source of friction. By applying a second chemical — simallarly-named “Protek” — to the three pinned bushings in the wippens, we could remove that friction. I began applying Protek to the balancier (aka repetition lever #9 in the diagram), jack (#5) and action center pins. The liquid is applied to several pins, and then treated with a heat gun — hot, but not the point of burning the wooden parts. This activates the chemical. After the process, the wippens seemed to move more easily around the action centers.

photo of Whippens rail suspended in a Jarrus Piano Cradle

Wippens rail suspended in a Jarrus Piano Cradle


There were still differences from wippens to wippens though. We suspended the wippens rail into a “Jarrus Upright Piano Action Cradle” so that they could be compared easily and I could mark any slow action centers with white chalk. A second application of Pro-Tech on these more stubborn wippens seemed to do the trick.

By the end of the day, I had several tasks queued up:

  • Profelt the key bushings (as mentioned earlier)
  • replace the let-off buttons
  • install the new hammers
  • replace the damaged hammer cushions
  • begin the process of key leveling
photo of a key weight

Key weight

During an earlier visit to Bernard’s shop, I got two boxes of lead weights that are hung on the back checks (#11 in the diagram) to simulate the resting key position with the rest of rest of the wippens and hammer stack. With them in place, a straight edge is placed on the key tops (and later sharps) to check for key level. The goal is to make all the keys exactly the same height by adding paper punchings to the key frame balance rail.

With these tasks (and refinishing the sharps) done, we will be ready for reassembly of the entire action to begin the regulation process. I can’t wait!

photo of Applying Protek to an action center

Applying Protek to an action center

Solar Powered Rainwater Project – The Chinese Connection

It’s been a while since I reported on the Solar Powered Rainwater Project. Since that early-April post, I’ve been gathering the rest of the parts, and we took our annual trip to the Red Power Roundup — this year at the Iowa State Fair in Des Moines. The biggest part of the delay was waiting for the inverter/charge controller. I wasn’t thrilled with the delivery delay on the device, but it turned out well in the end.

The Ideal Thing

Anyone who has used Amazon or eBay to order product from China has probably experienced the longer shipping times these Chinese products frequently have. Air shipping is expensive, and if you can wait for container shipping, the cost is small and often zero. The inverter/charge controller was ordered from Sun Gold Power in Shenzhen, China through Amazon. As I mentioned in an earlier post, The device combines a 4000 watt inverter, 40 amp MPPT charge controller, battery charger and A/C transfer switch in one package. It can be used off-grid, grid-assist or as a uninterruptible power supply (UPS). It even has provision for a backup generator. This is a very convenient product, and I was unable to find anything else like it in American product offerings. Unfortunately, there was very little information online and no product reviews in the Amazon store. I contacted the company directly through their Web site, and they responded quickly to my request for a manual. A good sign.


The product manual is reasonably good and answered many of my questions regarding how it would work in my application. The 4000-watt inverter was capable of producing a 12,000 watt surge for 20 milliseconds — hopefully enough to get the 2000-watt inductive load of the well pump started. In spite of sparse information from other customers, I decided to order one.

After placing the order with Amazon, I quickly got an email from the company. They wanted to confirm the choices I had made, and I was glad they did. I needed 240 VAC, split-phase output, and had inadvertently ordered 120 VAC instead. That would have been disappointing! They didn’t have the specific model in stock but would be making them in about two weeks. Would that be Ok? One more thing: the next production run would be for orange cases instead of blue. Was that a problem? That was fine with me, but perhaps they could speed up the shipping? Yes, and for a little extra cost ($55) they could ship 3-day by air with DHL. I paid the extra shipping and the deal was done. In about three weeks I would have a shiny new device.

This may take a while . . .

I got a product shipping notice from Amazon and the arrival date was more like six weeks in the future. Well, I had experience with variable arrival dates of Chinese product before, so I didn’t pay much attention. I was busy with other projects and didn’t need it right away.

Three weeks passed with no delivery. Checking the Amazon site, it showed my product in transit, but the shipper could not provide detailed information on the product location. Again, I had experienced this before, so no problem. Four weeks passed and then five, and I was starting to be concerned. I emailed the company again, and was relieved to get a quick response. Yes, they had shipped the inverter, unfortunately they had shipped it to the wrong customer. No wonder it wasn’t here! It was time to think about cancelling the order. Through emails to the company, it became clear that a mistake had been made, but they intended to make it right. If I could wait just a little longer, I would get the product. I gave them two weeks to get it here, or I would cancel the order. They said they would, and would sweeten the deal with the free addition of a battery temperature sensor (BTS, a $40 value).

All’s Well That Ends Well

Well, to my surprise and relief, the inverter/charge controller arrived two weeks later. It was carefully packed inside of wooden panels as advertised on their site, appeared to be in perfect condition and included the BTS. I wouldn’t be able to test it until we got back from the Red Power Roundup, but I now had the most important piece of the system. Next: batteries and putting it all together!

Weber Piano Action Rebuild – On to the Wippens!

photo of Applying was to rocker capstan screw holes

Applying was to rocker capstan screw holes

Another trip to Bernard’s piano shop and another list of tasks. I installed most of the rocker capstans (#2 in the diagram above) with the new square-drive screws at home, and used a little wax softened with mineral spirits to avoid damaging the keys. It took about an hour at the shop to finish them.

photo of installing a rocker capstan

installing a rocker capstan

Now it was time to cut the key end felt to size (shown below #14 in the diagram above). It comes as a long strip of felt about 2-tenths of an inch thick and has to be trimmed to the correct width. Because it is relatively thick, scissors are not an effective tool — they would distort the felt too much. Instead, we trimmed the felt along its length using a band saw.

photo of rocker capstans in place

rocker capstans in place

photo of cutting key end felt

cutting key end felt

Now the long strip of felt must be cut into individual pieces using a tiny little guillotine. This trims the felt strip into rectangular pieces that fit each key from side-to-side. Just to complicate matters a little, the keys are not all the same width. Celeste told me it was sometimes necessary to make felt pieces in three different widths. The Weber needed only two widths, and I’ll install them with hide glue at home in my shop.

photo of sizing key end felt

sizing key end felt

photo of bug damage on hammer rests

bug damage on wippens hammer rests

Next, Bernard discussed the wippens with me (#3 in the diagram, but also commonly includes #5, #9 and related parts). Bernard argued for the replacement of them with a more modern design, but this would have necessitated the replacement of the rocker capstans with capstan screws (#2) and a significant re-engineering of the action. Since the existing wippens are in relatively good shape and I wanted to preserve the period aspects of the instrument, my job would be to clean them, check for free motion on the various pivot points and replace insect-damaged hammer rests (blue felt in photo but shown as the red object in diagram between #3 and #10).

photo of measuring action center height

measuring action center height

It is finally time to order hammers and shanks (#10) — new parts that will have a big effect on the quality of the sound! It’s more than specifying a model number though, especially for an old piano like the 1893 Weber. Ideally, the hammer should strike when the hammer shank is exactly parallel to the string. Also, there is a “sweet spot” for the hammer to strike along the string’s length. We started by test fitting a new shank in place of an old one to confirm the hammer flange (#6) size. Then we needed string heights and the distance from the action center (pivot point between shank #8 and flange #6) to the key bed, as well as the bore distance of each hammer. I measured the string height before we started the rebuild, but we carefully measured each of the other dimensions. We confirmed the distance from the action center to the hammer as well — 5 inches instead of the more common 5 1/8th inches.

In most pianos, strings are not all parallel to each other, but instead cross to conserve space. In the Weber, the longer bass strings from notes 1 to 20 cross the remaining strings at about a 25º angle. This means the hammers must also be angled on the shank for the hammer to be parallel with the length of the string. We measured the angle for each section of the piano using the cut marks in the old hammers. Surprisingly, the hammers from note 21 to note 60 were not angled, and were not hitting the string in parallel. Did this explain some of the “tubbiness” in that range of notes? Might angling them properly further improve the piano’s tone? Perhaps. With all the measurements in hand, Bernard ordered hammers and shanks. The finale of the action rebuild is a little nearer . . .

photo of string cuts are not parallel to strings

string cuts are not parallel to strings

Weber Piano Action Rebuild – Keyframe with Refurbished Keys

I’ve been making progress on the Weber Project, but with a trip to New Mexico and various Spring activities, the pace has lagged. We’re back on track though. I mentioned in the last post that more work was needed on the bushings — they were too tight for the keys to move freely in the keyframe. This was because I had used too much material in each mortise, and it was binding. Initially we thought I might be able to cut some of the excess material away, but after struggling with it for a while, I decided to remove the new bushings and start over. Since I had done it once already it should go faster, and I wanted them to be right. With reasonable care and a little luck, they might not need replacement for another 100 years!

photo of Applying Windex to the bushing felt

Applying Windex to the bushing felt

This time I used the Windex method to remove the bushings instead of steaming them out. Taking a section of keys and clamping them together, I used a craft glue applicator to dose each bushing with a little Windex. It dissolves glue in a few minutes, and they pull out easily. This is much faster and easier than using steam.

Photo of keys clamped together

It helps to clamp the keys together

Bushing thickness varies from piano to piano, determined by the mortise and pin sizes. We found that two pieces of bushing felt (one for each side of the mortise) should be .072 inches. I had used all of the .036 inch bushing felt in my first attempt, so we decided to substitute two uneven sizes: .045 and .027 — totaling .072 inches. The other main difference with the first attempt was making each bushing as long as the width of the mortise — about 1/4 to 3/8ths of an inch. With the keys clamped together just like the removal process, the installation went very fast. I could complete 20 keys, both balance rail and key pin mortises, in about an hour. What a difference a little practice makes!

It takes a little time for the hide glue to cure, so I let them rest for 15-30 minutes before trying the keys out in the refurbished keyframe. At first, the installation was disappointing because the keys seemed just as stiff as the first time. They began to loosen as the glue continued to cure however, and by the next morning, most of the keys moved smoothly and easily within the keyframe. When I returned to Mollberg’s shop a few days later with all the keys installed, I was relieved to hear both Bernard and Celeste pronounce the work good enough and within “easing” range on the few sticky keys. The re-bushing was finally complete.

photo of Rail supports repaired with CA glue

Rail supports repaired with CA glue

Next, I re-laminated the plywood supports for the wippens and hammer rails using cyanoacrylate glue and accelerator. CA glue is perfect for this purpose as it easily penetrates the separated wood layers and cures almost instantly with the accelerator.

photo of Installing back rail felt

Installing back rail felt

With the danger of dripping CA glue past, I applied new lime-green back rail and red underlayment felt. The red felt lays under the back rail felt which is attached with a single line of glue along the front side of its length. Using hide glue makes this replaceable.

Before installing new key end felt, Bernard advised me to test fit the keyframe in the piano and confirm the spacing between the key end and the damper lever. With a sample of key end felt in place, the key should move about a third to a half of it total travel before engaging the damper lever. Changing the key end felt thickness is the easiest way to use the existing damper adjustment and thus simplify the damper timing regulation. Also, with the keyframe in the piano, I can confirm the fall board and end block clearance and that the keyslip covers an appropriate amount of the front ivory. If  they don’t, I can change the balance rail punching to correct the key height on a few sample keys.

photo of A damaged sharp

A damaged sharp

Finally, some cosmetics. A few of the sharps have divots where the soft wood below the ebony has been damaged, and these will be filled with a colored epoxy.

photo of Loose ivory

Loose ivory

Next, a few loose ivory veneers will be repaired using hide glue with whitening. Finally, the keys will be treated with wood bleach to remove the age darkening. No one will see this but me and a few other technicians, but hey, I’ll enjoy stripping away the years. I’ll also clean up and bleach the wooden rocker pieces.

photo of The old rocker screws

The old rocker screws

Lastly, I need to order replacement screws for the rockers. They were originally installed using #8 by 1 1/4″ round head slotted screws, and these are very difficult to adjust in the tight clearances of the piano action. Spax makes a torx-drive cabinet screw that looks promising. I’ll order some to try. If they fit well, they would be a real improvement and greatly simplify the regulation process.

If I have time this week, I’ll also install the new back checks. The keyframe and keys are nearly done. Next, we move on to the wippens rail!

photo of the New back checks

New back checks

Solar Powered Rainwater Project – Solar Panels Installed!

After designing the system and ordering parts for the project, the new Renogy 250-watt solar panels arrived very quickly. I was eager to install them, but they were just a bit larger than I could handle safely on the garden shed roof. Fortunately, I didn’t have to wait long, and my neighbor Jason was available a few days later. I had given the mounting some thought, and was ready when he arrived.

photo of Jason securing the first panel

Jason secures the first panel

The panels are about 65 by 39 inches, and weigh 40 pounds. The day was overcast, and not too windy (we didn’t want to mount a “sail” on the roof) so the work went quickly. The only hitch? One of the stainless nuts for mounting the aluminum “Z” brackets was SAE instead of metric like the others. We had to re-tap it to the 10 mm metric size. Not a big deal.

animated GIF of hoisting the second solar panel

Hoisting the second solar panel (click to view)

It was great to have Jason’s help, and we tied the center “Z” mounts into a rafter with 3-inch deck screws. The other mounting brackets didn’t line up with the 16-inch-on-center rafters, so we backed them up with a 2 by 2 making a “sandwich” of the roof decking.

photo of Jim securing the top mounts

Jim secures the top mounts

Although the brackets and screws seated tightly on the shingles, I’ll go back later and cover the mounting screws with roofing cement to prevent leaks.

photo showing generated power

Making power!

Once mounted, it was time for a test. Although the day was heavily overcast, a multimeter showed the expected unloaded voltage from each panel — between 33 and 34 volts DC. Connected in series like they’ll be used in our system, we measured over 67 volts. This will drop to around 60 volts under load. 500-watts of captured power!

We’re probably at least a week away from the charge controller/inverter delivery, so I’ll start buying the golf cart batteries. Eight will be needed for the 48-volt system and at around 60 pounds each it might be best to buy a few at a time. I’ll also have to reinforce the shelving in the garden shed to support the weight. Progress!

photo of the garden shed -- Ready for inverter and batteries

Ready for inverter and batteries

Solar Power Rainwater Project – Panels and Angles

Major equipment purchase done, it’s time to think about the installation. The solar panels will be arriving in a day or two, and I plan to install them right away. Solar panels produce maximum power when the sun is perpendicular to the panel surface. We talked about solar insolation in a previous post, but the sun moves. There are several strategies for dealing with this pesky movement, from doing nothing to 2-axis tracking systems. Or course, the sun moves across the sky, but it also changes elevation during the year. The simplest and most commonly-used strategy is to fix the panels in a single direction year-round. A worthwhile improvement in solar collection can be made by changing the panel elevation two or four times a year. This is relatively easy to implement and will improve solar collection by about 5-6%. The next improvement is to have a 2-axis tracking system. This is an expensive solution and probably not worth the investment on a small system like ours. It might result in a 20% improvement in array efficiency, but simply adding another solar panel can do the same and at much lower cost.

For example, the latitude here at Roy Creek Ranch is almost exactly 30º. Using figures supplied by we learn that the optimum year-round tilt is 25.9º which results in 71.1% of optimum panel insolation. By adjusting the panel elevation twice a year (6.9º in the summer and 45.5 º in the winter) we improve the insolation to 75.2% — about 5%. Adjusting the elevation four times a year (3.3º summer, 27.1 spring/fall and 50.7 winter) only offers a slight improvement to 75.7%. Just .5% more. With the tilt fixed to the winter angle (50.7º at 30º latitude) we get relatively even power throughout the year varying from 91% to 107% of winter insolation. Interestingly, the summer setting produces the least power with this arrangement. Since the need to pump water doesn’t vary much by season, the fixed winter elevation is probably best for our application.

With these figures in mind, we turn to the actual roof where the solar panels will be installed. The building is a 10-foot by 14-foot “salt box” garden shed. It has a peaked roof along the 14-foot axis with the smaller roof facing almost due south. Perfect. This roof is about 57-inches by 107-inches so we can fit four 250-watt (64 1/2-inches by 39-inches) panels. Now we need to know the roof angle or pitch.

photo of a framing square

Framing Square

There are three relatively easy ways to measure the roof pitch. The first way is from the building plans. Roof pitches are generally specified in terms of “rise” and “run”, with the “run” usually specified as 12-inches and the “rise” determining the pitch. In this case, the shed roof (both front and back) has a rise of 8 and a run of 12 (an 8/12 pitch). Charts can be found on the Internet that convert rise and run to degrees. They show an 8/12 pitch is equal to 33.75 degrees. The rise and run can also be measured with a framing square, as shown in this photo.

photo of a speed square used as a protractor

Speed Square protractor

A speed square or protractor can also be used to measure the pitch. A piece of string and a small weight such as a bolt or washer will turn the speed square into a protractor and the pitch can then be read directly. Note that the speed square reads 56º in this picture because it is measuring relative to the vertical plane. Simply subtract the reading from 90º (90º – 56º = 34º — very close to 33.75º) to convert that reading to the horizontal plane.

photo of Using the iPhone Compass app as an Inclinometer

Using the iPhone Compass app as an Inclinometer


Finally, and most simply, you can use a smart phone to measure angles. The current iPhone has a compass app (an excellent way to navigate in the woods!) and will also measure angles on the vertical and horizontal plane. Simply launch the compass, swipe left to access the second page of the app, and measure the roof angle directly.

So the garden shed roof is about 34º. Given the relatively small gains from adjusting the elevation of a solar array, I think I’ll simply mount the solar panels directly to the roof. That will optimize for the winter months and should result in relatively even insolation throughout the year. In this case, ease of installation is the overriding factor. One other thing — the panels get hot sitting in the sun all day, and that reduces their output. By leaving an air gap behind, convection draws fresh air behind and cools them. The panels will be mounted slightly separated from the roof.

Next: Installing the Solar Panels

Graphic or Solar Rainwater System

Solar Rainwater System Diagram

Solar Powered Rainwater Project – to Grid-Tie, or not to Grid-Tie. (that is the question)

Previously, we discussed the amount of power needed, and how that determines the size of the solar array. We also mentioned Days of Autonomy and how that informs battery storage. Now for Charge Controllers and Inverters — the brains that tie everything together.

There are two categories of charge controller, Pulse Width Modulation (PWM) and Maximum Power Point Transfer (MPPT). Both have advantages. PWM charge controllers are generally less costly, and are best for smaller (under 200 watt) systems. They work by turning the direct current (DC) from the solar panels into a stream of pulses that vary in width and thereby control the amount of power transferred. This is great for small systems, but there is a cost — some of the power from the panel is wasted as the controller interrupts power to safely charge batteries.

MPPT charge controllers on the other hand, are more efficient. By transforming the power from the solar panels into the optimum voltage and current for the load, they allow the panels to work at their most efficient voltage — usually several volts higher than the load needs. There is a closer match between power supply and load. Unfortunately they cost more. For larger systems though, the added cost is more than compensated by the increased efficiency. I decided on an MPPT model.

Now to convert power from the solar panels and batteries into Alternating Current (AC). Inverters do this but there are some special considerations. Normally, one would simply match the load, in this case less than 1000 watts, and get an inverter with at least that capacity or perhaps a little more. Pumps use capacitor-start motors however, and although my 3/4 horse pump may have a running load of only 900 watts, it’s starting current can be 5 to 7 times that while the motor spins up. Also, they run best on true sine wave power. Cheaper modified sine wave inverters may work, but the motors operate less efficiently and generate extra heat. In time, I want to run the well pump also. It has a 1 1/2 horse motor. With a run-time load of over 2000 watts, it might briefly need as much as 10,000 watts to start up.

There are quite a few inverters on the market that will supply the required power, but I was put off by their expense. Then I spotted a model from Sun Gold Power. It came in several power capacities, and combined a 40-Amp MPPT controller and a charger in the same unit. Why a charger? The price was attractive though, so I gave it a serious look.

It turns out the charger made the Sungold unit capable of Grid-assist. I wanted the rainwater system completely separate from the grid, because grid-tie systems are disabled during a power failure. This seems counterintuitive, but it prevents feeding power back into the grid and endangering electrical workers who aren’t expecting live power in the dead system. It’s an important safely feature, but it defeats my purpose: to have water no matter what.

Grid-assist systems work differently. Instead of feeding excess power back to the grid, that power is used to charge batteries. With a Grid-assist system, if the panels can’t produce enough power or there’s no sunshine, the grid can be used to run the system. The inverter can be set to prefer solar/battery power, so use of the grid is limited. It also acts like an uninterruptible power supply — loads are always powered. It can even start an outboard generator should both the solar and grid power fail at the same time. Also, unlike grid-tie systems, no approval is needed from the power company for installation.

So here’s what I bought:

I plan to buy eight golf cart batteries from Costco (about $900) with an initial storage capacity of 48 volts at 205 Amp-hours. I’ll have to buy some wire and other hardware, probably under $50. Total cost for the system: just under $3000. There’s room for expansion as well. I could add another pair of solar panels increasing the collection capacity to 1000 watts. Batteries could be doubled in capacity. I’ll hold off on those upgrades until I see how we do.

Next post: planning the solar panel installation.

Graphic or Solar Rainwater System

Solar Rainwater System Diagram

Solar Powered Rainwater Project – Panels, Batteries and Inverters — Oh My!

In a previous Project post, we determined our power needs – 1300 watt-hours per day to pump water. Now we can calculate the battery bank capacity,  solar panel wattage, and inverter capacity needed for our installation. Let’s start with the panel wattage. As I mentioned in the last post, the immediate goal is to power a 3/4 horsepower shallow-well pump, with a little power for the ultraviolet filter that cleans the water for house use. The filter uses 30 watts of power, so by multiplying that times 24 hours, we get 720 watt-hours. When added to the pump power consumption of 1300 watt-hours per day, our total is 2020 watt-hours per day. That’s what we have to collect from the sun.

How much sun is hitting the roof each day? Tables on the Internet show the insolation, or amount of solar radiation reaching a given area at various latitudes. The tables show that San Antonio (nearest datapoint to our home) averages 5.3 Kilowatt hours per square meter per day. Unfortunately, insolation varies during the year, and since we want our system to function even in the winter months, we’ll use the low figure for San Antonio of 4.65. That way, we will have plenty of power year around. The final factor is efficiency. No system is 100% efficient, with inverter losses, wire losses and even dirt on the panels playing a factor. It’s safe to assume an efficiency of around 67% for our system. Now we apply the formula to determine panel wattage:

  • (Watt-hours/day / insolation) / efficiency = panel wattage.

Inserting our numbers we get: (2020  / 4.65) / .67 = 648 watts of panel collection.

Since we need to pump water around the clock, we’ll have to supply power even when it’s dark. That’s where the batteries come in. There are several variables when figuring the battery bank capacity, including system voltage (12, 24 or 48 volt), watt-hours needed, days of autonomy, inverter efficiency, discharge limit, and temperature compensation. Let’s start with the watt-hours per day and inverter efficiency. We already determined a need for 2020 watt-hours per day for our pump and UV filter, but that will have to be converted from the DC voltage (12, 24 or 48 volts) to AC (120 or 240 volts). Inverters are about 92% efficient, losing about 8% of the applied power in the conversion process. This results in a formula:

  • Watt-hours per day / Inverter efficiency = Watt-hours per day of storage

Substituting numbers: 2020 / .92 = 2195 Watt-hours of storage per day

What if the sun doesn’t shine for several days? Here we build in “Days of Autonomy,” or the number of days the system can go without any sun. Three days is about right for our location, but your conditions may vary. Also, temperature can effect battery performance. A battery’s rated capacity is normally stated at 77 degrees Fahrenheit, but the same battery might fall to 80% capacity at 50 degrees. Finally, a battery can be discharged to it’s rated capacity, but only at the cost of drastically reducing it’s life. Normally  batteries are only discharged to 50% of their rated capacity to prevent damaging them. Another formula:

  • Average daily Watt-hours * Days of Autonomy * Battery Temperature Multiplier / Battery Discharge Limit = Battery Bank Capacity in Watt-hours

Substituting numbers: 2195 * 3 * 1.19 / .5 = 15,672 Watt-hours of storage

Finally, we need to determine the battery bank capacity in Ampere-hours. What system voltage you choose will depend on several factors, but generally speaking higher system voltages (like 24 or 48 volts) are preferred for high-wattage inverters. This is because we gain efficiency in the DC wiring and can buy smaller gauge copper wire to save money. Put another way, higher voltages require lower currents to achieve the same wattage and have lower resistive wire losses. I’m going to set mine up for 48 volts using this formula:

  • Battery Bank Capacity in Watt-hours / System Voltage = Battery Bank Capacity in Ampere-hours

Substituting: 15,672 / 48 = 327 Ampere-hours of Battery Bank Capacity.

How many batteries is that? Let’s start by imagining how they might be configured. We could just get four, 12 volt, 327 amp-hour deep-cycle storage batteries, but those are expensive batteries and I would rather use the more commonly available (and cheaper) golf cart batteries. They are typically 6 volt, 190 Amp-hour. Assuming we don’t want to have more than two parallel strings (to reduce trouble with charge equalization) can we get the required capacity? Once again, some formulas:

  • Battery Bank Capacity (Ah) / Maximum number of parallel strings = Minimum battery capacity

Substituting: 327 Ah / 2 = 163.5 Ah Minimum battery capacity. Yes. Since the golf cart batteries are 190 Ah, that will work.

Since we need to supply 48 volts, how many batteries will be in series?

  • DC System Voltage / Battery voltage = Number of batteries in each series string

Substituting: 48 / 6 = 8 batteries in each series string

Doubling that for two series strings, and we will need 16 total batteries. That’s a lot of lead.

Now we have some idea of solar panels and batteries needed for the project. In the next post, we’ll consider inverters, charge controllers and related equipment.

Graphic or Solar Rainwater System

Solar Rainwater System Diagram