Grid Tied Solar

I have been interested in solar power since I purchased my first house in Sacramento in 1995. I became aware of Sacramento Municipal Utility District’s (SMUDs) PV Pioneer program at that time, and eagerly applied to become one of the first to have SMUD install a retrofit system on my roof. However, my application was denied for failing to meet the rather stringent requirements at the time; my little house in Colonial Village didn’t have the right roof. I looked into the program again in 1998 after we moved to Elk Grove, but this time my tile roof prevented my participation in the program, as standardized mounting methods on tile roofs weren’t then available.

Now almost a decade later, with significant changes in the solar industry coupled with financial incentives available for this emerging renewable, I felt it was time to act. Continuous improvements in collectors, inverters, and mounting methods have made owner-installed systems achievable for people willing to invest some time in learning how these systems work and who are interested in meeting some of their energy needs through environmentally friendly means.

SMUD continues its residential solar retrofit program. Unlike in years past, the program today is fully contractor driven, meaning that SMUD no longer installs, but partners with participating contractors who bid, install, and service these systems, and apply to SMUD for incentives, passing these savings onto the homeowner. After soliciting bids for my retrofit, my own curiosity led me to seek out how these systems are designed (I am, after all, an electrical engineer), and how they are mounted. Quite comfortable with residential repair and building work and that electrically these PV systems are not outside my ability to understand, I decided to give it a shot. If it worked as I hoped, I should be able to build a 2kW system for less than $3.50/watt installed.


Design

The system I designed uses 12 Sharp 165 watt polycrystalline PV modules and a Sunny Boy 2100 watt inverter. This gives me a peak DC output rating of 1,980 watts. Sunny Boy inverters have been widely used in Europe, whose citizens have embraced solar technologies far more than in the U.S. Decades of experimentation led to the development of a string inverter that allows for vastly simpler systems compared to PV systems installed even just a decade ago.


Additionally, Sunny Boy’s U.S. affiliate is in Grass Valley, CA, only 40 miles from Sacramento. Because the inverter is such a key component and quite complicated, having a local supplier was important to me. This Sunny Boy is a grid-tied inverter only, with no provision for a battery back up. Having dependable electrical service from my local utility means the additional cost of battery backup does not make financial sense. As an aside, it also is worth noting that I work for SMUD as an engineer supporting the high voltage transmission system, so lack of dependable electric service is simply not an issue...My contacts with distribution engineers and residential services personnel at SMUD were instrumental in my decision to attempt this self-install.

The first step was to figure out how to mount as many PV panels as I could on my south-facing roof, to take full advantage of the sun. Unfortunately, most of my roof surface faces east and west, and while I could have much more easily mounted them there, I would either not be producing power at peak, or would do so only after 12:00 PM. Both east and west orientations provide only 88% of the annual energy a south facing system will produce. Considering that PV systems are still quite expensive relative to conventional energy sources, it’s important to maximize their efficiency. I only have 200 sq. ft of south facing roof above the garage at a 4/12 pitch (19º), on which I could fit 12 165-watt collectors (panels). Because this area is trapezoidal, I would have to stray from standard mounting methods. I needed more mounting rails that would otherwise be required if I could mount them in either a 6x2 or 3x4 pattern.

Shading

It is very important to keep PV panels unshaded. Even small shading can have a dramatic effect on system output. In my design, all 12 panels are connected in series (voltage additive) providing for a single input into the inverter. As each PV panel is rated at 34.6 volts and 4.8 A current (~165 watts DC), connecting 12 in series will produce 415 volts at peak output but the current is still 4.8 A, which flows through every panel. As even one section of a single PV panel is shaded by a tree, chimney, or other obstruction, the single panel will produce less voltage, but more critically will conduct less than 4.8 A, which has the effect of limiting the entire array. In some cases, 20% shading can lead to 50% reduced output.

Shading was a tough consideration for me. I’ve spent the last 9 years tending to a pin oak tree that had finally grown large enough to provide excellent shade, but alas, would need to be removed to make the PV system work. Ironically, this tree was planted as part of SMUDs Shade Tree program in 1993. At least there is no issue with removing the other 4 trees planted from this program that now provide a tremendous level of cooling for our house. If I had to guess, the yearly reduction in cooling costs from the trees alone is likely three quarters of my entire expected annual PV energy output, which is to say there is a cost-benefit ratio of 200:1 between PV and planting shade trees. Maximize tree shading first before considering PV. We have about as many trees as our tract will support. To compensate for the loss of the pin oak, we planted a red oak on the northeast corner that should provide nice shade near the end of the solar panel life.

Roof

I have a 15 year old concrete tile roof with no problems. A big concern for me was that I had planned an additional 3 lbs/sq. ft. dead load on roof trusses that are likely near their design limits with such a heavy roofing material, and with Elk Grove considered a category C wind area, live loading is substantially increased to more than 23 lbs/sq. ft. I decided that if the city building department had any structural issues with my plan, such as requiring structural engineering, I would avoid that expense and hassle and just re-roof that section over the garage with composite singles and add the PV standoffs during the re-roof.

One of the considerations for installing my own system is that the City of Elk Grove, just one month earlier, announced a partnership with SMUD that called for the waiving of permit fees, using a standardized application packet, and following a 24-hour permit review and inspection process. As I applied for my building permit that month, the City had yet to adopt a standardized application, which really meant that Elk Grove had not yet developed standards for rooftop PV installations. For comparison, the City of San Jose requires that PV modules be installed no more than 18 inches off the roof surface, and that the system be less than 4 lbs/sq. ft. (along with other requirements), but as far as I could determine I had no such requirements to meet. Unsure how Elk Grove was going to handle my application, I set my design standards to meet those of San Jose, and for any other city or district whose requirements I could find. This lays bare one difficulty; a lack of overall design guidelines that would standardize installs across the nation. It was hard enough to find information regarding the physical installation, let alone ensuring that my local authority would accept my design. This consumed an inordinate amount of my design effort, and in retrospect, probably wasn’t needed.

However, the application was painless; I spent no more than a half hour at the building department, and was issued my permit the following day. I submitted an overall sketch, a single line drawing (detailing the electrical conductors/components/conduit used), structural calcs for the roof and wires and cut sheets for the major components. There was no issue with live or dead loading increases on my roof. The permit fee would have been $430. Other authorities across California have also adopted or are considering similar partnerships.

Mounting System

Early on I decided to use UniRac PV mounting supports. While I could have saved by building my own racks using unistrut or similar material, I wanted to ensure I ground the system properly and also because UniRac provides for standoff systems with flashings that work well with tile roofs. However, the cost of the mounting system was roughly $950, 7% of the total cost, so it is not insignificant. This figure includes the mounting rails, clamps to hold down the modules, grounding lugs and clips, standoffs, and flashings. PV modules produce less voltage and current as cell temperatures increase. Sacramento Valley summers are hot and dry, with roof temperatures exceeding 85°C. It’s important to maintain airflow between the roof surface and the modules for cooling, which is why they are always ‘stood off’ from the surface. Standoffs provide for a minimum of 3” clearance and are screwed to the roof trusses directly, bearing the weight of the entire system. I elected to install ten rails using twenty 6” standoffs (two per rail), and I spent the greater part of two weekends cutting the roof tiles and mounting the standoffs. I am confident that the UniRac flashings will not develop leaks. The entire system is installed over the garage and not a living area, so even if I botched it, I joked with my wife, only my tools and I would suffer.

After the strongest wind storm in 55 years hit in January 2008, where my panels and mounts survived peak gusts of 75mph without incident, I am confident in the UniRac system. It should not expect to see such wind speeds again over the design life of this system.

Panels – Electrical

I chose Sharp PV modules that are a nominal 24 volts each. Most modules today come with multi-contact (MC) connectors that simply plug into the next module. This is the elegance of string systems, in that no complicated DC wiring is necessary. Because the voltage of PV modules is greater with decreasing temperature, I also needed to look to the lowest temperature expected in the Sacramento valley, as this condition will give the highest system voltage on a cold, sunny day. The Sunny Boy and connected DC wiring is constrained to 600 VDC, so there is a limit to the number of PV modules than can be connected together in series. If the string voltage exceeds 600 VDC under any condition, then the modules must be split into parallel and series combinations to limit the voltage, but other considerations for overcurrent protection must then be provided as the current in each branch and into the inverter will increase. So the choice of inverter and panels is not arbitrary. I first determined how many panels I wanted, and then chose the inverter based on its ability to accept any correct combination of series and parallel connections. Because the inverter also needs a minimum voltage to operate, most inverter manufacturers provide web-based string sizing programs to determine what combinations are acceptable. This Sunny Boy will accept 12 Sharp 165 panels in a single series string for expected temperatures in the Sacramento Valley.

Wiring

As mentioned before, if I had to parallel up two or more strings into the inverter, I would have to provide a means of overcurrent protection for each parallel connection, and then combine them into a single conductor to run to the inverter. But as I have only a single string, I avoided the need for fuse protection and combiner boxes by ensuring that I size the conductor to handle the full short circuit current. With a maximum panel short circuit current of 5.5 Amps, I sized the conductor to handle 125% of this current, with an addition 125% safety margin (the 1.56 rule) or 8.6 Amps. I used 10 AWG stranded copper conductors that will more than provide for this as well as minimize voltage drop on the DC side. Voltage drop must be considered. In my design, the garage roof is right above my service entrance, so I have only a fifteen foot run for the DC connection from the PV panels into the inverter, and a five foot run for the AC to the main service panel. These short distances do not cause any appreciable voltage drop.Safety considerations call for a DC disconnect switch between the PV array and the inverter. I used a three-pole lockable 600 VDC switch that will allow for future expansion if I need to increase the number of strings run into the inverter(s). Indeed, the Square D HU361RB is a common disconnect in PV systems as there are very few manufacturers making 600 VDC UL listed switchgear.

The AC side of the inverter is wired through a separate array meter, an AC disconnect, then into 15 amp 240V circuit breakers in the service panel. During a July 2006 heat storm, our original 100 Amp main service panel caught fire from overload, prompting the install of a new 200 Amp service panel. As it turns out, this provided additional breaker handle capacity that would have been required anyway. As I had yet to finish the stucco around the new panel, I had virtually no issues with PV wiring. I installed all conductors in raceways as exposed high DC voltages pose a serious safety hazard.



Grounding

Grounding the PV system proved to be the most challenging aspect of the design. To be sure, there are quite a few on-line resources available for PV design in the U.S., the foremost being the Southwest Technology Development Institute (SWTDI) in New Mexico, a renewable research and development center. They have made available a large archive of information that is quite readable. Still, I had many questions regarding how my particular system should be grounded.

U.S. installations are almost always grounded; one pole of the array (either positive or negative) is bonded to ground. However, European systems are typically ungrounded, and considering that inverters are universal, this led to confusion on how I needed to connect it. In discussions with the manufacturer I discovered that the negative pole of the array is referenced to ground inside the inverter. Because this is a roof mounting on a dwelling, I am required to have a ground fault detection and interruption circuit, which the Sunny Boy provides internally. This also constitutes the single DC bonding point between the array and ground.



Secondly, all PV systems need an equipment ground, whether or not one pole of the array is grounded. This is a grounding connection between each solar panel frame and rail, used to provide a safe path for current to flow to ground in the event of a fault. I sized the equipment grounding conductor the same as the PV conductors to handle the maximum anticipated short circuit current should a fault occur. The equipment grounding conductor is connected to a protective earth (PE) terminal inside the inverter.

NEC 250.166 called for an 8 AWG or larger grounding electrode conductor, yet all the on-line inverter manuals only indicated a single protective earth connection that can handle up to a 10 AWG wire. Upon inspection the inverter indeed has a grounding electrode conductor terminal sized to fit up to a 6 AWG connection. The grounding electrode conductor runs unspliced to the grounding bus bar in my main service panel. I ran a separate AC equipment ground from this grounding bus bar to the exposed metallic AC enclosures. The main service panel already has a 4 AWG grounding electrode conductor to the grounding electrode, the UFER, which is where the DC and AC systems are bonded, although the actual connection is at the grounding bus bar in the main service panel. I chose not to employ a separate grounding electrode for the array, as lightning is not common in the Sacramento Valley.



Click here for a larger image of the electrical schematic.

Inspection

I called for my electrical inspection, and to be honest, was hoping the inspector would validate my self installation. I had spent inordinate time and energy ensuring the system was built to code, but I still had (and have) reservations about grounding. Indeed, it was his first solo inspection of a PV system. The inspector’s reaction upon learning it was owner installed immediately raised a warning, as it was clear his aim was to lean on a licensed installer’s experience. It became immediately evident that I much better understood PV systems.

My goal was to pass inspection, so I offered little. The inspector was able to ensure correct wiring methods, but beyond that, did not review conductor ampacities, system grounding, overcurrent protection, required disconnects, or labeling. This was a disservice, not because I question my own install, but because there will likely be many more owner installs to come, and not all will be as diligent. And having had the opportunity to review dozens of schematics and installs by licensed contractors both before and after my own, I now know that minimal scrutiny leads to installers that cut corners. This is a harsh statement but I know how this works. I am much more appreciative of why building and electrical codes exist. PV systems will produce power for 40 years; they should all do so safely.

Throwing the Switch

After verifying that the inverter was receiving 240V from the utility side, and that I had correct DC polarity, I closed in during the late afternoon. Surprisingly uneventful, the inverter ran through its checks and produced 2 kWh that first day. I assumed that the system would not be producing power under incident light. It develops sufficient DC voltage well before and after direct sunlight on the panels, although the energy output is certainly minimal. The AC power delivered will drop to 1 watt before the inverter finally shuts down for the night.

System Performance

The first full day of operation in mid-April produced just over 8 kWh, on a cooler but very sunny day. As my system is oriented south at a 19º angle, my output drops gradually until the beginning of summer as the sun continues to rise in declination.



Financial Considerations

With the price of professionally installed PV systems ranging from $7 to $12 per watt, they are not competitive with utility service. Indeed, while my system size is limited by my roofline, a 2kW system is roughly the size I need to completely eliminate me from top tier energy rates. Considering SMUDs rates are still among the lowest in California, financially it doesn’t make sense to offset 100% of our energy with a larger system. I expect to never reach tier 3 (the highest level in SMUD), and should considerably reduce my tier 2 charges.

That PV is still an emerging technology, the system costs are indeed high but are currently offset by federal tax credits, CEC or utility rebates, and the City of Elk Grove waiving the permit fee. Additionally, a self-install doesn’t incur SMUD engineering costs when applying for the rebate. All told, these constitute a 40% reduction to the total cost. The utility rebating is particularly interesting. If I lived in an area served by the three investor owned utilities in California, I would apply to the Energy Commission for these emerging incentives. As a municipal utility, SMUD is not a party to CEC rebating, yet independently developed a rebating structure based on the CECs program. Without such programs PV systems would be far less attractive. One source of frustration is that as an owner-builder, I am not entitled to the same incentive (currently $2.50/watt) than if I contracted the work through select solar installers. I am instead limited to $2.15/watt. I have yet to see any reasoning behind this, as there is no incremental difference on the CECs part whether a solar installer or an owner-builder applies for the rebate. As SMUD has adopted the CECs program for equipment certification and incentive levels, it appears I’m not going to win this battle.

Based on expected system output and future rates, I expect my system to pay for itself in 16 years. If I lived across the river serviced by PG&E, at my current usage I would expect the break-even point to be half that. With possible future time-of-use metering and personal consumption changes, I may be able to reduce the break-even point even further.

However, the cheapest kWh is the one that is never used. For every dollar spent to reduce consumption, there is a savings of three to four dollars over PV generation. It makes far better sense to spend money to reduce electricity usage, such as installing CFLs, timers, or energy star appliances. Solar energy, combined with efficiencies that I’ve already made, provides me an economical system that is a good example for others. Our family of four uses 900 kWh a month on average. I am hopeful to reduce that by 145 kWh per month from solar energy. One apparent inequity in my calculations is that if we were only using tier 1 energy to begin with (i.e., we were even more energy efficient and conservative), I might never reach break-even. For this, I am much more in favor of European feed-in standards, encouraging energy efficiency along with fixed solar incentives that are persuading many Europeans to install PV.

Self Installation vs. Contractor Installations


Hopefully the preceding sections illustrate the time and effort involved in a retrofit. It is uncommon for individuals to take on the responsibility of self installation. If a contracted installation is more attractive, consider the information here as a supplement to the growing knowledge base available in print and on the web. Also consider that the contractor is not the one living in your solar home, so it pays to be actively involved in the planning and design phases. Knowledge of how these systems work will go a long way to understand what is being sold, and to ensure you get a system that works for you.

Personal Goals

With a conservative 16-year projected payoff, photovoltaics make little financial sense. However, I’m committed to more than just savings. I hope that my PV installation will:

Reduce carbon emissions by 1.6 tons per year.
Make a small dent in reducing our dependency on imported oil.
Serve as an example for others, that ordinary homeowners can either contract or self-install their own systems.
Allow me to keep any Renewable Energy Credits (REC) that I generate.

I have long been a SMUD greenergy subscriber, paying a premium for electrical service provided by renewable resources. I commute by bicycle and bus four times a week. All told, I have reduced my total energy consumption by almost 40% from a few years ago with hardly any sacrifice. Even modest reductions in consumption, if applied on a larger scale, could drastically improve our quality of life. There are much less obvious changes that can have an impact as well; refer to several excellent resources listed in the Contacts section.

Contacts

SMA America, Inc. 12438 Loma Rica Drive, Grass Valley, CA 95945 · 530.273.4895 http://www.sma-america.com/ · Sunny Boy Inverter

Sharp Solar Systems of America · 630.378.3357 http://www.sharpusa.com/ · PV panels

UniRac, Inc. Albuquerque, NM USA · 505.242.6411 http://www.unirac.com/ · PV Mounting Systems

Southwest Technology Development Institute (SWTDI), New Mexico State University, Las Cruces, New Mexico, USA· 505.646.1049 http://www.nmsu.edu/~tdi/· PV Design Information

Sacramento Municipal Utility District (SMUD), 1601 S Street, Sacramento, CA, 95817 USA· http://www.smud.org/ · Shade Tree Program, Solar Residential Retrofit Program

California’s Solar Initiative · http://www.gosolarcalifornia.ca.gov/ · CEC rebating, Solar Information