Perspectives on PV Blog Home All Blogs RSS
The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous “Perspectives on PV” articles for easy downloading. A color copy of the latest version (1.93) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

 Search all posts for:

Top tags: Perspectives on PV  Featured  January-February 1999  January-February 2005  January-February 2006  January-February 2007  January-February 2008  January-February 2009  January-February 2010  January-February 2011  January-February 2012  January-February 2013  January-February 2014  July-August 2004  July-August 2005  July-August 2006  July-August 2007  July-August 2008  July-August 2009  July-August 2010  July-August 2011  July-August 2012  July-August 2013  March-April 2005  March-April 2006  March-April 2007  March-April 2008  March-April 2009  March-April 2010  March-April 2011

## PV Math

Posted By John Wiles, Thursday, January 01, 2009
Updated: Monday, January 21, 2013

As we look at the PV array in a PV system, we find that many installers and inspectors are confused by the new system voltage calculations that may be required by the Code specific to PV systems. Code fine print notes (FPN) also address voltage drop that may be applied to the dc wiring from the array to the inverter. This article will cover both of those subjects.

Photo 1. Cold weather increases module open-circuit voltage

PV Math—Module Open-Circuit Voltage

A PV module or a string of series-connected modules has a rated open-circuit voltage (Voc) that is measured (and labeled) at 25 degrees Celsius (C) [77 degrees Fahrenheit (F)]. This voltage increases from the rated voltage as the temperature drops below 25°C. It is necessary to calculate this voltage at the expected lowest temperature at the installation location to ensure that it is less than the maximum input voltage of the inverter and less than the voltage rating of any connected cables, switchgear, and overcurrent devices (usually 600 volts). Since parallel connections of strings do not affect the open-circuit voltage, the number of strings connected in parallel is not involved with this calculation.

Section 690.7 in the 2008 NEC requires that the open-circuit voltage (Voc) of a PV array be determined at the lowest expected temperature at the installation location where module temperature coefficients are available. In previous editions of the NEC, Table 690.7 could be used to determine a multiplier that was applied to either the module or string (series connection of PV modules) rated Voc. The table can also be used under the 2008 NEC where module temperature coefficient data are not available.

The rated Voc is measured at 25°C (77°F) and is printed on the back of the module and in the technical literature of the module. To use the table, all one has to do is to determine the lowest expected temperature, look up the factor from the table for that temperature (which ranges between 1.02 at 24°C to 1.25 at -40°C), and multiply the factor by the rated Voc.

For example, a module has a Voc of 35 volts and is going to be installed where the temperature dips to -17°C. The factor from Table 690.7 in the 2008 NEC is 1.16 and the cold temperature Voc for this module is 35 x 1.16 = 40.6 volts.

If 12 modules were going to be connected in series, the string Voc in cold weather would be 12 x 40.6 = 487.2 volts.

We could also calculate the string voltage at rated conditions first and then apply the temperature factor. In this case, the 12 modules in series would have a string open-circuit voltage of 12 x 35 = 420 volts at 25 degrees C. Then we apply the 1.16 factor and get 1.16 x 420 = 487.2 volts; the same answer as before.

While the table is still valid and has been refined with 5°C increments, new modules may have different technologies than the silicon module technology used to develop the table.

NEC-2008 Requirements Differ

Table 690.7 is based on an average type of crystalline PV module that has been the most widely used over the last thirty years. However, we now have modules with different internal types of PV cells, and the table may not apply very well to these newer modules. Section 690.7 in the 2008NECrequires that where the module manufacturer’s temperature coefficients data are available they will be used. These temperature coefficients are found in the technical literature of nearly all modules and can also be obtained directly from the manufacturer. Unfortunately, different manufacturers present the temperature coefficients in two different forms.

Percentage Coefficients

One way of presenting these data is to specify them as a percentage change, and they are expressed as a percentage change in Voc for a change in temperature measured in degrees C. Note that the temperature used is a change in temperature from the rated 25°C.

For example: The Voc temperature coefficient is given as

-0.36% per degree C or -0.36% / °C.

The module has a Voc of 45 volts at 25°C (77°F) and is going to be installed where the expected lowest temperature is -10°C (14°F). Because the temperature coefficient is given in degrees C, we must work in degrees C. The change in temperature is from 25°C to -10°C. This represents a change in temperature of 35 degrees. The minus sign in the coefficient can be ignored as long as we remember that the voltage increases as the temperature goes down and visa versa.

If we apply the coefficient, we can see that the percentage change in Voc resulting from this temperature change is

0.36% / °C x 35°C = 12.6%.

This percentage change can now be applied to the rated Voc of 45 volts. And, at -10°C, the Voc will be 1.126 x 45 = 50.67 V.

Eleven of these modules could be connected in series and the cold-weather voltage would be 11 x 50.67 = 557.37 V, and that voltage is less than the 600-volt equipment limitation.

Millivolt Coefficients

Other PV module manufacturers express the Voc temperature coefficient as a millivolt coefficient. A millivolt is one, one-thousandth of a volt or 0.001V.

A typical module with an open-circuit voltage (at 25°C) of 65 volts might have a temperature coefficient expressed as

-240 mV per degree C or -240 mV / °C.

If we install it where the expected low temperature is -30°C (-22°F), then we have a 55°C degree change in the temperature from 25°C to -30°C. Again, we must work in degrees Celsius since that is the way the coefficient is presented.

Millivolts are converted to volts by dividing the millivolt number by 1000.

240 mV / 1000 mV/V = 0.24 volts

The module Voc will increase 0.24 V/°C x 55 °C = 13.2 volts as the temperature changes from 25°C to -30°C.

The module Voc will increase from 65 volts at 25°C to 65 + 13.2 = 78.2 volts at the -30°C temperature.

Let us suppose that the inverter maximum input voltage was listed as 550 volts. How many modules could be connected in series and not exceed this voltage? We take that maximum inverter voltage of 550 volts and divide it by the cold-weather open-circuit voltage for the module of 78.2 volts.

550 / 78.2 = 7.03 modules and the correct answer would be seven (7) modules.

7 x 78.2V = 547.4V

Eight modules could not be used because the open-circuit, cold-weather voltage would exceed 550 volts.

8 x 78.2V = 625.6V

Expected Lowest Temperature?

Where do we get the expected lowest temperature? Normally, this temperature occurs in the very early morning hours just before sunrise on cold winter mornings. The PV modules are, in many cases, a few degrees colder than the air temperature due to night-sky radiation effects. The illumination at dawn and dusk are sufficient to produce high Voc, even when the sun is not shining directly on the PV array and has not produced any solar heating of the modules. A conservative approach would get weather data that show the record low temperatures and use this as the expected low temperature. Other data show more moderate low temperatures associated with the data used to size heating systems. However, these data are not widely available. The National Renewable Energy Laboratory (NREL) maintains data on a web site that shows the record lows for many locations in the US.

http://rredc.nrel.gov/solar/old_data/nsrdb/1961-1990/redbook/sum2/state.html

Local airports and weather stations may have historical data on low temperatures

Also, weather.com has some of these data on file accessed by zip codes

http://www.weather.com/weather/climatology/monthly/zip code

PV Math—Module Short-Circuit Current

In most silicon PV modules, the module short-circuit current does increase very slightly as temperature increases, but the increase is so small as to be negligible at normal module operating temperatures. It is normally ignored.

Fine Print Notes—Voltage Drop

Fine print notes are not part of the Code—at least until the AHJ reads them, and then they become part of his or her personal code.

In the common, utility-interactive PV system, the PV array may operate at a nominal 48 volts to voltages near 600 volts. With nominal, peak-power, and open-circuit voltages to deal with, the installer and inspector are sometimes in a quandary as to how to calculate the voltage drop from the PV array to the inverter.

The utility-interactive inverter will normally operate in a manner that keeps the array voltage near the peak-power voltage (also called the maximum power point). While this voltage can vary with temperature, and temperatures vary considerably, using the rated maximum power point voltage and current of the modules results the easiest method of calculating voltage drop.

A typical PV array may have a single string of ten modules in series connected through 200 feet of 10 AWG USE-2/RHW-2 conductors to the inverter. The maximum power point numbers for the module are:

Vmp = 55V Imp = 5.5 amps, where the subscript mp means at maximum power.

For a single string of 10 modules, the string maximum power point numbers are:

Vmp = 550V Imp = 5.5 amps.

Table 8 in Chapter 9 of theNECgives conductor resistance per 1000 feet at 75°C.

For an uncoated, stranded 10 AWG conductor, the resistance is 1.24 ohms per 1000 feet.

The total conductor length (both ways) must be used in the calculation and this is 400 feet.

The resistance for 400 feet of a 10 AWG conductor is 400 / 1000 x 1.24 = 0.496 ohms.

The current at the maximum power point is 5.5 amps. Voltage drop is found by multiplying this current by the conductor resistance:

5.5 x 0.496 = 2.728 volts.

Expressed as a percentage, 2.278/550 x 100 = .496% or about 0.5% and that is much less than the FPN recommendation of three percent for most circuits. Of course, the losses in the PV dc disconnect were not counted, but they are typically less than one percent on these circuits.

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.eduPhone: 575-646-6105

A color copy of the latest version (1.8) of the 150-page,Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published inHome Power Magazineover the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## A Top to Bottom Perspective on a PV System

Posted By John Wiles, Saturday, November 01, 2008
Updated: Monday, January 21, 2013

Photovoltaic power systems can be examined in a number of different ways as we have done in the last few years in the "Perspectives on PV” series of articles. In this article and the next few articles in the series, let’s start at the modules at the "top” of the system and progress through the system to the grid interconnection at the "bottom.” A utility-interactive PV system is a series-connected system, so where we start is not important and if you are in a hurry for information on some part of the system that we have not gotten to, you can review past articles in the series that are archived on my web site.

The PV Array—Mechanical Considerations

The PV array consists of individual PV modules attached to a mechanical frame, usually attached to the structural members of the roof in a typical rooftop-mounted residential utility-interactive PV system. Although not an electrical-code issue, some attention must be given to the attachment of the PV array to the building structure.

Most roofs in recent years have been built using span tables in the building codes or using trusses designed by professional engineers. PV arrays may add up to 4–5 pounds per square foot of dead weight to the roof structural members, and that weight will be concentrated through the rack mounting feet. Also, because the PV array is mounted above the roof some distance (zero to six inches or more), the roof may be subjected to both uplift and down-force wind loadings—again concentrated through the mounting feet of the rack. If the roof has several layers of old shingles under the array, the structural limit of the roof may be approached. Leaving as many as two layers of old shingles in place is a common practice during re-roofing, so we can assume that the basic roofing structure has a safety factor allowing the extra load of old shingles or the PV array, but possibly not several layers of shingles and a PV array.

Photo 1. Array rack attachment point - used in dry climates

Array racks must be attached to the structural elements of a roof (trusses or rafters), and this will require penetrating the roofing surface material in a manner that is weatherproof for the life of the PV array or the life of the roof—whichever is shorter (see photo 1).

Stainless steel hardware is usually used to connect the modules to the racks. Galvanized hardware is frequently used to bolt the racks to the roof. In both cases, corrosion resistance is a must in most climates.

PV Array—Electrical Requirements

Photo 2. Module with attached cables and connectors

Electrically, the PV array consists of PV modules connected in series using exposed single-conductor cables with "finger safe” connectors (see photo 2). The conductors are typically USE-2 as allowed by NEC Section 690.31. In the 2008NEC, a new PV Wire (a.k.a. PV cable, photovoltaic wire, or photovoltaic cable) is also allowed. This conductor is a "super” USE-2 that has a thicker jacket (the conduit fill tables cannot be used), passes a 720-hour accelerated UV test (is marked Sunlight Resistant), and has the flame and smoke retardants of RHW-2. It can be used under and within the PV array for the module interconnections and in raceways in other locations. This new cable will soon be appearing on all modules because it facilitates the use of ungrounded PV arrays and transformerless inverters (lower cost, less weight, higher efficiency) (NEC690.35).

Photo 3. Modules in landscape orientation

Although the electrical connectors attached to the ends of the module cables are "finger safe” when new, if they are opened under load, the dc arc may damage the insulation and the connectors may then pose a shock hazard. Therefore, there are new requirements in 690.33 for locking connectors in the 2008NEC. A tool will be required to open these locking connectors. They will also soon be appearing on most, if not all, PV modules, although they are only required when the PV array wiring is operating above 30 volts and is readily accessible (690.33).

Photo 4. Modules in portrait orientation

Another 2008NECrequirement that applies to readily accessible PV source and output circuit conductors operating at over 30 volts is found in 690.31(A). These conductors must be installed in raceways. Unfortunately, as mentioned above, most PV modules do not have junction boxes with knockouts that would accept a raceway. They come with permanently attached exposed, single-conductor cables and connectors with no provision for attaching a conduit or other raceway. Fortunately, most residential rooftop PV arrays are not readily accessible. A few manufacturers can provide conduit-ready modules on special order, but many module manufacturers have no such option.

The solution, as noted in the 2008NEC Handbook, is to make this wiring not readily accessible by placing some sort of barrier behind the modules that prevents the wiring from being touched without removing the barrier. Fences with locked gates may not be a solution, because a basic maintenance requirement for the readily accessible ground-mounted PV array is keeping the grass mowed—a task usually done by people not qualified to be near PV or other electrical systems.

Photo 5. Pipe clamp used to secure module conductors

The conductor leads attached to the modules are 40 inches long or longer to allow the series connection of modules when they are mounted in a landscape orientation (see photo 3). When the modules are mounted in portrait orientation (see photo 4), the excess lengths of conductors must be securely fastened against the module racks to resist abrasive damage due to wind, sleet, and ice. Many use plastic cable ties, but unless they are of very high quality, they may not last the required 40 years or more when exposed to the extremes of heat and ultraviolet radiation from sunlight. Some people use a stainless steel pipe clamp (loop strap) with an EDPM insert (see photo 5).

Section 690.74(D) requires that the PV array metal surfaces be connected directly to earth via a separate grounding electrode. This requirement provides a greater degree of lightning protection for PV systems than other Code requirements provide. This new requirement is in addition to the normal equipment-grounding conductors that run with the circuit conductors and which are connected to earth (grounded) at locations remote from the PV array. If the array is on the same building that contains the inverter and the existing ac grounding electrode, then the array may be connected directly to that electrode and a separate electrode is not required. If the connection to an existing electrode requires a horizontal extension greater than six feet from the closest earth contact point, a separate electrode is required. This new array-grounding electrode does not have to be bonded to any other electrode.

Photo 6. Grounding a metal roof -Oops, outdoor rated lug and wire needed

Module grounding has been discussed recently in previous articles and will not be covered here in any detail. Suffice it to say that the module frames must be effectively grounded, and that is not always easy with aluminum frames and copper conductors. Those single-conductor exposed module wires are bound to touch the roof, if not on initial installation, sometime over the life of the system. The racks must also be grounded, and if the PV array is mounted on a metal roof, that metal roof should be grounded (see photo 6). Rodent damage and abrasion could very well cause the roof to be energized (see photo 7).

Photo 7. Rodent-damaged conductors

The single-conductor exposed wiring (USE-2 or PV Wire) is allowed only in the near vicinity of the PV array to interconnect the modules and to return the end of the string conductor to the origination point of the string wiring. At this point, the exposed wiring must transition to one of the more common wiring systems found in chapter 3 of the Code. Typically, this will be some form of conduit such as EMT. If the array output conductors penetrate the surface of the structure before reaching the first readily accessible dc PV disconnecting means, then they must be in a metal raceway where inside the structure. Metal raceways include the rigid metal conduits and flexible metal conduit (FMC), but do not include metallic cable assemblies like Type MC and Type AC cables. The transition fitting keeps water, dirt, rodents, and other material out of the conduit. A rain head or a cord grip might be used (see photo 8).

Temperature Corrections

Modules can operate at very high temperatures (70–80 degrees C), the exposed wiring will come into contact with the hot surfaces, and the conductors originate in the hot termination boxes attached to the backs of the modules. Field-installed wiring (and the leads connected directly to the module) must be evaluated for temperature and ampacity corrections applied.

Photo 8. Cord grip transition. Can you spot the violation - allowed by the local inspector?

In most locations in the United States, a 75 degree C temperature correction factor is suggested for conductors near PV modules that are mounted roughly four inches or less from a surface like a roof. The distance in not exact and is normally measured from the back of the module frame to the surface. Four inches or less is insufficient clearance to allow cooling air to flow behind the modules mounted in an array.

If the air space behind the modules is greater than four inches, then a 65 degree C temperature-correction factor is suggested. Again, these are not hard and fast numbers, and the individual installation location and microclimate (Death Valley or Nome) may affect them.

Photo 9. Conduits on hot roof

Conductors in conduit on roofs (and possibly elsewhere) in sunlight are also exposed to solar heating, and 310.15(B)(2) in the 2008NECprovides the temperature additions above the expected average high temperatures (see photo 9). These temperatures apply not only to PV systems but any conduits on the roofs of buildings exposed to sunlight. In many cases, where the high average temperatures are in the 40–45 degree range and the conduits are close (1/2″ or less) to the roof, again a 65–75 degree C temperature correction factor applies. Those PV circuit conductors are going to be delivering energy for the next forty years or more, so we really need to carefully apply these temperature correction factors to ensure that the insulation does not suffer premature degradation.

In the next article we will move away from the PV array and on to other parts of the system.

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.eduPhone: 575-646-6105

A color copy of the latest version (1.8) of the 150-page, Photovoltaic Power Systems and the 2005National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last ten years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## Are We Grounded Yet?

Posted By John Wiles, Monday, September 01, 2008
Updated: Monday, January 21, 2013

Photovoltaic (PV) systems will be producing hazardous voltages and currents for 50 years or more. Over that period of time, they may or may not be operational and they may or may not be maintained. Proper grounding of all exposed metal surfaces in the system that may be energized by internal faults, poor terminations or failing conductor insulation is one of the most important requirements in a code-compliant system. Even in a failing or failed system, maintaining all metal surfaces at ground (or earth) potential will minimize the possibility of electrical shocks. Those grounding connections must be maintained in a harsh outdoor environment where they are exposed to heat and cold, solar radiation (ultraviolet radiation and added heat), dirt, rain, wind, ice, sleet, and snow.

Photo 1. Aged PV modules - still producing power

Why is grounding PV different from grounding HVAC?

Heating ventilation and air conditioning (HVAC) systems are exposed to the same environmental conditions as PV systems, but there are significant differences in the grounding requirements and procedures between the two systems.

Photo 2. Failing PV module - still producing power

PV modules, with an active life measured in many decades, will be in place longer than the outdoor unit of a HVAC system. When the performance of an HVAC system deteriorates, it is usually inspected and repaired promptly. PV systems suffer gradual degradation that is not usually monitored, and the PV array may remain installed on the roof even after the system has been decommissioned or abandoned when the inverter fails—a common occurrence when sufficient funds are unavailable to make the inverter repair or replacement. (See photos 1 and 2).

Photo 3. HVAC outside unit, well maintained and with single grounding point

HVAC units have only a few separate pieces, the equipment grounding points are well marked, and the factory installed bonding jumpers and screws effectively bond all parts of the listed device together (photo 3). HVAC components are typically made of steel and the equipment-grounding terminals are electrically and chemically compatible with copper conductors.

Photo 4. Thousands of PV modules in a large array, each to be grounded

On the other hand, PV systems have numerous modules (tens to thousands) and mounting racks that must be individually grounded (photo 4). PV modules and mounting racks are typically made of aluminum and are not compatible with copper conductors. See "Perspectives on PV” in the September-October 2004 and March-April 2008IAEI Newsfor issues related to grounding aluminum-framed PV modules.

Are we getting better?

Inspectors throughout the country have been flooding Underwriters Laboratories (UL) with e-mails about poor PV module and PV system grounding techniques and equipment that they are seeing in the field (photo 5). And, as a result, UL is getting tough on grounding. In the fall of 2007, UL issued an "Interpretation” of the existing standard for PV modules (UL 1703). The current state of affairs with respect to grounding is in part due to a confusing section in UL 1703 that combines bonding requirements and instructions

Photo 5. Poor PV module grounding

(connecting the module frame parts together in the factory) with grounding requirements (installing the external equipment-grounding connection in the field) in one section. Methods and equipment used in the factory to bond the module frame sections together are evaluated during the listing process. However, the same level of scrutiny cannot be applied to field-installed, equipment-grounding methods, and the same parts and techniques used in the factory are generally not appropriate in the field.

The Interpretation clarifies the intent of the standard in several areas:

1. Dissimilar metals, like copper and aluminum, cannot come into contact with one another at the equipment-grounding connection point. A chart is provided showing numerous metals and which types can be in contact without galvanic corrosion problems.

2. Any threaded fastener used for grounding must pass the same durability tests as any threaded fastener used for other electrical connections. It must be fastened and unfastened ten times without damage to the threads. This requirement will probably result in the demise of the use of thread-cutting or thread-forming screws for module grounding because threads in soft aluminum cannot pass this requirement.

3. The module manufacturer must provide or designate the specific hardware and methods used to ground the module, and those instructions must be included in the module instruction manual. UL will evaluate the grounding hardware and methods throughout the entire testing and listing/certification process on new modules and also when existing modules come up for recertification.

UL is also working on changes to UL 1703 that will clarify the requirements, markings, and instructions for grounding PV modules. At some point, they will develop a separate standard that will allow the evaluation and listing of various universal PV module grounding methods and devices that will work with a number of different module frame geometries. The use of this standard will allow grounding-device manufacturers to meet the standard without having to be tested with each and every separate type of PV module.

In the meantime…

As theCoderequires, instructions and labels provided with a certified/listed product must be followed [110.3(B)]. The listing and certification process is slow, and modules only come up for review every five years. Therefore, it may be some time before all of the instruction manuals meet the clarified intent of UL 1703. And this brings us to the question of new grounding devices

New grounding devices

With respect to new PV module grounding methods and devices, such as clips and washers, the situation is somewhat murky. Of course, the local AHJ can call it as they see it and some jurisdictions have accepted these new devices.

As mentioned above,NEC110.3(B) requires that the instructions and labels provided with a listed product be followed. PV modules are marked for grounding at specific points. Hardware (when provided) and these instructions require the use of the marked points. The instructions do not generally address grounding the module at the mounting holes or at other locations.

A few manufacturers may have tech bulletins that show other methods. These tech bulletins may or may not have been reviewed by UL where they differ from the listed grounding points. UL is attempting to review new manuals and directions submitted by the manufacturer, but at times, the manuals get published without a proper review. Also, even if reviewed, they may not be in compliance with allNECrequirements or may show grounding techniques that have not withstood the test of time. The future UL Standard for PV Module Grounding Methods/Devices will evaluate the long-term durability and reliability of the various grounding methods and devices.
When using a new grounding method, other than a separate wire to each PV module, grounding continuity must be addressed. One of the oldest requirements in the Code is to make a grounding connection first and break it last [250.124(A)]. Consider a module with an internal ground fault to the frame. If the circuit conductors are left connected and the module is unbolted from the grounded rack (disconnecting the frame grounding first rather than last), the module frame may be energized with up to 600 volts to the grounded rack.

Photo 6. One module grounding method

A few PV module manufacturers have listed their grounding devices and racks with specific PV modules so they have a listed combination. Rack manufacturers also are developing grounding devices, but they are not associated or listed at the present time with any particular module.

See Appendix G in the latest version (1.8) of the PV/NEC Suggested Practices Manual for the grounding method we currently use at SWTDI (photo 6). This method is used only if it does not conflict with the module instructions and when those instructions allow the use of a properly listed lug attached to the marked grounding points after appropriate surface preparation has been accomplished.

I have long been encouraging module manufacturers to get their modules tested with these new grounding products and get that information into the instruction manuals, so the AHJs won’t have any questions and the installation will be code-compliant.

TheNECis not holding up any new grounding device or method and Section 690.43 ofNEC-2008 allows the use of these new devices as soon as they have been listed/certified and identified for the use and appear in the module instruction manuals.

We aren’t connected to Mother Earth yet

A related question that will eventually have to be addressed is: To what are these new grounding devices attached? It is necessary to first verify that they can make a durable connection with the module frame, and then the device must make a connection to an acceptable grounding electrode (such as building steel) or to an accepted equipment grounding conductor such as a copper conductor? Aluminum module mounting racks are not currently listed as equipment grounding conductors, but some of the rack manufacturers are getting such a certification/listing. This is necessary because the racks are typically designed for mechanical durability and not electrical connections. Joints may be designed to allow for thermal expansion and contraction, and with aluminum, such "slop” does not make for good electrical conductivity. As theCoderequires for loosely jointed metal raceways (250.98), a provision for electrically bonding the sections of the rack together must be incorporated into the design.

Where we want to be

In the future, modules will either come with an integral mounting rack (there are a few now, see photo 7) or they will be easily attached to a rack in a manner that provides both robust mechanical and electrical connections. One point on the rack will allow for the connection of the equipment grounding conductor for all modules and for a grounding conductor routed directly to earth (where required). Installations will take less time, will cost less, and will keep those module metal surfaces grounded ‘till the cows come home.

Photo 7. Integrated modules, rack, conductors, and grounding connections

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 575-646-6105

A color copy of the latest version (1.8) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published inHome Power Magazineover the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## Grid Interconnections – Then [2005] and Now [2008]

Posted By John Wiles, Tuesday, July 01, 2008
Updated: Monday, January 21, 2013

The final connection between the photovoltaic (PV) power system and the electrical utility grid is always an area of high interest to both inspectors and to the utility, because both agencies are responsible for safety. These connections vary significantly from PV system to system due to the size of the PV system and to the configuration of the existing service-entrance equipment. These variations are made more complex because of differences in Section 690.64 in the National Electrical Code (NEC) between the 2005 and 2008 editions.

Section 690.64(B) establishes the requirements for connections of the output of utility-interactive inverters on the load side of the main service disconnect. The key to understanding this section is in carefully reading 690.64(B)(2) and noting that the ratings of overcurrent devices supplying a busbar or conductor must be added so that the sum of these ratings does not exceed the rating of the busbar or conductor. Note that overcurrent devices supplying loads are not counted. Also note that the overcurrent-device (normally a circuit breaker) rating is used in this calculation and not the current flowing through the circuit. Overcurrent devices that would be counted are the main breaker and all breakers being back fed from utility-interactive PV inverters. We can use an equation of breaker ratings to express this requirement:

PV + Main <= Bus or Conductor

In the 2005NEC, this requirement applies to commercial (nonresidential/dwelling unit) PV installations. The requirement essentially says that if, for example, the site has a 400-A main service panel with a 400-A main breaker, then no (zero) PV can be added to the panel. In many commercial installations, this limitation forces the installer to a supply-side connection discussed below.

For residential installations a 120% allowance is added and to make the installations somewhat easier. The equation for residential looks like this:

PV + Main <= 120% Bus or 120% Conductor

In the residential example, a 200-A panel with a 200-A main breaker could have up to 40 amps of backfed PV breakers connected.

In the 2008NEC, 690.64 was rewritten and the 120% allowance was applied to commercial installations if an additional requirement was met. That requirement [690.64(B)(7)] says that the PV backfed breakers must be mounted at the opposite end of the bus from the main breaker or feeder. This location prevents overloading the busbar. If this requirement cannot be met, then the sum of the breakers will be limited to no more than the busbar rating on commercial installations. Note that the requirements apply to both the busbars in a panel or load center and to any conductor that is fed by overcurrent devices from multiple sources.

Figure 1. Section 690.64(B)(2) requirements in the 2005 NEC

In figure 1, the requirements in the 2005 NEC are applied to a multi-story building where a PV system requiring a 15-A circuit breaker is needed in a 100-A main lug panel on the tenth floor. This panel is fed through a 100-A breaker in a 400-A main lug panel on the fourth floor which is, in turn, fed by a 400-A circuit breaker in the 1000-A main distribution panel, which has a 1000-A main disconnect. Since this is a load-side connection, 690.64(B) applies to each panelboard and conductor supplied through an overcurrent device from multiple sources. To meet the requirement in the top floor panel, the panel would have to be removed and replaced with at least a 115-A panel. The feeder between the 100-A panel and the 100-A breaker would also be required to have an ampacity of at least 115 amps. If that top floor panel had a 100-A main breaker, then the feeder would need to be rated at 200 amps to meet the 690.64(B)(2) equation.

At the fourth floor panel, the sum of the rating of the breakers is 100 + 400 = 500 and this exceeds the panel rating of 400 amps. The panel would have to be replaced with a panel having a rating of at least 500 amps. The feeder between the fourth floor panel and the main panel would also have to be rated at 500 amps with a main lug panel. If that fourth floor panel had a 400-A main breaker, then the feeder would be required to be rated for 800 A. Now look at that 1000-A main service panel. The sum of the rating of breakers supply it is 400 + 1000 = 1400, which is significantly larger than the 1000-A rating. It needs to be replaced with at least a 1400-A rated panel.

Yes, life is tough and seemingly unfair, but these requirements were established in 1984 with the concept that they would protect those buses and conductors from overloads even when the PV system was enlarged, the panels had excessive loads placed on them or when the feeders were unknowingly tapped. The 2008 NEC provides some relief as shown in figure 2 and even more relief might come in the 2011 NEC.

Figure 2. Section 690.64(B)(2) requirements in the 2008 NEC

In figure 2, the 120% allowance is put in the calculations for this commercial installation as allowed by 690.64(B)(2) in the 2008 NEC. That 100-A panel on the top floor is OK, because 100 + 15 = 115 which is less than the allowed 120 amps. The same equation applies to the cable when the top floor panel is a main lug panel and the feeder does not need to be changed. If the top floor panel had a 100-A main breaker, then the equation for the feeder conductors would still be 15 + 100 = 115 <= 120 A, and the conductor would remain unchanged because a new sentence in 690.64(B)(2) requires that only the first overcurrent device connected to the inverter output be counted in subsequent equations.

At the fourth-floor 400-A panel, the allowance would be 480 A (120% of 400 = 480), but the additional rule in 690.64(B)(2) requires that only the first overcurrent device connected to the inverter output be counted in subsequent equations. The equation becomes 15 + 400 <= 480 and no changes in the panel are required. With a main lug 400-A panel, the same equation applies to the feeder to the main panel. Also, even if a 400-A main breaker were installed in that 400-A panel, then the cable ampacity would not need to be changed.

Even with the allowances in the 2008 NEC for the load-side connections in 690.64(B), many systems are large enough that ripping out existing load centers and feeders is required and that becomes costly. The supply-side connections allowed by 690.64(A) are used.

Supply Side Connections

The supply-side connection is essentially a second service entrance on the facility that is connected on the load side of the existing meter to allow for net metering. See "Perspectives on PV” in the September/October 2005 and January/February 2006 IAEI News for more details on the Code requirements for these connections found in Article 230.

Section 240.21 tap rules don’t apply to these service-conductor taps, because the 240.21 requirements were developed over a number of years for a circuit with currents flowing one way from a single source protected by an overcurrent device. The service-entrance tap with a utility-interactive PV inverter may have currents flowing both directions from two sources, and one of them (the utility) has very limited overcurrent protection.

Actually making the tap will depend on the type of equipment involved. Many load centers do not have adequate space to splice to the incoming service conductors. The same holds true for the limited space in meter sockets. In these cases, the supply-side tap will require that a new enclosure be added between the meter and the separate load center.

Photo 2. Much space but tap would violate the listing on the device

Combined meter/load centers like the one shown in photo 1 can only be tapped with permission and instructions supplied by the manufacturers. The cables and busbars (photo 2) may be exposed with plenty of room for the tap, but in most cases, the manufacturer will not allow them to be tapped because it would violate the UL listing on the device. To add a supply-side tap to this type of installation may require adding a new external meter socket and a tap enclosure before the existing meter. Then the existing meter is bypassed with an appropriate set of jumper bars.

Photo 3. Main-Lug-Only Panel

One situation that arises in many parts of the country is the dwelling that has a main-lug-only panel (see photo 3). There is no single main breaker feeding the panel, but up to six main breakers are allowed. Where these panels have one or more empty breaker positions, they can be used as a supply-side connection. The basic restriction (not in the Code—wait for 2011) that would apply to this type of main service panel is that the sum of the overcurrent devices from the PV inverter(s) not exceed the rating of the panel bus or the rating of the service-entrance cables.

Summary

Connections from the PV system to the utility are still somewhat complex. However, the requirements in the 2008 NEC have allowed smaller systems to be more easily connected in the commercial environment. In either residential or commercial PV installations, the requirements of the Code should be studied in some detail to ensure that a safe and durable system is planned and installed.

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 575-646-6105

A color copy of the latest version (1.8) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner ” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## Questions from the AHJ – To Fuse or Not to Fuse?

Posted By Laura Hildreth, Thursday, May 01, 2008
Updated: Monday, January 21, 2013

Nearly everyone agrees that the National Electrical Code gets better with every edition. However, new technologies like photovoltaic (PV) power systems and fuel cells are still evolving with new equipment, new wiring procedures, and new installation requirements being developed every week. With new inspectors and new installers coming into the field every day, questions are bound to arise. The question addressed below is very common and is frequently posed by both oldtimers and newcomers. The answer is not directly found in the Code but must be evaluated on a case-by-case basis by examining the system.

When are overcurrent devices (fuses or circuit breakers) needed in the direct current circuits between the PV modules and the utility-interactive inverter?

Before answering this question directly, we first should address the issue that properly rated fuses and properly rated circuit breakers are equivalent in this application and are collectively known as overcurrent protective devices (OCPD). This is true even though the required label on the back of certified/listed PV modules says "Fuse.” In general, PV arrays operating at dc voltages above about 150 volts (cold-weather, open-circuit voltage) may use fuses, and those operating below this voltage may be using either fuses or circuit breakers. These applications are due to the ratings, availability, and cost of the different devices.

In most electrical systems, theNECrequires that every ungrounded circuit conductor be protected from overcurrents that might damage that conductor. Overcurrent protective devices, either fuses or circuit breakers, provide that function. However, some of the smaller utility-interactive PV systems may not need OCPD in the dc circuits that are connected to the PV modules.

PV modules are current-limited devices, and their worst-case, continuous outputs for Code calculations are 1.25 times the rated short-circuit current. An exception to Section 690.9(A) allows conductors to be used with no OCPD where there are no sources of external currents that might damage that conductor.

The module series fuse requirement

Additionally, Underwriters Laboratories (in UL Standard 1703) has established that modules must have an external series OCPD if there are external sources of current that can damage the internal module conductors. The module can be damaged if reverse currents are forced through the module (due to an external or internal fault) in excess of the values of the maximum series fuse marked on the label on the back of the module (see photo 1). Again, if there are no sources of external currents that exceed this marked value, then no OCPD is needed to protect the internal module wiring.

Photo 1. Label on back of PV module showing series fuse rating

External sources of current vary from system to system. These external currents can originate from modules or series-connected strings of modules that are connected in parallel to the module of interest, from batteries in the system backfeeding through charge controllers, or from utility currents backfeeding through utility-interactive inverters. The material below will deal only with the utility-interactive PV system with no batteries in the system.

Where required, only one OCPD will protect all modules connected in a single series-connected string of modules [690.9(E)]. A properly rated and located OCPD can protect the modules and properly rated conductors from external overcurrents.

Utility-interactive inverters and backfeed currents from the utility

Many of the smaller utility-interactive inverters (below about 10 kW) are designed so that they cannot backfeed currents from the utility into array faults. However, there are currently no normal operation tests in UL 1741 to validate the lack of backfeeding from the utility, so a manufacturer’s certification should be obtained that the inverter cannot backfeed from the utility into an array fault. Yes, there are abnormal operation tests for backfeed, but theses tests do not rule out backfeed during normal operation of the inverter. Larger inverters and inverters designed for transformerless or bipolar operation may require additional certification that they cannot backfeed.

The most common case—systems with inverters that cannot backfeed from the utility

The most common situation occurs in systems where there are multiple strings of modules connected in parallel (see photo 1a). The non-faulted strings may be able to supply sufficient overcurrents (through the parallel connection) to damage either the conductors or the modules in the faulted strings.

Photo 1a. Large system with multiple strings of modules will require OCPD.

A basic question is, How many PV modules or strings of modules can be connected in parallel and still meet the National Electrical Code and Underwriters Laboratories requirements (marked on the back of each module) before an OCPD is needed on each module or string of modules?

UL marks the modules based on reverse-current tests as described above. The NEC requires that the manufacturer’s instructions and labels be followed [110.3(B)]. This is a maximum value for the OCPD. Lesser values can be used as long as they meet the NEC requirement of 1.56 times the module short-circuit current (1.56 Isc) to protect the conductor associated with the module or string of modules [690.8(A)&(B)].

In a few cases, module manufacturers have not met (or understood) the Code requirements, and the value of the module protective overcurrent device marked on the back of the module is less than 1.56 Isc (see photo 1). This poses a Code conflict 110.3(B) vs. 690.8,9 and is an issue for UL to rectify.

One string of modules

It is easy to see that in a one-string system, no fusing would be required since there are no external sources of overcurrents. An unfused dc PV disconnect would be used on this type of system (see photo 2). The maximum series fuse rating that is marked on the back of the module is at least 1.56 Isc and there are no sources of external currents that could damage the modules or the connecting cables (also rated at 1.56 Isc or higher).

Photo 2. Unfused dc disconnect

Now let’s look at a PV system with several strings of modules connected in parallel. Keep in mind that we are not determining the rating of any required OCPD, we are merely making some calculations that determine whether or not an OCPD is needed on each string of modules.

Two strings of modules in parallel

Consider two modules or two strings of modules connected in parallel, then connected to the inverter input. Each string of modules can generate a maximum of 1.25 Isc. If a fault occurs in one string, the second, unfaulted string can try to force 1.25 Isc amps into the faulted string. However, we know that the modules in the faulted string can withstand currents up to at least 1.56 Isc or higher (if their marked series fuse rating is higher), and the conductors have an ampacity of at least 1.56 Isc or greater. Therefore, with only two strings of modules, no currents exist in the PV array that can damage the modules or the wiring and no OCPD are required.

Three strings in parallel

Now let us consider a system with three strings of modules connected in parallel. A fault in one string could see currents from the two other unfaulted strings. Each of these unfaulted strings could deliver up to 1.25 Isc under worst-case conditions for a total of 2 x 1.25 Isc =2.5 Isc. Suppose that the module manufacturer had a value of the maximum series fuse marked on the back of the module of exactly 1.56 Isc and the wiring was sized at exactly 1.56 Isc. The currents from the two unfaulted strings at 2.5 Isc would be greater than the series fuse rating of the module and ampacity of the conductors at only 1.56 Isc and they could be damaged. Fuses in all three strings at a minimum value of 1.56 Isc would be required.

However, the module manufacturer does not usually have a marked maximum fuse value of exactly 1.56 Isc. Typically, the module will pass the UL reverse-current tests at a higher current such as 15 amps. As an example, let’s take a module that has a short-circuit current (Isc) of 5 amps and a marked value of the maximum series fuse of 15 amps. The interconnecting conductors between the modules must also have an ampacity of 15 amps, after the appropriate deratings for conditions of use have been applied if the conductors are to be protected. In a system with three series strings of this module, the two unfaulted strings could deliver 2 x 1.25 x 5 = 12.5 amps. Since this current is less than the 15-amp ampacity of the conductors and is also less than the 15-amp maximum series fuse requirement marked on the back of the module, no fuses are required because no damage can be done by overcurrents.

In another example, the module wiring is still 15 amps, as is the fuse rating marked on the back of the module. However, this module has a short-circuit current of 8 amps. The two unfaulted strings could send up to 2 x 8 x 1.25 = 20 amps. This 20 amps exceeds both the conductor ampacity and the ability of the module to withstand reverse currents, so fuses are required in each string of modules. The OCPD must be at least 1.56 Isc (1.56 x 8 = 12.48 amps) and not more than 15 amps. A 15-amp OCPD would normally be used.

Modules with low Isc and high series fuse ratings

Some modules have a low, short-circuit current and a high, maximum series fuse rating. For example, a module with a 1.5-amp Isc and a 20-amp maximum series fuse can have up to 11 strings of modules in parallel without any OCPD. The reader is encouraged to verify this—the author may be wrong.

As can be seen from these three examples, when more than two strings of modules are connected in parallel, a calculation should be done to see if the OCPD is required in each string. When three strings of modules are connected in parallel without fuses, the conductor ampacity may have to be greater than the normal 1.56 Isc.

Larger systems and possible inverter backfeed

If the inverter can backfeed utility currents into the dc PV wiring, the NEC requires that an OCPD be installed in series with the output of all strings (or modules) to protect the cables and the modules from reverse currents from any back feed of ac currents through an inverter. In many cases where there are fused combining boxes mounted at the array, an OCPD may also be needed at the inverter input, since we are assuming that the inverter is a potential source of overcurrents (see photo 3). This OCPD will have a minimum rating based on the number of strings connected in parallel on that circuit and the short circuit current of each string. This OCPD is sized to allow maximum forward currents from the array (all strings of modules) to pass through without interruption and to keep the overcurrent device from operating at more than 80% of rating.

Photo 3. Large system fused combiner box

Summary

Most utility-interactive PV systems with only one or two strings of PV modules will not require OCPD in the dc wiring between the PV array and the inverter. Systems with three strings or more will require a simple calculation to determine the OCPD requirements. Most current inverters rated at less than 10 kW are not able to backfeed currents from the utility into the dc wiring, but larger inverters and inverters that may be transformerless or designed for bipolar operation should be certified for no backfeeding. For a slightly more technical approach to these requirements and calculations, see Appendix J in the author’s PV Power Systems and the 2005 NEC: Suggested Practices manual (below).

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 575-646-6105

A color copy of the latest version (1.7a) of the 150-page,Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## Common PV Code Violations

Posted By John Wiles, Saturday, March 01, 2008
Updated: Monday, January 21, 2013
As we move into 2008, the PV industry continues to grow by leaps and bounds. New module and inverter manufacturers are entering the industry, and the number of individuals and organizations installing PV systems is growing right along with the demand. Numerous small 2 kW residential and large megawatt commercial PV systems are being installed in many states, and all need to be inspected for code-compliance. With new people entering the industry every day, the common code violations we have seen in the past will continue. Here are some of the most prominent ones that have been repeatedly observed throughout the country.

Photo 1. Grounded PV source circuits, but no white conductors

DC Module Wiring Color Codes

Back in ’97—that is 1897—when the first edition of the Code was being drafted, Thomas A. Edison was generating power. And it was direct current (dc) power, not that alternating current (ac) stuff with those heavy, costly transformers developed by Westinghouse and/or Tesla. AC came later, and the early Code dealt with direct current, including color codes for that dc power. If the conductor is a grounded circuit conductor, the insulation or marking on larger conductors must be white or gray. If the conductor is an equipment-grounding conductor, it must have green or green with yellow stripe insulation or be bare.

Those color codes apply to both ac and dc electrical systems. There is no special color code for dc systems. Nearly all past PV systems and those being currently installed are grounded systems, and one of the conductors in the dc parts of the system should be white. PV installers insisting that red is positive and black is negative are to be relegated back to their electronics workbenches where such color codes originated.

Yes, in the future, we will see the installation of ungrounded PV arrays (see 690.35) that will be used with transformerless inverters, and those systems will not have a grounded PV dc conductor. Then red and black conductors may become more common; but on the current grounded systems, they are incorrect. (See photo 1).

### Module Grounding

Module grounding still continues to be an issue with many inspectors, and rightly so, as the PV installers attempt to take time and materials short cuts when grounding modules. (See photos 2a and 2b). Underwriters Laboratories (UL) has issued an Interpretation of the UL Standard 1703 for PV modules in September 2007 that requires that the module manufacturer identify the grounding method and materials to be used in grounding the module. UL will then test and evaluate those methods and materials during the listing of new modules and the periodic recertification of existing modules. It is likely that the common use of a thread-cutting screw will not survive these new evaluations which require that all threaded electrical connections be installed and removed ten times without damage to the threads.

Photo 2a. Improper module grounding dissimilar metals

Photo 2b. Dry location lug in wet location

Photo 3. Module grounding hardware and tools

Until those more definitive instructions start appearing, NEC 110.3 requires that the labels and instructions provided with the listed/certified modules be followed. That usually means attaching a conductor or tin-plated copper, direct-burial lug to one of the four grounding points marked on the module frame. Attaching lugs properly is a time and materials intensive process, and it is hoped that new procedures and materials are approved quickly. (See photo 3).

### Enclosure and Conduit Grounding

Most utility-interactive PV systems operate at dc voltages between 300 and 600 volts. The metallic enclosures used for disconnects and source-circuit combiners must be properly grounded. TheNECdoes not recognize the use of sheet metal screws to ground enclosures (250.8), but PV installers and electricians continue to use them. (See photo 4). In listed safety disconnects, there is usually a label requiring the use of the appropriate listed, ground-bar kit to ground the enclosure. There are designated areas of the enclosure where the metal has been swaged thicker to allow two full threads of the thread cutting screw provided with the ground-bar kit to be cut into the enclosure. (See photo 5).

Photo 4. Improper grounding of enclosure; wrong device, wrong location

Photo 5. Listed ground-bar kit in the proper location

Failure to use the proper ground-bar kit would appear to violate 110.3(B) and could result in an enclosure that is not properly grounded.

Typical, residential, utility-interactive PV systems operate at voltages up to 600 volts. NEC 250.97 requires that metal conduits operating over 250 volts be properly bonded to the enclosures, particularly when concentric and eccentric knockouts are involved.

Photo 6. Bonding brushing on 500 VDC conduits

The large enclosures used for disconnects have not be evaluated for grounding/bonding where concentric or eccentric knockouts are used. (See photo 6).

### Disconnect Connections

The typical fused and unfused disconnects (a.k.a. safety switches) usually have the line terminals (usually the top set of terminals) shielded by an insulator so that these terminals, when energized by the source, cannot be easily touched when the cover or door is open. These disconnects normally have a mechanical interlock between the handle and door that requires that the disconnect be turned "OFF” before the door can be opened. With the disconnect "OFF,” the blade contacts and the lower set of load terminals are supposedly safe and are not energized. They are exposed and not covered with insulation. This works well when the only source of power is connected to the line terminals and loads are connected to the lower load terminals. However, PV systems with multiple sources of power and power flows confuse the issue somewhat.

Photo 7. Warning label for PV dc disconnect

The dc PV disconnect should have the line terminals connected to the incoming PV output conductors, and the inverter dc input should be connected to the load terminals on the disconnect. However, there are energy storage and filtering capacitors in the inverter that can energize the inverter dc input terminals and the disconnect load terminals up to five minutes after the disconnect is opened. These energized load terminals are the reason for the requirement in 690.17 for a warning label on the disconnect saying that all terminals might be energized when the disconnect is opened. (See photo 7).

Sometimes, installers (and inspectors) get confused when a safety switch is used as the ac inverter disconnect. These disconnects are frequently required by the local electric utility or may be part of the service-entrance tap for the PV system. The power flows from the inverter to the utility (usually through a backfed circuit breaker) and some installers and inspectors want the upper line-side terminals of the disconnect to be connected to the source of energy, the inverter. However, the normally energized conductors from the utility are the most dangerous and should be connected to the upper or line terminals of the disconnect. When the disconnect is opened, the inverter immediately ceases producing power and the load terminals and the blades of the disconnect have no voltage on them. Because the load terminals are de-energized when the disconnect is opened, there is no requirement for a 690.17 warning label on this disconnect when it is connected properly.

Improper Conductors

PV modules operate in extreme outdoor conditions where the temperatures on and near the modules may range from -40 to +80 degrees Celsius. There is always an abundance of ultraviolet (UV) radiation (remember, it comes from sunlight) and wind, rain, snow, and ice depending on location. NEC 690.31 allows single-conductor, insulated cables to be installed as connections between PV modules and from the modules to a transition box under the PV array where a more conventional wiring system starts. The use of the wrong conductors in exposed locations such as THHN/THWN, RHW, THW, or others that are intended for use in conduit will result in rapid deterioration of these conductors that have no UV resistance. (See photo 8).

Photo 8. THHN conductors deteriorating due to outdoor UV exposure

Conductors marked USE-2 with or without RHW-2 markings should be used for exposed module interconnections. Newer cables marked "PV Wire,” "PV Cable,” "Photovoltaic Wire,” or "Photovoltaic Cable” are coming to the market, and they too will be acceptable since they have superior sunlight resistance to USE-2 and a thicker jacket, plus some other good features. Where used in conduit (it has the necessary properties for that application), the conduit fill will have to be calculated manually because of the thicker jacket.

Summary

Photovoltaic power systems are a mature, but evolving, technology. While seasoned inspectors and PV installers are meeting Code requirements, there is a continual influx of new equipment and new, inexperienced installers. Inspectors must keep up with the new equipment installation requirements while maintaining a firm but fair vigilance for the Code violations that have been seen in the past and that will continue to be seen. Inspectors should also be vigilant for unexpected hazards—(photo 9).

Photo 9. Unexpected hazard

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. Phone: 575-646-6105

A color copy of the latest version (1.7a) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## Ground-Fault Protection for PV Systems

Posted By John Wiles, Tuesday, January 01, 2008
Updated: Monday, January 21, 2013

Once upon a time (the 1987 Code cycle) in the land of Quincy, a group of alchemists from a national laboratory was elaborating on the excellence of their photovoltaic (PV) test facility in the distant Land of Enchantment. They showed some senior firefighters a picture of a burned PV module that had been subject to a ground fault and had subsequently melted down. The alchemists failed to mention at the time that this was a prototype, unlisted PV module, that the module was on a concrete pad, and that ground faults in PV systems were somewhat rare. These firefighting pros said to themselves, "PV ground faults lead to fires. Fires on the roofs and in the attics of dwellings are very hard to fight.” They then told the PV industry to propose Section 690.5 for the 1987 NEC to require a ground-fault protection device (GFPD). The proposal was accepted and the requirement was established, but no hardware existed.

Photo 1. One-pole, ground-fault protective device for 48-volt PV system

In 1989, I joined the PV industry as a full-time employee at the Southwest Technology Development Institute. One of my first projects was to develop prototype hardware that could be used to meet the new Section 690.5 requirement. This effort was funded under contract to Salt River Project, a Phoenix, Arizona, electric utility. In the 1987 Code, the requirements for this fire-reduction device were to:

1. Detect ground faults in PV arrays mounted on the roofs of dwellings
2. Interrupt the fault current
3. Indicate that a ground fault had occurred
4. Disconnect the faulted part of the PV array
5. Crowbar or short circuit the PV array

Figure 1. Ground-fault currents go through the bonding conductor

The original GFPD prototype was developed in two versions that were similar except for voltage rating. The basic concept was to insert a 0.5 or 1.0 amp circuit breaker in the dc system-bonding conductor connecting the grounded circuit conductor (usually the negative) to the grounding system (the point where equipment grounding conductors and grounding electrode conductor are connected together). Any ground-fault currents must flow through this bond on their way from the ground-fault point back to the driving source, the PV module or PV array. (See figure 1). When the current in this bond exceeds 0.5 or 1.0 amp, the circuit breaker trips to the open position. This action interrupts the fault current, even when the fault is many feet away on the roof of the building and provides the indication that a ground fault has occurred. Requirements 1, 2, and 3 are satisfied by these actions.

This small circuit breaker is mechanically linked to one to four large, 100-amp circuit breakers and they open when the 0.5 amp circuit breaker opens. These added breakers are connected in series with each of the incoming ungrounded conductors from the PV array and when they open, the PV array is disconnected from the rest of the system, thereby meeting requirement 4.

Requirement 5 was added to reduce the PV array voltage to zero by shorting the positive and negative conductors together to minimize a potential shock hazard. In the original GFPD design, this was accomplished either by using a motor-driven circuit breaker on 48-volt systems or by using a solenoid-driven (closed) shunt-trip breaker on the higher voltage systems. This fifth shorting requirement was later removed from the NEC when it was determined that it might be possible to damage a "new technology” PV module by short-circuiting it. The module was never produced, but the crowbar requirement was not reintroduced even though the PV wiring can handle the worst case short-circuit currents and is oversized by a factor of 125 percent. It is an impressive demonstration when circuit breakers rated at 750 volts close and short circuit a 100-amp PV array that has an open-circuit voltage of 600 volts.

Modern Ground Fault Protection Devices

The early designs of the prototype GFPDs were released to the PV industry in 1991. Finally, in 1997, a GFPD was manufactured for the 48-volt and below PV systems, and that device used the exact design and components as the prototype. (See photo 1). Other ground-fault devices for the low-voltage systems soon followed as these off-grid, stand-alone systems became more common and were inspected more frequently. (See photos 2 and 3).

Photo 2. Two-pole, ground-fault protective device for 48-volt PV system

Photo 3. Four-pole, ground-fault protective device for 48-volt PV system

As the higher-voltage, utility-interactive PV inverters became available in the late 1990s, it was more cost-effective to use a 0.5 or 1.0 amp fuse as the sensing element and use the control electronics in the inverter to monitor the fuse, indicate that a ground fault had occurred (light or display), and shut down the inverter (effectively disconnecting the equipment). (See photo 4).

Coming in 2008

NEC-2008 will require GFPDs on nearly all PV systems including those mounted on commercial buildings (non-dwellings) and on the ground. This requirement was added to theNECbecause on the larger PV arrays, a ground fault, if not interrupted, can continue as long as the sun is shining, and may not be detected until significant damage has been done. The possible arc from the ground fault and the overloaded equipment grounding conductors may each pose hazards.

Sizing equipment grounding conductors usingNEC250.122 for PV systems with fuses does not always result in a conductor size that can withstand continuous ground-fault currents. The conductor and overcurrent sizing requirements for PV source and output circuits and the current-limited nature of PV module outputs cannot ensure that overcurrent devices will open properly in a very short time as they do on ac circuits. Therefore, a requirement was added to interrupt the ground-fault currents on all PV systems when they exceed the low 0.5 or 1.0 amp value.

Before May 2007, inverters larger than about 10 kW had only partial GFPD functionality. They detected the ground faults, indicated that the fault had occurred, and shut down. However, they did not interrupt the fault currents. Now, with a change in UL Standard 1741 for PV inverters and the 2008 NEC, all utility-interactive inverters will have full functionality with respect to ground faults and will act in a manner similar to the smaller residential units. Off grid, PV systems with batteries operating at 48 volts nominal, or less, will have the GFPD built into the charge controller, or an external device will have to be added. Small, dc off-grid systems that have no dc or ac wiring inside or on a building will be exempt from the GFPD requirement if the equipment grounding conductors are oversized by a factor of about two.

The AC Ground-Fault Issue

The common alternating-current ground-fault circuit interrupters (GFCI) are not designed to be backfed. The output of a utility-interactive inverter connected to the load terminals and backfeeding a receptacle or breaker GFCI, a 30-milliamp equipment protection ground-fault breaker, or even a 600–1200 amp main breaker with ground-fault elements may damage that device with no external indication of a problem. Anytime a utility-interactive PV system is installed, the entire ac premises wiring system should be examined all the way from the PV inverter output to the service entrance to ensure that there are no ground-fault devices in that circuit that may be subject to backfeeding. Some of the newest ground-fault breakers in the 1000 amp and larger sizes are listed as suitable for backfeeding, but that information must be obtained on any ac ground-fault device that could be subject to backfeeding.

Photo 4. Ground-fault fuse on high-voltage inverter

Arc-Fault Circuit Interrupters

Arc-fault circuit interrupters (AFCI) are, in some ways, similar to GFCIs and should not be backfed by PV inverters unless listed and identified for backfeeding. They are being required in many locations thereby increasing the safety of electrical systems here in the U. S. DC arc-fault circuit interrupters are not currently available.

However, we know that there is some danger associated with a line-to-line fault in the dc wiring of a PV array. The PV industry and Underwriters Laboratories are studying the issue to determine the signature of a typical dc arc originating from a PV system and how, if possible, to detect, control, and extinguish that arc. This is not an easy task because the electrical sources (the PV modules) in any system are widely dispersed and numerous.

Summary

The number of PV installations is increasing at more than 20% a year. Nearly all PV systems will soon be required to have a ground-fault protective device that will minimize the possibility of fires starting from ground faults in PV arrays. Efforts are continuing to enhance the safety of PV systems for the general public through revisions and additions to the National Electrical Code and UL Standards. The goal is to have safe, reliable, and cost-effective PV systems. The green future must be a safe future.

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 575-646-6105

A color copy of the latest version (1.7a) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this website: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the IEE/SWTDI web site.

## Why Inspect PV Systems?

Posted By John Wiles, Thursday, November 01, 2007
Updated: Tuesday, January 22, 2013

Photovoltaic power systems are a rapidly growing (30+ percent/year) segment of the residential and commercial electrical systems market. These systems operate up to 600 volts and, in the larger commercial systems, the dc and ac currents can range up to 1000 amps. These levels of voltage and current, if not properly managed, pose both shock and fire hazards. The electrical inspection is a key element to minimizing these potential hazards.

Previous articles in this "Perspectives on PV” series have covered the details of the Code requirements for these systems, and copies of those articles are available onwww.iaei.organd on the author’s web site—link is below. In a perfect world, all of those requirements have been fully met or exceeded, and the installation has been executed with exceptional workmanship.

Quality Systems Are Coming

Eventually, we may approach this ideal level of quality in PV installations. That could happen in a few years when the well-seasoned PV systems integrators, installers (hopefully also trained electricians), and electrical inspectors have worked together on hundreds of the same type of systems and installations. Those old-timers in the PV industry will be working with well-established products. They will have worked closely with the plan reviewers, permitting officials, and inspectors to submit a well-documented electrical design package that includes a clear, three-line diagram, the calculations used to determine ampacity, conductor types and conduit fill, and a copy of the equipment specifications and manuals. Inspections of PV systems will be as quick and routine as for any residential or commercial electrical system.

But Not All Systems Are Code-Compliant and Durable

Unfortunately, we are still a long way from that ideal scenario. While there are a few PV systems integrators (the larger companies) and other PV installers who have done dozens and possibly hundreds of PV installations, they are not common. PV installers, normally with little electrical installation experience, abound. They are familiar with neither Article 690 in the NEC covering PV systems nor the first four chapters of the Code that deal with the basics. On the other side of the installation/inspection equation, inspectors and plan reviewers have had little experience with the unique nature of PV systems and have not worked extensively with these new PV companies. New equipment (inverters and PV modules) is being introduced continually, and all involved with PV systems are hard-pressed to keep up with the ever-changing installation requirements due to the unique nature of each piece of equipment. Unfortunately, even a PV installer who has obtained the NABCEP (North American Board of Certified Energy Practitioners,www.nabcep.org) certificate by passing a 60-question written examination may not have extensive experience installing conventional residential or commercial electrical systems.

Safety First for the Inspector

This state of affairs should lead the inspector initially to conduct cautious, thorough inspections. Remember, there are old inspectors and there are bold inspectors, but there are few old, bold inspectors. When inspecting that first PV system from an unknown company/installer, personal safety (for the inspector) is a first consideration. Proper signage placed by the installer might indicate that attention was given to even the smaller details (see photo 1/photo 1A). But are that metal switchgear and inverter housing properly grounded? Will it be OK to touch that switchgear and possibly the inverter to check the workmanship on the connections? (see photos 2 and 3). It always pays to look at the external workmanship, grounding and bonding before touching anything.

Photos 1 and 1A. Required placards may indicate attention to details

There are many items that should be inspected on a PV system. See the "Perspectives on PV” articles in the IAEI News for March/April and May/June 2006 for a more complete list. We need to keep in mind that these systems are unique in that they will be producing significant amounts of energy for the next 40–50 years whenever they are exposed to sunlight. The inspector should evaluate the overall workmanship with this timeframe in mind.

Photo 2. Are they properly grounded?

Will the System Be Safe 10 Years from Now?

Will those exposed cables on the roof be secure and not allowed to move around in the wind or when ice slides under the PV array? Are the conductors the right type for the exposed conditions? If squirrels are scampering around the roof and near the PV conductors, possible insulation damage may be expected (see photo 4). No, we don’t have a solution for this problem, nor do we know how prevalent it is. Are the conduits securely fastened to the building? Do the conductor sizes and ampacities reflect appropriate temperature deratings for conductors in conduits in sunlight per 310.15(B)(2) inNEC-2008?

Photo 3. Must have been done by a "Grounding Guru"; violation of 250.53(G)

Are the covers on enclosures like junction boxes and combiners firmly attached with screws so that a tool is needed to open them? (see photo 5).

Inspectors in some areas are reporting that conductors can frequently be pulled loose from terminals because they have not been properly tightened. A few "pull tests” on unenergized conductors will reveal whether or not the installer used a torque screwdriver (see photo 6). Many installers (and electricians) don’t have torque screwdrivers, even though every electrical terminal has a torque specification and that torque specification should be followed for a durable electrical connection.

Photo 4. Rodent damaged cable and connector

If the system is ground-mounted or readily accessible (Code definition) from a readily accessible flat roof, have those exposed single-conductor cables been treated properly and made not readily accessible with a barrier? If not, the NEC-2008 will require that they be installed in a raceway, and that will be hard to accomplish since few modules have provisions for using conduit. At 600 volts, the general feeling is that the unqualified person should not have ready access to these conductors and the currently used, pull-apart connectors. Connectors for these exposed conductors will be locking and require some sort of tool to open starting sometime in 2008. The most likely solution in these readily accessible systems will be to put some sort of barrier behind them, possibly just a wire mesh or screen that would prevent ready accessibility.

Photo 5. How long has that cover been off?

After looking at the workmanship issue and the long-term durability of the system, the inspector can then concentrate on ensuring that the electrical components have been connected properly. For some reason, NEC 690.64 seems to be frequently abused. This may be because it is somewhat complex and the requirements hard to meet with the larger PV systems on residential services. See the "Perspectives on PV” in the September/October 2005 and January/February 2006 IAEI News for the full story.

Photo 6. Torque screwdrivers are needed for electrical connections

If the electrical transmission and distribution system suffers any long duration blackouts, then there may be an increase in the number of PV systems that, in addition to being connected to the utility grid, will also have batteries for energy storage. These systems require special inverters that can disconnect from the utility grid during the outage and supply part of or the entire house loads with power from the PV modules and/or the batteries. These more complex systems will require additional time for the inspections.

Summary

PV systems are like other electrical power systems. When they are installed incorrectly and not in compliance with the requirements of theNECand local codes, they can pose hazards, not only to the owners/users of the systems but also to inspectors. Inspections are definitely required. Teamwork between the designers, installers, and inspectors of these systems is a necessity. Teamwork coupled with increased familiarity with the equipment and the Code requirements based on experience will yield safe, durable, and reliable systems. Concern for the environment and significant financial incentives are producing significantly more photovoltaic power systems installations. A fast train is just leaving the station. Hop on board. The trip should be interesting

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 505-646-6105

A color copy of the latest version (1.6) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.htm

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the SWTDI web site.

## The Nature of the PV Module: Limited Currents Have Benefits and Drawbacks

Posted By John Wiles, Saturday, September 01, 2007
Updated: Tuesday, January 22, 2013

The currents in a PV system are somewhat different from the currents traveling through a typical alternating current (ac) electrical system. Yes, the PV system has ac circuits and they are somewhat like a typical ac load circuit, but the direct current (dc) circuits are a little unusual. This article will address the unique aspects of these dc currents and how the Code handles them.

Current Sources

While PV modules produce volts, amps, and watts, they are considered to be current sources and operate differently than the normal voltage sources commonly experienced in the 120/240-volt ac circuits in our homes or the 12-volt dc circuits in our automobiles.

A voltage source can have very high available short-circuit currents. If it were not for the overcurrent devices in the load centers and main disconnects, the typical utility transformer feeding a residence could deliver short-circuit currents approaching 10,000 amps. The larger transformers feeding commercial buildings with 480-volt ac can deliver even higher short-circuit currents. The typical 12-volt automotive battery can send several thousands of amps into a short circuit.

PV modules as current-limited current sources have a limited capability to produce high currents. A typical 208-W PV module might have an operating current of 7.5 amps and be able to deliver a short-circuit current of only 8.1 amps. The amount of current a single PV module can deliver is limited by the size of the cells in the module, the method of internal wiring, and the brightness of the sunlight falling on it.

Modules are rated in the laboratory at a set of standard test conditions (STC) that include, among other things, a standard sunlight (solar) intensity of 1000 watts per square meter (W/m2). At 1000 W/m2, the module is tested and the values of short-circuit current (Isc), and operating current (Imp) are recorded. Average production values for these two parameters are marked on the back of the module along with other items. On most modules, the short-circuit current at STC will be only about 10–15% higher than the operating current.

When modules are connected in series to form what the PV installer calls a "string” of modules, the operating and short-circuit currents do not change. The string currents are the same as the values for a single module. However, the voltage that each module produces does add up in the string and many typical residential PV systems operate with voltages in the 400- to 600-volt range.

A 480-Volt PV System is NOT like a 480-Volt Feeder

Yes, PV systems can and do operate over 480 volts, but unlike the 480-volt ac feeder (a voltage source) the available short-circuit current in a typical residential system will be less than 10–20 amps. At this voltage, shock hazards definitely exist, and while arcs are possible, arc-blast hazards are not possible since the available current is insufficient to produce them. However, even in residential systems, installation errors can result in damaged equipment (see photo 1). Arcs involving direct currents are somewhat more difficult to extinguish than arcs in alternating currents because the dc arcs do not self-extinguish 120 times per second as do ac arcs.

Photo 1. Burned combiner box caused by wiring error

As the PV systems get larger (commercial systems), the operating voltages will usually be limited to no more than 600 volts, except in a few experimental systems operated by utilities on utility property and behind secure utility fences. But, as the system size and power increase, the current will increase above the few tens of amps into hundreds of amps. There are now single, utility-interactive inverters rated at 500 kW ac output and they will have dc operating currents and short-circuit currents at the inverter input approaching 1000 amps (see photo 2). These large PV systems, and the high levels of dc current, while not as large as the

Photo 2. High-current dc fusing

hundreds of thousands of amps associated with 480-volt ac feeders, must be treated with extreme caution when working on such a system when energized (see photo 3).

Fortunately, most residential PV systems rated at power levels of 2500–5000 watts have dc currents in the 5–15 amp range and are proportionately less dangerous from an arcing point of view.

Working Safely on PV Arrays

With a solar energy source, it is somewhat difficult to turn off the current from an illuminated PV module. If the attached leads from each module were available, they could be disconnected (open circuited) or connected together (short-circuited) to reduce either the current or the voltage from the module to zero. However, in most PV systems, these leads and their connectors are not readily accessible, so other means must be used to work on active systems. One method is to cover the modules with an opaque surface, but this is rarely

Photo 3. 250 kW inverters

done on any but the smallest systems. The area of a PV array in a typical residential PV system might be hundreds of square feet as shown in photo 4.

In the typical PV module, the output conductors are terminated in connectors that are insulated and are, to a limited extent, considered "touch safe”— photo 5. These connectors can be plugged together safely, and, in the usual wiring sequence, at least one pair of these connectors is left open until all other wiring is done on the PV array. By leaving one connection open in each string of the array while making terminal connections to disconnects at the end of each string, the possibility of getting shocked is somewhat reduced. However, ground faults and stray leakage currents can always present a shock hazard, so insulated gloves and tools are recommended when making terminal connections.

Photo 4. Residential rooftop PV array

Continuous Currents

From the onset, the PV Industry has been aware of the nature of the PV module as a current source, the limited nature of that current, and its relationship to the intensity of the solar radiation. Article 690 of the National Electrical Code was written to address this unique nature of the PV module and the PV system. Several requirements are in the Code that will allow PV systems to be safely designed and installed in a manner that should provide an essentially hazard-free system for the life of the PV module—a life that may approach 50 years.

For purposes of calculating the ampacity of conductors, the Code requires that the rated dc short-circuit current (not the operating current) from a single PV module or the current from a combiner box that carries the paralleled output of several strings of modules be multiplied by 125 percent [690.8(A)]. This 125 percent factor is applied to address the fact that the solar irradiance on clear days in many parts of the country may exceed the standard test condition value of 1000 W/m2 by a varying amount. Values of 1150 W/m2 (115 percent) are not uncommon for periods that can last three hours or more. This 125 percent factor ensures that on a continuous basis (as defined by the NEC), the currents used in ampacity calculations are the worst possible currents that could be generated by the PV module. Also, by this calculation, we are implying that these currents exist continuously 24 hours per day when, actually, they will never reach the 125 percent level, may only be above 1000 W/m2 for only a few minutes each day, are lower most of the time, and drop to zero at night. Also note that the calculation involves the short-circuit current, not the operating current which is typically 10–15% lower, and that the short-circuit current would normally only flow under a fault condition.

Photo 5. "Touch Safe" module connectors

Then, the normal Code provision (applied to all conductors and most overcurrent devices) is added where they must have a basic ampacity of 125% of the continuous load (already 125% of Isc) [690.8(B)]. Since this is an energy supply system, we use the term source currents rather than load currents, but the same calculations apply. The Code has now required that a 156 percent (125% x 125%) factor be applied to the rated short-circuit current for the module or modules. This ensures that, after appropriate corrections have been made for conditions of use, the conductors carrying the dc currents from PV modules will never be overloaded and will be operated within their ratings for the life of the system.

Conservative Designs Yield Long, Safe Operation

PV modules may be producing hazardous voltages and currents for 40 years or longer. Using the 156% factor on the short-circuit current to size cables and rate overcurrent devices helps to ensure that these devices will safely and reliably carry the normal PV currents for many years. Even in the extreme outdoor environment (hot, wet, and ultraviolet) that provides challenging operating conditions, conservative ratings keep conductors and overcurrent devices well within their operating tolerances.

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 505-646-6105

A color copy of the latest version (1.6) of the 150-page, Photovoltaic Power Systems and the 2005 National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPrac.htm

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.htm

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the SWTDI web site.

## Disconnect, Disconnect, Where For Art Thou?

Posted By John Wiles, Tuesday, May 01, 2007
Updated: Tuesday, January 22, 2013

The requirements and necessity for, and the location of disconnects in a photovoltaic (PV) power system are always of great interest. While PV equipment manufacturers, designers, installers, and electrical inspectors are all interested in getting safe PV systems, there are usually some "friendly” discussions on the whys and hows of disconnects needed to achieve those ends. The following information may shed a little light on those sometimes elusive disconnect requirements and how they can be addressed.

Photo 1. Small system dc PV disconnect

Disconnects defined

Article 100 in the NEC defines disconnecting means as a device or devices that could be used to disconnect circuits. Switches, circuit breakers, screw terminals, and bolted connections fall under that definition (see 690.17).

Why are they needed?

PV disconnects are generally required on both small (photo 1) and large (photo 2) systems for two reasons. The first reason is to disconnect the external power source conductors from the circuits in the building or structure (690.13, 230.70). A common disconnect of this type is the ac service-entrance disconnect for a house. On a PV system, the main PV dc disconnect falls into this category if the PV dc conductors penetrate the house. Although batteries are not power generators, they can source energy, so a battery disconnect might also fall into this category.

Photo 2. Large system dc PV disconnects

Secondly, disconnects are required to remove power from a device that needs maintenance. I could use the word "service” instead of "maintenance,” but I am trying to make these articles clearer than the Code. Of course, all of the main-power disconnects could be opened to remove all power from a building, but disconnects associated with equipment that must be maintained provide a degree of safety without shutting down the entire electrical system for maintenance on a single piece of equipment (see 690.15).

Disconnects for PV systems, let me count the ways

A main dc PV disconnect is required where the PV dc circuits from the PV array enter the building (690.13, 690.14).

A main ac PV disconnect is required where the dc PV circuits do not enter the building, but the ac output of the inverter does. No, you won’t find this one explicitly listed in 690, but see figures 1 through 4 (690.13, 690.14).

Figure 1. All components outside the building

A dc inverter maintenance disconnect is required and more than one may be required if the system has batteries (690.15).

A battery disconnect is normally required on stand-alone PV systems with batteries or utility-interactive PV systems with battery backup.

An ac inverter maintenance disconnect is required for utility interactive inverters (690.15).

Stand-alone inverters with generator inputs may also require a generator disconnect at the inverter input (690.15).

Figure 2. Main load center inside building

Charge controller input and output disconnects are required for maintenance on systems with batteries (690.15).

Systems with backup generators will normally require a generator disconnect both outside at the generator location (point of entry power disconnect) and inside near the inverter and other power processing equipment (maintenance disconnect).

The ac point of connection will require a disconnect on utility-interactive systems [690.64(B)(1)].

Many utilities will require a lockable open, visible blade ac disconnect for the PV system, and this disconnect will usually be located near the utility revenue meter.

Where art thou?

Utility personnel and emergency responders such as fire fighters like to know where the main-power disconnects are located. The generalNECrequirements for these disconnects are discussed below, but the local jurisdiction may have differing requirements.

Figure 3. Inverter inside the building

Although there are two separate requirements for disconnects, in some cases a single disconnect, properly rated and located, may solve both requirements. In other cases, due to equipment placement and the necessity for grouping the maintenance disconnects, two or more disconnects may be needed in a single circuit (690.15).

With the introduction of PV and Article 690 into the 1984 edition of the NEC, the original intent of the requirements for the PV disconnect was to match them with the existing requirements for the ac service disconnect as established by Article 230. In fact, 690.14 in the 1984 NEC referred the reader directly to Article 230 Part F. Unfortunately, most PV installers did not follow this guidance because they were not electricians familiar with installing ac service-entrance conductors and service disconnects. The PV installers frequently penetrated the roof with energized PV source and output conductors and routed these conductors to the main dc PV disconnect just about anywhere in the structure they pleased. Complaints from electricians and electrical inspectors caused the NFPA (without any help from the PV industry) to rewrite Section 690.14 in the 2002 NEC. In this revised section (which mimics 230 Parts IV, V), the requirement was firmly established to install the PV disconnect in a readily accessible location at the point where the PV conductors first penetrate the structure. This requirement effectively keeps the energized PV conductors outside the structure until reaching that disconnect.

Figure 4. Inverters on the roof

The NEC does not specify whether the main ac service disconnect or the main dc PV disconnect is to be located inside or outside the structure at the point of penetration of these circuits. That is left to the local jurisdiction and the requirement for locating these disconnects varies throughout the country. Figure 1 shows the simplest configuration of a utility-interactive PV system where the local jurisdiction requires all disconnects to be on the outside of the building, the ac load center is mounted on the outside of the building, and the inverter is also mounted on the outside of the building. This meets the K.I.S.S. principle.

An addition to the 2005 NEC [690.31(E)] allows the PV source and output conductors to be routed inside the building (the dotted line in the figures) before they reach the main PV disconnect if they are installed in a metal raceway. Metal raceways include metal conduits and flexible metal conduit. Metallic cable assemblies are not allowed so the installations cannot yet use Type MC and Type AC cable assemblies—maybe in 2008?

In figure 2, the main load center is inside the building and this contains the backfed PV circuit breaker. Where the utility requires an external disconnect (usually lockable open), the utility may also allow this disconnect to be used as the grouped ac maintenance disconnect for the inverter. If the utility disconnect is not required or it cannot be used as a code-required maintenance disconnect, then a separate ac disconnect will have to be mounted in this circuit next to the inverter on the outside of the building.

In figure 3, the local jurisdiction requires that the main ac and dc power disconnects be located outside the building and the main load center containing this disconnect is also outside the building. For architectural reasons, the inverter is located inside the building. To provide for safe maintenance of the inverter, added dc and ac maintenance disconnects are needed inside the building on either side of the inverter.

While lock-out tag-out procedures might be used in an industrial environment to use just the external disconnects to de-energize the inverter safely, these procedures are not easily adapted in the residential or commercial environments.

Where batteries are located in a separate room or at some distance (typically, five feet or more) from the inverter and charge controllers, a disconnect is required at the battery location, and this disconnect is usually merged with an overcurrent protective device.

If a backup generator is used in the system, it is generally located outside the structure. A disconnect will be required at the generator and then again inside the building near the inverter or power distribution panel.

Contrary to the understanding of some inspectors, there is no requirement for a disconnect at the PV array [690.14(C)(5)]. Such a disconnect serves no safety purpose for the user or PV installer since the PV array is always energized when illuminated even if the disconnect were opened.

There are some PV installations, both residential (flat roofs) and commercial, where the inverters are mounted near the PV arrays on the roof in not-readily-accessible locations. NEC 690.14(D) addresses these systems and requires ac and dc disconnects at the inverters and an additional ac PV disconnect at ground level. Figure 4 shows this system where all of the equipment is outside the building.

For a discussion on the use of disconnects inside the inverter, see the "Perspectives on PV” in the September-October 2006 issue of IAEI News.

To disconnect or not to disconnect…

That is not the question. Disconnects are required throughout the PV system with the proper ratings and in the code-required places. As the system complexity increases with batteries, generators, and possibly wind or hydropower inputs, the number of disconnects increases. The basic disconnect requirements were in the Code long before PV systems arrived, and following those requirements as well as the newer requirements for PV systems will make for safe installations.

If this article has raised questions, do not hesitate to contact the author by phone or e-mail. E-mail: jwiles@nmsu.edu Phone: 505-646-6105

A color copy of the latest version of the 150-page, 2005 edition of the Photovoltaic Power Systems and the National Electrical Code: Suggested Practices, published by Sandia National Laboratories and written by the author, may be downloaded from this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/PVnecSugPract.html

The Southwest Technology Development Institute web site maintains a PV Systems Inspector/Installer Checklist and all copies of the previous "Perspectives on PV” articles for easy downloading. Copies of "Code Corner” written by the author and published in Home Power Magazine over the last 10 years are also available on this web site: http://www.nmsu.edu/~tdi/Photovoltaics/Codes-Stds/Codes-Stds.html

The author makes 6–8 hour presentations on "PV Systems and the NEC” to groups of 40 or more inspectors, electricians, electrical contractors, and PV professionals for a very nominal cost on an as-requested basis. A schedule of future presentations can be found on the SWTDI web site.