Print Page   |   Contact Us   |   Sign In   |   Join
Perspectives on PV
Blog Home All Blogs
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:


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 

Microinverters and AC PV Modules Are Different Beasts

Posted By John Wiles, Tuesday, May 01, 2012
Updated: Wednesday, January 16, 2013

Microinverters and PV modulesMicroinverters and AC PV modules are becoming very common in residential and small commercial PV systems. See photos 1 and 2. They have even been used in PV systems rated at 60 kW and above. They have some common features. For example, microinverters and AC PV modules have similar ac output characteristics, connections and code requirements. However, they are different from the typical PV string inverters that use multiple modules connected in series and have dc voltages in the 200–600 volt range. The microinverters and AC PV modules typically operate with just one PV module and the dc voltages are less than 100 volts.

Instructions supplied with these listed products should be followed [NEC110.3(B)]. The suggestions below do not substitute for compliance with theNECor local codes.


Both the AC PV module and the microinverter will require equipment-grounding connections where there is any exposed metal in these devices. A grounding electrode conductor (GEC) connection will be required when the microinverter operates the module in a grounded manner.

Equipment/Safety Grounding

Photo 1. PV microinverter with exposed dc cables and connectors to PV module

Photo 1. PV microinverter with exposed dc cables and connectors to PV module

The ac output circuit cable of some microinverters and AC PV modules does not have an ac equipment grounding conductor (EGC). This EGC conductor must be started (originated) in the transition box on the roof where each set of inverters has the final factory ac output cable connected to another wiring system. The ac equipment grounding conductor should also be attached to the microinverter enclosure. This ac EGC must be routed all the way back to the service-entrance bonding point as it is in any other ac circuit. There is no requirement that it be unspliced and the size will typically be 14 AWG per Table 250.122.

System/Functional Grounding

True AC PV modules where there are no readily accessible dc conductors or dc disconnect will normally not require a grounding electrode conductor. Since both the requirements in the 2005NEC690.47(C) and the permitted 690.47(C) in the 2008NECare both based on Article 250, the provisions of either editions of the Code appear to be applicable in jurisdictions using either edition. Section 690.47(C) in the 2011NECcombined and clarified 2005 and 2008 code requirements in this area.

Under UL Standard 1741 the microinverter, if it isolates the dc grounded input conductor (assuming a grounded PV module) from the ac output, must have a dc grounding electrode conductor (GEC) running from the grounding electrode terminal on the microinverter case to a dc grounding electrode. If the microinverter operates the PV module as an ungrounded system (neither positive nor negative connected to ground), then no grounding electrode conductor would be required.

Section 690.47(C) in the 2008 NEC permits the use of a combined ac EGC and dc grounding electrode conductor (GEC) from the inverter. The 2011NEChas this requirement in 690.47(C)(3). UL 1741 requires the dc GEC terminal on the outside of the inverter. If this option is elected, then the 8 AWG minimum (250.166) conductor from each inverter must be bonded to the input and output of each metal conduit and metal box that it travels through until it gets to the main grounding bar in the service entrance equipment. The bonding requirement and 8 AWG size would appear to rule out the use of 10-3 with ground type NM cable for the ac output circuit inside the building. The bonding requirement may also be cumbersome to implement multiple times and the routing of this combined conductor may induce lightning surges to enter the main load center and other branch circuits. The permissive method of grounding described in 690.47(C) in the 2008NECmay also be used under the 2005NEC.

Photo 2. AC PV module. No exposed dc cables or connectors. Courtesy Exeltech.

Photo 2. AC PV module. No exposed dc cables or connectors. Courtesy Exeltech.

Alternatively, the permissive grounding method described in the 2005 NEC 690.47 may also be used under the 2008NECas an alternative to the 2008 NEC 690.47. Section 690.47(C) in the 2005NECand 690.47(C) in the 2008 NEC are based on the general requirements of Article 250.

Section 690.47(C) in the 2011 NEC combines and clarifies the grounding methods described in the 2005 and 2008 NEC.

The Exception in 690.47(D) in the 2008 NEC regarding array grounding is not clear. The subject of the section refers to array grounding electrodes. It is not clear if the Exception removes the requirement for an additional array grounding electrode only and leaves the requirement for the array GEC or removes the requirement for both. The intent of the submittal was to use a new array GEC to ground the array to an existing electrode or for a ground-mounted array, to a new grounding electrode at the array location. This would be particularly important in a high lightning area, but that is a performance issue, not a safety issue. This section was not in the 2005 NECand was removed from the 2011 NEC. An auxiliary grounding electrode is always an option under 250.54.

The size of the dc grounding electrode conductor is determined by 250.166, and this section has been clarified in the 2008NEC. In many cases, but not all, a 6 AWG bare copper conductor will meet the requirements. Where a UFER (concrete-encased electrode) is used, a 4 AWG grounding electrode conductor will usually be required. A short 6 AWG conductor may have to be irreversibly spliced to the 4 AWG conductor at each microinverter and connected to the microinverter grounding terminal if the inverter grounding terminal will not accept a 4 AWG conductor directly. An alternative would be to drive a single ground rod six or more feet from the UFER ground, ground the inverters and modules as described below with a 6 AWG bare copper grounding-electrode conductor and then bond the ground rod to the UFER with a 4 AWG bonding jumper (690.47(C)(1) in 2005 and 2011NEC).

The dc grounding electrode conductor may terminate at the service-entrance grounding electrode or at a grounding electrode associated with any subpanel where the inverter dedicated circuits end in backfed breakers under the 2005NEC. Under the 2008NEC, the combined conductor dc GEC/ac EGC can be terminated at the main service grounding bus bar or at any subpanel bus bar that has a grounding electrode attached and where the inverter backfed breaker terminates. The 2011NECallows either location to be used.


The microinverters should be installed in compliance with 690.14(D) of theNEC. As noted in this section, there are requirements for dc and ac disconnects on the roof in this not-readily accessible area, and an additional ac disconnect in a readily accessible location.

The relatively low dc voltage (usually less than 70 volts) and currents (less than 8 amps) may allow the dc connectors on the microinverter inverter to serve as the dc disconnects for servicing the inverter. In a similar manner, the ac connectors on the microinverters and AC PV modules could be used as the maintenance disconnects required by 690.15. Microinverter and AC PV Module manufacturers can have the ac and dc connectors designed and listed with the microinverter or AC PV module as load break rated disconnects and this will allow the use of these connectors to meet Code requirements (690.14, 690.15 and 690.17).

Even with load break rated ac connectors, a transition box is needed to change from the flexible ac output cable to the code-required fixed wiring system that will enter the building. An inexpensive unfused 60-amp 240-volt air conditioning pull out disconnect would serve nicely and is already in a NEMA 3 R enclosure. It will also serve as an ac disconnect that when pulled, will shut down the microinverters or AC PV modules and opening the ac circuit will reduce the dc currents in the microinverter input cables and connectors to very near zero permitting safer opening of the dc disconnects.

Such a disconnect can also be used to meet some AHJ requirements for a non-connector disconnecting means on the roof.

Section 690.14(D)(3) requires an additional disconnect and that disconnect requirement may be met by the backfed breaker in the load center where the load center is positioned to meet the accessibility and location requirements of 690.14(C)(1). Some jurisdictions are requiring that this second ac disconnect be on the outside of the building and any utility-required disconnect on the inverter output circuit would usually meet this requirement.

AC Output Circuits

The output circuit of any utility-interactive inverter up to the first overcurrent protection device (OCPD) is very much like an ac branch circuit. If the utility voltage is removed from this circuit (for any reason), the circuit becomes de-energized (dead) — just like a branch circuit. If there is a line-to-line or line-to-ground fault on this circuit, the OCPD responds in a normal manner to the fault currents generated by the utility. The inverter(s) can generate no more than its rated current per UL Standard 1741 and when the fault occurs, the drop in line voltage will normally cause the inverter to shut down. And when the branch circuit breaker opens in response to the fault, the inverter shuts down.

It would appear that these inverter output circuits could be wired using any Chapter 3 wiring method suitable for the environment (hot, wet and UV outside and hot in attics). Grounding requirements or methods used for microinverters may dictate conductor sizes too large for 10 AWG type NM conductors.

An ac GFCI device should not be used to protect the dedicated circuit to the microinverter or ac PV module even though it is an outside circuit. None of the small GFCI devices (5 ma–30 ma) are designed for back feeding and will be damaged if backfed. In a similar manner, most ac AFCIs have not been evaluated for backfeeding and may be damaged if backfed with the output of a PV inverter.

Combining Multiple Sets of Microinvertersor AC PV Modules

In multiple strings of these inverters, there is no NEC requirement that an ac combining panel (load center) be located on the roof. In fact, most NEMA 3R load centers must be mounted against a surface to keep water from penetrating holes in the back panel and they must be mounted within 30 degrees or vertical. Such a surface may have to be added in order to properly mount a 3R load center on the roof. And then there might be problems meeting 110.26 clearance requirements. A further issue with OCPD on the roof is heating of the device over its rated 40 degrees Celsius operating temperature. Gray load centers in the sun will normally operate 10–20 degrees C hotter than the ambient temperature. This may be difficult to compensate for when considering available equipment, the size of the ac conductors attached to the inverters, and listing restrictions on the inverters. Nevertheless, it is possible to mount an ac load center on the roof with proper solar shielding and use it to combine the outputs of U-I inverters or sets of microinverters.

The rating of any combining panel and the ampacity of conductor from that panel to the backfed breaker in the main load center as well as the rating of the main load center and the backfed breaker must meet 690.64(B)/705.12(D) requirements. This requirement will require a combining panel and conductor with a rating nearly twice sum of all of the 15-amp or 20-amp backfed breakers used for each output. See the 120% allowance in 690.64(B)(2)/705.12(D)(2) and 690.64(B)(7)/705.12(D)(7).

The ac output conductor for a set of inverters must have an ampacity of 125% of the continuous currents for all of the inverters on that circuit. The backfed circuit breaker in the panel must be rated the same and if an odd current rating is determined, the breaker rating should be the next larger size. The breaker must protect the conductor under the conditions of use and the conductor ampacity must be derated for those conditions of use.

The ac output circuit from each set of inverters must have an equipment grounding conductor to facilitate OCPD operation during ac ground faults. Some microinverters have a three-wire output through a four-contact connector. The unused terminal in the connector is reserved for future use. The three active pins in the connector are 240-V L1 and L2, and a neutral. There is no ac equipment grounding conductor. This lack on an equipment grounding conductor in the cable requires that the equipment grounding conductor for the microinverter or ac PV module be an external connection to the inverter case, where the case is metal. This external equipment grounding conductor must be connected to the fixed wiring system (usually, but not always conduit) where that wiring system originates.
Unless the microinverter bracket has been designed and evaluated as a grounding/bonding jumper, grounding the microinverters does not ground the rack or the modules and visa versa.

There is only one ac neutral-to-ground bond in an ac electrical system. That bond is made in the existing service entrance equipment. No additional neutral-to-ground bonds should be made when installing a PV system unless a supply-side service entrance connection is made.

AC PV Module Grounding — A Gray Area

Combinations of PV modules and microinverters combined/assembled in the field or at the dealer or distributor do not meet the intent, definition, or requirements associated with true AC PV Modules as defined in 690.2 and in 690.6. As of early 2012 there is no specific size associated with either microinverters or ac PV modules. The power outputs are increasing with nearly every new product and are now in the 190–220 watt range.

Combinations of a microinverter and a PV module with exposed dc connectors and dc conductors between the PV module and the microinverter are being certified/listed as ac PV modules. Some of these products have instruction manuals that say the microinverter may not be removed from the PV module. Other manuals give specific instructions for removing the microinverter from the PV module for repair. At issue is the definition of an ac PV module as a factory assembled unit and the potential need to meet all dc code requirements for these products with exposed dc connectors and dc conductors. Connectors are subject to loosening or being opened in the field. Connectors and conductors are exposed to environmental degradation, ground faults, and animal damage.

Also at issue in the ac PV module is the microinverter-to-PV module frame bonding when the mechanical/electrical connection is broken in the field. When the microinverter is replaced, how is the bonding connection quality verified and how is the certification/listing maintained without NRTL evaluation?

At some point, these issues will be addressed in UL Standard 1741 and possibly in theNational Electrical Code.

For Additional Information

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

See the web site below for a schedule of presentations on PV and the Code.

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 (1.91) 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:

Read more by John Wiles

Tags:  Featured  May-June 2012  Perspectives on PV 

Share |
PermalinkComments (0)

More Questions from Inspectors Numerous PV Systems Pose Issues

Posted By John Wiles, Thursday, March 01, 2012
Updated: Wednesday, January 16, 2013

PV SystemsPhotovoltaic (PV) systems prices continue to drop and inspectors are getting numerous requests for inspections. The questions that I receive indicate that this is new territory for many inspectors. These questions also indicate a few "holes” in the National Electrical Code, which we hope to plug in the 2014 NEC.

Questions on Grounding

Question: Does the NEC require that a grounding electrode conductor (GEC) and a grounding electrode (ground rod) be connected to the new transformerless inverters? See photo 1. Section 690.47 in the 2011 National Electrical Code (NEC) does not exactly address this issue.

Answer: If the listed transformerless inverter (also called a non-isolated inverter) adheres to the requirements of Underwriters Laboratories Standard 1741 for PV inverters, the inverter will not even have a terminal for a grounding electrode conductor. These inverters are used with an ungrounded PV array. The UL standard requires a grounding electrode conductor terminal and the Code would require a grounding electrode conductor only when there is a bonding jumper in the direct current (dc) side of the inverter. In normal transformer-type of inverters (also called isolated inverters), this bonding jumper is part of the required 690.5 ground fault detection and interruption (GFDI) circuit.

Photo 1. Transformerless inverter. Looks like many other inverters that have transformers, but may not have a GEC terminal or a 690.5 Warning. Photo courtesy SMA Technologies

Photo 1. Transformerless inverter. Looks like many other inverters that have transformers, but may not have a GEC terminal or a 690.5 Warning. Photo courtesy SMA Technologies

Transformerless inverters do not connect one of the dc circuit conductors in the PV array to ground (as allowed byNEC690.35) and have no internal bonding jumper. Therefore, there will normally be no terminal to connect the GEC to and theNECdoes not require a dc GEC. Unfortunately, Section 690.47 does not specifically say this, so a proposal has been submitted for the 2014NECthat hopefully clarifies the issue. Here is the wording of that proposal for 690.47(B).

Add a new third paragraph to 690.47(B) as follows:

Ungrounded DC PV arrays connected to utilization equipment with common ac and dc equipment-grounding terminals shall be permitted to have dc equipment-grounding requirements met by the ac equipment-grounding system without the requirement for a dc grounding electrode conductor or grounding system.

We have been asking PV installers to get that dc GEC connected to the inverter for many years. Now, on these new systems, it will no longer be required. But, be advised that not all inverter manufacturers, nor their certification agencies, will read all of the fine print in the standard and some transformerless inverters will have terminals or instructions for a GEC. This terminal will be, as it is in other inverters, connected internally to the dc and ac equipment-grounding conductor terminals. And, if desired, this terminal may be used with a GEC routed to a grounding electrode. This would essentially be a 250.54 optional grounding electrode and that electrode does not have to be bonded to any other grounding electrode. It is connected only to the equipment-grounding system in the inverter.

The 690.47(B) proposal for the 2014NECindicates that since the ac and dc equipment-grounding conductor terminals are common in the inverter, the ac equipment-grounding system (grounded at the service-entrance equipment) can be used to provide the array equipment-grounding function.

However, this may route lightning induced surges on the array equipment-grounding system through the inverter and into the service equipment. Far-thinking PV installers may elect to install optional 250.54 grounding systems at the array and possibly also at the inverter to better protect against these surges.

Question: What is the proper method of grounding the modules and microinverters that are not manufactured or certified/listed as an AC PV module?

Answer: These microinverters are essentially small inverters. It is difficult to precisely define them as a unique device since they are continually getting larger (now 380+ watts) while some normal "string” inverters are down to 700 watts and below. The micro-inverter/ PV module combination has many of the characteristics of any other inverter when it comes to grounding. There are usually exposed metal surfaces on both the inverter and the module that must be grounded (i.e., connected to earth through an equipment-grounding conductor/system). The microinverter may cause the module to operate as an ungrounded module, as a positively grounded module (most common), or as a negatively grounded module. This form of grounding refers to how the dc circuit conductors are referenced to ground and is called system or functional grounding. When the module is operated in a grounded manner, there will be a dc bonding jumper inside the inverter and this fact will require that the inverter have a dc grounding electrode conductor terminal. The dc grounding electrode conductor (GEC) will have to be 6 AWG in exposed locations and at least 8 AWG inside conduit. It will have to be unspliced or irreversibly spliced from the microinverter all the way to the grounding electrode or the grounding bus bar in the equipment that has a connected grounding electrode.

Photo 2. Microinverter with single grounding terminal for both equipment grounding and dc grounding electrode conductor connections. Photo courtesy Enphase.

Photo 2. Microinverter with single grounding terminal for both equipment grounding and dc grounding electrode conductor connections. Photo courtesy Enphase.

The inverter should also have an ac equipment-grounding conductor that will be routed with the ac output circuit conductors. With a dc input from the module, there should also be provisions to accept a dc equipment-grounding conductor from the PV module. However, in many cases, a single external terminal on the microinverter can meet both the equipment-grounding terminal requirements (ac and dc) and the grounding electrode conductor terminal requirement. See photo 2

In general, the module will require an equipment-grounding conductor attached to the frame following the instructions provided in the module instruction manual and sized per 690.45. In many cases this can be as small as 14 AWG. Of course, 690.46 may apply or the AHJ may require a larger conductor to provide greater mechanical integrity. In these cases, a 6 AWG conductor is frequently used.

In some cases, an electrical connection (not just a mechanical attachment between inverter and module frame) between the module frame and the microinverter enclosure will enable a single equipment-grounding conductor to be used for both devices.

Creative sizing (6 AWG) and routing of a single unspliced conductor can be used to meet all module and microinverter grounding requirements.

Question: What type of grounding is required on modules with plastic frames (also known as industrial composites) and these new dc-to-dc converters that are attached to the module outputs that have plastic enclosures? See photo 3

Answer: My favorite type a question — an easy one. If there are no exposed metal parts on a module, a microinverter, or a dc-to-dc to dc converter, there will be no requirement for an equipment-grounding conductor and probably no place to attach such a conductor. However, we may get a plastic encased dc-to-dc inverter or a microinverter that has a dc grounding electrode conductor requirement and there will be a terminal for that conductor. The manual for these certified/listed products will have the instructions for this connection.

Questions on Overcurrent Protection

Question: When do multiple strings of modules require a fused combiner box or a set of fuses inside the inverter?

Answer: The number of strings of PV modules that can be connected in parallel without a fused combiner is determined by the short-circuit current (Isc) rating of each module and the maximum series fuse. Each string of modules can, under worst-case conditions of sunlight, generate 1.25 x Isc of current into a fault in a parallel-connected string of modules. If we have "n” strings connected in parallel, then "n-1” strings can send fault current into a faulted string. The total fault current would be (n-1) x 1.25 x Isc. That fault current must be less than the rating of the module protective fuse marked on the back of the module. If the fault current were greater than the value of the module protective fuse, then the module and its cable could be damaged where there was no fuse.

A little PV math shows that:

(n-1) x 1.25 x Isc < F where F is the value of series fuse marked on the back of the module.

If we solve this for n, the total number of strings in parallel, we get:

n< (F+1.25 x Isc)/(1.25 x Isc)

Example 1: Module W has an F of 15 amps (pretty common) and an Isc=8 amps.

n<(15+1.25 x 8)/(1.25 x 8) = 25/10 = 2.5, and the total number of strings (n) for this module can be 2.5; and since n has to be a whole number, two strings of modules can be connected in parallel.

Example 2: Module Y has F = 20 and Isc = 3. n < (20+1.25 x 3)/(1.25 x 3) = 23.75/3.75 = 6.33 and six strings of these modules can be connected in parallel.

For many PV modules in the 180–300 watt range, only two strings can be connected in parallel because of these constraints.

Question: Can two sets of 15 microinverters be connected in parallel without overcurrent devices?

Answer: In short — No. The microinverters are tested and certified/listed to be used as a set with the cable or wiring harness provided with them. The instruction manual will specify how many microinverters can be connected to the factory cable and the rating of the required circuit breaker for that set on a single cable. This is consistent with NEC 705.12(D)(1) that requires a dedicated circuit breaker for utility-interactive inverters.

Question: How does the short-circuit current from a PV module affect the output current of the connected dc-to-dc converter? How is the PV module open circuit voltage used to calculate the voltage rating of any combiner or inverter downstream.

Photo 3. Dc-to-dc converter. Photo courtesy Tigo Energy

Photo 3. Dc-to-dc converter. Photo courtesy Tigo Energy

Answer: This new technology of dc-to-dc converters and other PV module power processors has evolved in numerous configurations. Some converters are required on the output of every PV module; some are required on only a few modules. Most are connected in series, but some are connected in parallel. Some of the devices are "smart” and must be used with "dumb” inverters. All of these devices must be certified/listed to UL Standard 1741. There are and will be too many variations to address the specific connection requirements of each product in the NEC directly. The outputs of these devices are decoupled from their inputs, so PV module short-circuit currents and voltages cannot directly be used to meet any Code requirements that are associated with the circuits connected to the output of these devices. Essentially the installers and the inspectors will have to rely on 110.3(B) where these certified/listed devices must be installed following all instructions provided with the product and all labels on the product. A proposal for the 2014NECwill reinforce this requirement in Article 690.

Questions on Large Systems

Question: What needs to be addressed concerning the ground-fault protective device connected to the inverter output to meet the exception on 690.64(B)(3)/705.12(D)(3)? The exception requires that loads be protected from all sources of ground-fault currents.

Answer: This particular area is beyond the scope of the NEC. The load circuits must be protected from ground faults originating from the utility service and also from ground-fault currents originating from the load-side connected PV inverter. The fault currents reaching the load circuits will be shared between these two sources. It would take engineering analysis to determine how the two sources will share the fault currents under various situations and how the settings of each ground fault device will be determined.

Question: The service disconnect is at 12 kV (12.47 kV) for a large facility and the PV system will be connected at 480 volts on a feeder. For this load-side connection, how do we apply 705.12(D)(2) to determine conductor and busbar ampacities when transformers are involved?

Answer: The voltage ratio of the transformer is used to adjust the various overcurrent device ratings and ampacities to an equivalent set of numbers at a single voltage, either at the 12 kV or the 480 V level.

For example, a 25-amp fuse on the 12 kV side of the transformer would translate to about a 650-amp (25 x 12470/480) overcurrent device when referenced to the 480 V feeder. Then the requirements of 705.12(D) may be applied. In these large facilities, keep in mind that the PV inverter output connection to the existing system must be made at the end of a feeder or busbar opposite the utility feed end before the 120% allowance can be used. If the sum of the overcurrent devices exceeds 120% of the ampacity of the feeder or the rating of the busbar, or the PV connection cannot be properly located, a 100% factor must be used. Any circuits (conductors and busbars) not protected by a single overcurrent device that may carry current from the PV system may have to be increased in size.

Keep those questions coming. The holes in the 2014 NEC have not yet been addressed and thatCodeis more than two years away.

For Additional Information

See the web site below for a schedule of presentations on PV and theCode.

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 (1.91) 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:

And yes, it may be updated to the 2008 and 2011 Codes sometime this year.

Read more by John Wiles

Tags:  Featured  March-April 2012  Perspectives on PV 

Share |
PermalinkComments (0)

Questions from Inspectors — Inquiring Minds Need to Know

Posted By John Wiles, Sunday, January 01, 2012
Updated: Wednesday, January 16, 2013

The following questions and answers result from some of the more common situations that many inspectors face throughout their working day when seeing a new PV installation or reviewing a set of plans for a PV system. The questions are simplified versions of questions I receive in e-mails and from questioned plan sets as well as sometimes long, involved phone calls.

Service Entrance Questions

Question: I am looking at a diagram of a PV system where the main service is a 100-amp main-lug-only (MLO) panel with six breakers. One of the six breakers is rated at 40 amps and is being backfed from a utility-interactive PV inverter. Doesn’t this 40-amp breaker exceed the 120% allowance of 690.64(B)/705.12(D) that would limit the backfed breaker to 20 amps on a 100-amp panel?

Answer: These six breaker MLO service-entrance panels are common in many areas of the country, primarily in older homes. There is no main overcurrent device or disconnect ahead of the MLO panel busbar, so each of these six breakers represents a service disconnect. That backfed 40 amp represents a supply-side connection allowed under 690.64(A)/705.12(A) and the load-side requirements of 690.64(B)/705.12(D) do not apply. The limit of the breaker rating in such a supply-side connection would be the rating of the MLO panel, the rating of the panel busbar (usually the same as the panel rating), or the rating of the service, whichever is less.

Photo 1. Main-lug-only panel — PV breaker rating?

Question: The PV installer has made a supply-side connection between the meter base and the load center by cutting the EMT, installing a pull box and making the PV connection inside the box. He has installed a fused disconnect adjacent to the pull box and has run EMT to the inverter. Workmanship looks good, but what else should I be looking for?

Answer: The NEC treats these supply-side connections as additional services as allowed by 230.2(A)(5). As services, the various requirements of services should be followed including conductor type between the connection point and the disconnecting means, routing and protection of this service-entrance conductor, bonding neutral to ground at the new service disconnect, and running a grounding electrode conductor from the bonding jumper to the existing grounding electrode. Yes, it appears that there may be some parallel paths for the neutral currents, but they do not appear objectionable since similar multiple bonding jumpers in close proximity are shown in Article 250 in the NEC Handbook where multiple services are involved.

Photo 2. Utility-required disconnect — PV AC disconnect too?

Question: We have numerous commercial buildings in our jurisdiction with 480-volt, 4-wire services that are over 1000 amps and have main service-entrance disconnects as main breakers with attached or internal ground-fault protection devices. What are the issues that should be considered when looking at a plan to backfeed a panel on the load side of this main GFP breaker with the output of a photovoltaic inverter?

Answer: Briefly: (1) Has the GFP device been evaluated for backfeeding? Most new ones are, but older units may not have been evaluated. UL does not do this particular evaluation; only the manufacturer can provide the necessary information. The breaker may not be marked "Line” or "Load,” which indicates that it has been evaluated for backfeeding, but this has no bearing on the suitability for the GFP device for back feeding. (2) Does the inverter ac output circuit have a ground-fault protection device connected to protect loads from ground-fault currents originating from the inverter? The internal dc ground-fault protection device does not meet this function. (3) Has a fault analysis been accomplished to determine how ground-fault currents will divide between the main GFP and the inverter GFP and what the proper trip settings for each should be? See a White Paper on this subject on the author’s web site below.

Photo 3. Inverters with internal AC and DC disconnects plus external disconnects

Disconnect Questions

Question: Can an unfused disconnect used to meet a local utility requirement be also used as the 690.15 maintenance disconnect for the inverter? The disconnect is not locked by the utility and is located near the service disconnect and the meter on the outside of the building.

Answer: Usually the utility will have no objections to this dual use of the utility-required disconnect, but it never hurts to verify. In order to meet the intended safety requirements of 690.15, the disconnect should be located near or at least within sight of the inverter. This location requirement would allow the inverter to be maintained in a safe manner by opening this ac disconnect, opening the dc disconnect, verifying that both are open and then working on the inverter as necessary. An inverter that is not mounted within sight of this utility-required disconnect may require that an additional ac disconnect be mounted adjacent to the inverter location.

Question: Can the disconnects, either ac or dc or both, that may be internal to the inverter be used as the 690.14 dc PV disconnect and/or the 690.15 required disconnects?

Answer: If the inverter is mounted in the location required by 690.14 for the dc PV disconnect, an internal dc disconnect might meet that disconnect requirement. However, meeting the 690.15 maintenance disconnects with any internal disconnects may pose certain problems. This is a discussion that the PV installer and the AHJ will have to have. Where the internal disconnects are mounted in a section of the inverter that is separate from the inverter electronics and the inverter electronics section can be removed for service while the disconnect section remains attached to the wall and the dc and ac conduits, then it would appear that the safety intent can be met. However, if the internal disconnects are in one enclosure with the inverter proper, there is the possibility that a less-than-fully-qualified person might run into trouble by unintentionally pulling live dc cables through the conduit knockout when removing the inverter for service. Recent internal disconnect failures, a few disconnect fires, and recalls of some inverters for problems in the disconnect section have caused many AHJs to reevaluate their position on the internal disconnect.

Photo 4. Microinverters — disconnects required?

Question: Microinverters are mounted on roofs in not readily accessible areas. How can the disconnect requirements of 690.14 and 690.15 be met? It would appear that 690.14(D) would apply since the inverters are mounted in these roof top areas, but there are no disconnects being used. Should I require ac and dc disconnects for each microinverter?

Answer: Before requiring large and expensive ac and dc disconnects for each inverter, check with the microinverter manufacturer to determine if the connectors on the microinverter have been evaluated as load-break-rated disconnects. While the typical MC 3 or MC 4 PV disconnect on a PV module is only a recognized component because it cannot pass the listing requirements at 600 volts dc, those connectors can be evaluated as load-break disconnects at the lower operating voltages (typically less than 80 volts) of the microinverters. At least one manufacturer of microinverters has had the ac and dc connectors so evaluated.

Photo 5. FMC from the roof

Where an additional ac disconnect is deemed necessary, the common 60-amp pullout ac HVAC unfused disconnect can meet the requirements and provides a transition point between the microinverter cable and the circuit to the ac panel. It is usually cheaper than many other outdoor-rated pull or junction boxes.

Circuit Questions

Question: The electrician ran flexible metal conduit from the roof penetration through the house to the dc disconnect and the inverter located in the basement. Is this type of installation permitted per the NEC?

Answer: Yes, as of the 2005 NEC, Section 690.31(E) allowed metal raceways to be used for this interior circuit run between the rooftop mounted PV system and the readily accessible dc disconnect/inverter. This would include flexible metal conduit (Type FMC). In the 2011 NEC, a metallic cable assembly, Type MC was added. Type AC metallic cable assemblies, particularly those with aluminum outer jackets, are not approved or listed for use in direct current (dc) circuits.

Question: Is the inside of a house or building with locked doors and windows considered a readily accessible location for meeting the Article 230 service entrance and Article 690 PV disconnecting means location requirements?

Photos 6A and 6B (inset). Steel door and high security lock — readily accessible?

Answer: Excellent question, and one that needs further clarification in the Code. Fire fighters will usually call the utility to have the ac power disconnected from a building before entering an area that might have energized circuits. When the utility is unable to get to the location in a timely manner, the fire fighters are reluctant to remove the utility meter due to the safety hazards and legal issues involved. In life safety issues, they will pull the utility meter thereby de-energizing the ac circuits.

But what about that inside-the-house dc disconnect for the PV system? They know that it is there because of the code-required directories and placards on the outside meters and service equipment. Fire fighters have told me that they have master keys for many locks; and for the high security locks, there is always the fire axe. However, the answer to this question remains unclear in the NEC. Is the inside of a locked building considered a readily accessible area in which an ac service-entrance disconnect or a dc PV disconnect can be located?

Question: Section 690.47(C)(3) in the 2011 NEC allows the function of the PV inverter dc grounding-electrode conductor to be combined with the function of an inverter ac equipment grounding conductor in a single conductor meeting the most stringent requirements of either conductor. In many older electrical systems and in some newer ones, an outbuilding such as a barn or garage is connected to the main service panel with a feeder that uses the neutral as both the grounded circuit conductor and as the equipment grounding conductor as allowed by 250.32(B) Ex. If a utility-interactive PV system is installed on the outbuilding, can that combined neutral/ac equipment grounding conductor be used as the 690.47(C)(3) "grounding” conductor for the inverter?

Answer: Section 690.47(C)(3) addresses only the grounding-electrode and equipment grounding conductors from the inverter. Under normal operation, neither of these conductors carries current, whereas the combined ac neutral/equipment grounding conductor allowed by 250.32(B) Ex would normally be a current-carrying conductor. Although the NEC does not explicitly address this combination, I tend to think that these two functions should not be further combined into a single conductor in that feeder between the main panel and the outbuilding. One reason that comes to mind is that lightning surges induced from the PV array now have a relatively easy path along the neutral into the service equipment. However, the next question may have some bearing on this issue.

Answer: Although 690.47(C) in the 2008 is a bit murky, I believe both editions of the Code allow this combined conductor to be terminated at a grounding bus bar in the nearest ac panel that has an ac grounding electrode conductor connected to a grounding electrode that meets the requirements of the Code. Such a panel would certainly include the main service-entrance panel and also any feeder panel that has the necessary grounding. With respect to the previous question, the remote building that has the 250.32(B) Ex "grounding” system is required to have a grounding electrode at the outbuilding. It would appear that the PV inverter could be mounted in this location with the combined dc grounding electrode conductor/ac equipment grounding conductor terminated at the grounding bus bar in the outbuilding panel. In this case, the combined neutral/equipment grounding conductor between the buildings would not be involved in the inverter grounding requirements.

Ratings and Calculations Questions

Question: I’m a building inspector and I have a few questions regarding STC ratings. I know that the NECrequires all PV modules to be marked with its maximum voltage, open-circuit voltage, short-circuit voltage, etc., and common sense will tell me that the conductors and OCPD must be sized based on that info. The problem is, I can’t find anywhere in the NEC that states exactly that, other than the word "rated” in 690.8 and 690.9. So I guess I’m asking: What forces us to use STC ratings when sizing a PV system? And are STC ratings the only ratings marked on modules? If another testing standard was marked on the modules and the modules were listed, would the Code require the wires and OCPD to be sized based on that info instead of STC?

Answer: The key is NEC Section 110.3(B), which requires that we use the instructions and labels on a listed product. The label on the back of a PV module is required by UL Standard 1703 and the values on that label are based on testing under the Standard Test Condition as required by the standard. As far as I know, UL Standard 1703 is the only standard being used in the U.S. to certify/list PV modules and that standard is being harmonized with the European IEC standards. In both the UL and the IEC standards, Standard Test Conditions are used to rate the module. NEC Informative Annex A lists UL Standard 1703 as the applicable standard for flat plate PV modules. There are no values on the back of the module other than the STC values. So, the rated values required in 690.8 are the values marked on the back of the module and they would be used in the circuit sizing and overcurrent protection. In a similar manner, the motor nameplate ratings in terms of locked-rotor current and full-load current would be used in determining the circuit sizing for that motor. Yes, there are other specifications sometimes listed for modules in the technical specification sheets or in other documents. For example, the temperature coefficients are listed in specification sheets and used to calculate the cold weather, open-circuit voltage as required by 690.7. In some cases, PVUSA Test Conditions (PTC) are given, but these typically are used for performance estimations and are not involved with Code calculations.

Photo 7. Cold weather Voc calculations are important.

Question: I am checking a set of plans for the calculations on the cold-weather open-circuit voltage (Voc) and I find that some of the module specification sheets show a Voc temperature coefficient in degrees K. In the January-February 2009 IAEI News article on "PV Math,” you described the method of using coefficients with degrees Celsius (C). But what do I do with these numbers in degrees K?

Answer: You use the numerical values in coefficients that are based on degrees Kelvin (K), in the same way you use the coefficients based on degrees Celsius (C). A change in temperature of one degree K is the same as a change in temperature of one degree C. The difference is that the Kelvin temperature scale is based on zero being at an absolute zero temperature where all molecular motion stops, but the Celsius temperature scale has a zero based on the freezing point of water. The zero point on the scale does not affect our calculations.

For Additional Information

See the web site below for a schedule of presentations on PV and the Code.

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 (1.91) 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:

And, yes, it may be updated to the 2008 and 2011 Codes sometime this year.

Read more by John Wiles

Tags:  Featured  January-February 2012  Perspectives on PV 

Share |
PermalinkComments (0)

Inspecting PV Systems

Posted By John Wiles, Tuesday, November 01, 2011
Updated: Wednesday, January 16, 2013

Plan Reviewers and Inspectors. What Do You Need?

Photovoltaic (PV) power systems are becoming more numerous, larger and more complex. Inspectors and plan reviewers have limited time to deal with these new systems and still carry on the routine electrical system inspections that have been done for 100 years or more. I intend for this "Perspectives on PV” articles to provide you with information on the Code requirements for these systems and also give you information on how to make the plan reviews and inspections easier and faster.

Inspecting PV Systems

Inspecting PV Systems

What do you need to know about concerning PV systems? Give me a call or drop me an e-mail and let me know what you would like to see in these articles. There will be a time delay since I am writing this November-December 2011 IAEI News article in August. In a hurry for an answer? Try the e-mail and I’ll try to get a fast response.

On the Front Lines

Plan reviewers and inspectors bear a heavy responsibility for the safety of the public when it comes to electrical systems, including PV systems. While most residential and small commercial electrical systems have not changed much over the past few decades or so, PV systems now have transformerless inverters for ungrounded PV arrays, microinverters, AC PV modules, dc to dc converters in the PV array and dc PV arc fault circuit protection. Couple those new "toys” with the dc current-limited outputs from the PV modules and we have a very dynamic, constantly evolving situation.

I know that many jurisdictions do not have a plan review section or person and that many inspectors only have 15–30 minutes allocated to perform a residential inspection. We all know that there are both qualified and unqualified people doing electrical installations, including PV systems. And with the significant amounts of money flowing into green electrical systems, there are many people jumping on the bandwagon that should not even be near the parade.

In this Perspectives on PV, I will share with you a PV installer checklist that covers the more import Code requirements for PV systems. The checklist will show 2005, 2008 and 2011 requirements and the differences will be noted.

Photo 2. AC or DC disconnect?

Photo 2. AC or DC disconnect?

Since jurisdictions vary in the availability of a plan review department and the time available for the inspection differ, I will not attempt to separate the items that would be accomplished at the plan review stage and those that need to be done at the on-site inspection. And, yes, I have tried many times to read a conductor size and type on a hot sweaty day when the conductors are cut to minimum length inside a disconnect—it sometimes is just not possible.

The following checklist is available on the author’s web site (see below) and it is double spaced for better readability.



  • PV modules listed to UL Standard 1703? [110.3] [690.4(D)]

a. Mechanical Attachment

  • Modules attached to the mounting structure according to the manufacturer’s instructions? [110.3(B)]
  • Roof penetrations secure and weather tight? [110.12]

b. Grounding

  • Each module grounded using the supplied hardware, the grounding point identified on the module and the manufacturer’s instructions? Note: Bolting the module to a "grounded” structure usually will not meetNECrequirements [110.3(B)]. Array PV mounting racks are usually not identified as equipment-grounding conductors. [Note 690.43(C) and (D) in 2011 have additional provisions and allowances for grounding with mounting structures.]
  • Properly sized equipment-grounding conductors routed with the circuit conductors? [690.45] Note differences between 2005, 2008 and 2011NEC.

c. Conductors

  • Conductor type? —If exposed: USE-2, UF (usually inadequate at 60°C), or SE, 90°C, wet-rated and sunlight-resistant. [690.31(B)] (2008 NEC restricts exposed single-conductor wiring to USE-2 and listed PV/Photovoltaic Wire/Cable)—If in conduit: RHW-2, THWN-2, or XHHW-2 90°C, wet-rated conductors are required. [310.15]
  • Conductor insulation rated at 90°C [UL-1703] to allow for operation at 70°C+ near modules and in conduit exposed to sunlight (add 17–20°C to ambient temperature-2005NEC)[see Table 310.15(B)(2) in the2008 NEC] [Table 310.15(B)(3)(c)]
  • Temperature-corrected ampacity calculations based on 156% of short-circuit current (Isc), and the corrected ampacity greater than 156% Isc rating of overcurrent device? [690.8,9]

Note: Suggest temperature derating factors of 65°C in installations where the backs of the module receive cooling air (4″ or more from surface) and 75°C where no cooling air can get to the backs of the modules. Ambient temperatures in excess of 40°C may require different derating factors.

(2011 690.8 substantially updates ampacity calculations to parallel calculations in other sections of theNEC.)

  • Portable power cords allowed only for tracker connections? [690.31(C), 400.3,7,8]
  • Strain reliefs/cable clamps or conduit used on all cables and cords? [300.4, 400.10]
  • Listed for the application and the environment? Fine stranded, flexible conductor cables properly terminated with terminals listed for such conductors? [690.31(E)(4)]
  • Cables and flexible conduits installed and properly marked? [690.31(E)]
  • Exposed conductors in readily accessible areas in a raceway if over 30 volts? [690.31(A)] Note: Raceways cannot be installed on modules. Conductors should be installed so that they are not readily accessible.


  • Overcurrent devices in the dc circuits listed for dc operation? If device is not marked dc, verify dc listing with manufacturer. Auto, marine, and telecom devices are not acceptable.
  • Rated at 1.25 x 1.25 = 1.56 times short-circuit current from modules? [UL-1703, 690.8, module instructions] Note: Both 125% factors are now in theNEC, but the duplicate 125% should be removed from the modular instructions in calendar year 2011. Supplementary listed devices are allowed in PV source circuits only, but branch-circuit rated devices are preferred. [690.9(C)].
  • Each module or series string of modules have an overcurrent device protecting the module? [UL-1703/NEC110.3(B)] Note: Frequently, installers ignore this requirement marked on the back of modules. Listed combiner PV combiner boxes meeting this requirement are available. One or two strings of modules do not require overcurrent devices, but three strings or more in parallel will usually require an overcurrent device. The module maximum series fuse must be at least 1.56 Isc.
  • Located in a position in the circuit to protect the module conductors from backfed currents from parallel module circuits or from the charge controller or battery? [690-9(A) FPN, NEC-2008] Informational Note, 2011.
  • Smallest conductor used to wire modules protected? Sources of overcurrent are parallel-connected modules, batteries, and ac backfeed through inverters. [690-9(A)]
  • User-accessible fuses in "touch-safe” holders or fuses capable of being changed without touching live contacts? [690.16] Strengthened for 2011 to include distance between overcurrent device and disconnect.
Photo 3. Double Lugging
Photo 3. Double Lugging


  • Pressure terminals tightened to the recommended torque specification?
  • Crimp-on terminals listed and installed with listed crimping tools by the same manufacturer? [110.3(B)]
  • Twist-on wire connectors listed for the environment (i.e., dry, damp, wet, or direct burial) and installed per the manufacturer’s instructions?
  • Pressure lugs or other terminals listed for the environment? (i.e., inside, outside, wet, direct burial)
  • Power distribution blockslistedand not just UL Recognized?
  • Terminals containing more than one conductor listed for multiple conductors?
  • Connectors or terminals using flexible, fine-stranded conductors listed for use with such conductors? [690.31(F), 690.74]
  • Locking (tool-required) on readily accessible PV conductors operating over 30 volts [690.33(C)]


  • Charge controller listed to UL Standard 1741? [110.3] [690.4(D)]
  • Exposed energized terminals not readily accessible?
  • Does a diversion controller have an independent backup control method? [690.72(B)(1)]


  • Disconnects listed for dc operation in dc circuits? Automotive, marine, and telecom devices are not acceptable.
  • PV disconnect readily accessible and located at first point of penetration of PV conductors?
  • PV conductors outside structure until reaching first readily accessible disconnect unless in metallic raceway? [690.14, 690.31(F)]
  • Disconnects for all current-carrying conductors of PV source? [690.13]
  • Disconnects for equipment? [690.17]
  • Grounded conductorsnotfused or switched? Bolted disconnects OK.

Note: Listed PV Centers by Xantrex, Outback, and others for 12, 24, and 48-volt systems contain charge controllers, disconnects, and overcurrent protection for entire dc system with possible exception of source circuit or module protective fuses.

6. INVERTERS (Stand-Alone Systems)

  • Inverter listed to UL Standard 1741? [110.3] [690.4(D)] Note: Inverters listed to telecommunications or other standards do not meetNECrequirements.
  • DC input currents calculated for cable and fuse requirements? Input current = rated ac output in watts divided by lowest battery voltage divided by inverter efficiency at that power level. [690.8(B)(4)]
  • Cables to batteries sized 125% of calculated inverter input currents? [690.8(A)]
  • Overcurrent/Disconnects mounted near batteries and external to PV load centers if cables are longer than 4–5 feet to batteries or inverter?
  • High interrupt, listed, dc-rated fuses or circuit breakers used in battery circuits? AIR/AIC at least 20,000 amps? [690.71(C), 110.9]
  • No multi-wire branch circuits where single 120-volt inverters connected to 120/240-volt load centers? [100—Branch Circuit, Multi-wire], [690.10(C)]


  • None are listed.
  • Building-wire type cables used? [Chapter 3] Note: Welding cables, marine, locomotive (DLO), and auto battery cables don’t meetNEC. Flexible, listed RHW, or THW cables are available. Article 400 flexible cables larger than 2/0 AWG are OK for battery cell connections, but not in conduit or through walls. [690.74, 400.8] Flexible, fine stranded cables require very limited, specially listed terminals. See stand-alone inverters for ampacity calculations.
  • Access limited? [690.71(B)]
  • Installed in well-vented areas (garages, basements, outbuildings, and not living areas)? Note: Manifolds, power venting, and single exterior vents to the outside are not required and should be avoided.
  • Cables to inverters, dc load centers, and/or charge controllers in conduit?
  • Conduit enters the battery enclosure below the tops of the batteries? [300.4]
    Photo 4. Undetected Ground Fault

    Photo 4. Undetected Ground Fault

    Note: There are no listed battery boxes. Lockable heavy-duty plastic polyethylene toolboxes are usually acceptable

8. INVERTERS (Utility-Interactive Systems)

  • Inverter listed to UL Standard 1741 and identified for use in interactive photovoltaic power systems? [690.4(D), 690.60] Note: Inverters listed to telecommunications and other standards do not meetNECrequirements.
  • Backup charge controller to regulate the batteries when the grid fails? [690.72(B)(1)]
  • Connected to dedicated branch circuit with back-fed overcurrent protection? [690.64]
  • Listed dc and ac disconnects and overcurrent protection? [690.15,17]
  • Total rating of overcurrent devicessupplyingpower to ac load center (main breaker plus backfed PV breaker) must be less than load-center rating (120% of rating in residences) [690.64(B)(2)]. The2008 NECallows the 120% breaker total on commercial installations and residential system ONLY if the PV breaker is at the opposite end of the busbar from the main utility breaker. No change for 2011.


  • Only one bonding conductor (grounded conductor to ground) for dc circuits and one bonding conductor for ac circuits (neutral to ground) for system grounding? [250] Note: The main dc bonding jumper will generally be located inside inverters as part of the ground-fault protection devices. On stand-alone systems, the dc bonding jumper may be in a separate ground-fault detection and interruption device or may be built in to the charge controller.
  • AC and dc grounding electrode conductors connected properly? They may be connected to the same grounding electrode system (ground rod). Separate electrodes, if used, must be bonded together. [690.41,47] The 2008NECin 690.47 allows a combined dc grounding electrode conductor and an ac equipment-grounding electrode, but the conditions and requirements are numerous. [690.47]. (2011NECclarifies and combines 2005 and 2008 690.47(C) requirements.)
  • The 2008NEC690.47(D) array grounding requirement was removed in 2011NEC.
  • Equipment grounding conductors properly sized (even on ungrounded, low-voltage systems)? [690.43, 45, 46]
  • Disconnects and overcurrent in both of the ungrounded conductors in each circuit on 12-volt, ungrounded systems or on ungrounded systems at any voltage? [240.20(A)], [690.41]
  • Bonding/grounding fittings used with metal conduits when dc system voltage is more than 250-V dc? [250.97]

10. CONDUCTORS (General)

  • Standard building-wire cables and wiring methods used? [300.1(A)]
  • Wet-rated conductors used in conduits in exposed locations? [100 Definition of Location, Wet]
  • Insulations other than black in color will not be as durable as black in the outdoor UV-rich environment.
  • DC color codes correct? They are the same as ac color codes—grounded conductors are white and equipment-grounding conductors are green, green/yellow, or bare. [200.6(A)] Ungrounded PV array conductors on ungrounded PV arrays willnotbe white in color.

For Additional Information

The US Department of Energy funding for providing inspectors and the PV Industry with telephone and e-mail support from the author was terminated on March 1, 2011. Answers to your questions may be delayed or not answered at all depending on future funding. Consultation services are available on a contracted basis. E-mail: Phone: 575-646-6105

See the web site below for a schedule of presentations on PV and theCode.

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 (1.91) 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:

Read more by John Wiles

Tags:  Featured  November-December 2011  Perspectives on PV 

Share |
PermalinkComments (0)

Two Important Inspection Areas & One for the Plan Reviewers

Posted By John Wiles, Thursday, September 01, 2011
Updated: Wednesday, January 16, 2013
Photovoltaic (PV) power systems have PV modules and PV arrays that will be producing dangerous amounts of voltage and current for the next 50 years or more. If the inverters in these systems do not fail or are maintained in operating condition, significant amounts of energy will be supplied to local loads and to the connected utility grid. There are two areas of PV systems that deserve the attention of inspectors to ensure the safety of the public over these very long periods of time. One is proper grounding of the PV array and the entire system and the other is ensuring that the ac output connections have been properly made to the existing premises wiring. Plan reviewers can look at conductor types with an eye to durability and longevity.

Grounding, Grounding, Grounding

Grounding is particularly important to the long-term safety of a PV system. See the "Perspectives on PV” article in the May-June 2010 issue of theIAEI Newsfor some history and a more complete discussion of this subject. In areas where heavy rains are infrequent, the PV modules will accumulate layers of dirt, soot and bird droppings that will reduce the electrical output. Where these modules are visible, where conscientious (green minded) people are involved, or where power purchase agreements are involved, these modules will get washed (photo 1). This is done generally with a garden hose and sometimes at close range. Few people realize that the conductive connections inside the modules and the exposed single conductor cables from the modules to other portions of the PV system operate from 60 to almost 600 volts direct current (dc) — depending on system design and configuration — and some of the newer systems with micro inverters or AC PV modules have 120- or 240-volt alternating current (ac) circuits on the roof. Standards written by Underwriters Laboratories have established that shock hazards can exist at voltages as low as 30 volts in wet conditions.

Photo 1. Cleaning the PV array may be hazardous to your health.

Photo 1. Cleaning the PV array may be hazardous to your health.

People, who grew up on a farm, learned at an early age whatnotto do against an electric fence. Damage to the PV module or to any exposed conductors may pose similar shock hazards to unwary people washing their PV modules. Also, workers on the roof repairing the roof, gutters, HVAC equipment and the like could also be exposed to shock hazards, especially if the roof and the PV array are wet from recent rains.

When PV systems are installed in full compliance with the requirements of theNational Electrical Code(NEC), and any local codes, and with high levels of workmanship, these PV systems will be essentially hazard free for many years. It is up to the inspector to ensure that the installation is code-compatible and that the workmanship is high.

Module Grounding

As a first requirement, the grounding instructions and labels provided with or on the PV module should be followed. If any type of listed grounding device is suggested, used, or supplied, that device must be used in accordance with the instructions provided and the instructions and labels for the module. Unfortunately, installation instructions for installing some of the common, over-the-counter grounding devices, like the lay-in lug, are not easy to find, even when available.

Photo 2. Grounded for 50 years?

Photo 2. Grounded for 50 years?

The instruction manuals for many modules have not been reviewed recently by the listing/certification agencies (UL, ETL, CSA, TUV). This is done every five years and in many cases, the grounding instructions were not properly reviewed initially because of the way in which UL Standard 1703 (Flat Plate PV modules) is written with respect to grounding the module. In late 2007, UL issued a Critical Requirements Decision on UL 1703 with several reinterpreted and reemphasized requirements.

Dissimilar metals should not come into contact. At the grounding point, the field-installed copper conductor should not touch the aluminum surface of the module frame. If these two metals come into contact, and there is moisture in the air, the aluminum surface may be eaten away causing the contact/connection to fail.

Where an electrical contact is made to an aluminum framed PV module, the clear coating anodizing and oxidation should be penetrated or removed. In some cases, listed grounding devices have sharp contact points that can penetrate those insulators. In other cases, the module frame must be prepared to remove these insulators before the grounding device is used.

Photo 3. Load side, supply side or both?

Photo 3. Load side, supply side or both?

Although stainless steel can come into contact with aluminum, a stainless steel washer may not be adequate to isolate a copper wire from aluminum when the electrical connection is through a screw holding the assembly together. Normally in electrical equipment, mechanical fasteners like screws, washers, and nuts are used to provide mechanical force to hold the electrically conductive parts together. The screw is not normally intended to carry current and steel is not a very good conductor. When a stainless steel flat washer is placed against a module frame, there may be little current flow through the washers unless the module frame coatings have also been removed from the aluminum surface under the washer. The same situation applies when a grounding lug is attached to a module frame. The module coatings must be removed.

Many grounding devices are tin-plated copper. The instructions, when they can be found, for attaching these listed devices show flat washers against the grounding device to prevent split ring or star lock washers from digging into the relatively soft copper thereby losing their compression force. Also, the use of any type of washer that digs into the tin plating on the grounding device may remove that plating, exposing the underlying metal to corrosive/cathodic action.

Inverter AC Output Circuits

Electrical power systems are constantly being changed in both residential and commercial locations. Not only are loads being changed (without any qualified supervision), but additional circuits may be added at any time. The Smart Grid and the Smart Home may have significant impacts on these wiring systems. See "Perspectives on PV” in the July-AugustIAEI Newsfor more details on the future. If the requirements of theNECfor connecting the ac outputs of the PV systems are not carefully followed, there may be the possibility of inadvertent overloads due to the future changes. Although load circuits may be impacted by the Smart Grid and Smart Home, the utility-interactive inverter may be a unique device for some time to come and the typical electrician may not be familiar with the inverter ac output characteristics that drive theCoderequirements. Taps may be added for new loads and these taps can be detrimental to the electrical system if shortcuts are taken today in the installation of the PV system.

Photo 4. Incorrect conductor being used outdoors

Photo 4. Incorrect conductor being used outdoors

First, inspectors and PV installers should recognize that the requirements of 690.64(B)/705.12(D) will generally apply to the ac output circuits and panelboards/load centers of both load-side PV connections (690.64(B)/705.12(D) and to supply-side connections (690.64(A)/705.12(A). The application of load-side requirements to supply-side connections is not widely realized, but as soon as the new supply-side service disconnect is passed, all circuits toward the inverter may have to meet load-side Code requirements (Photo 3). See "Perspectives on PV” in the November-December 2010 issue of theIAEI News.

As noted in that earlier article, conductors involved in these load-side circuits may be larger than normal, but this is not always a bad thing for the installer. The anti-islanding circuits react to the ac voltage at the inverter output terminals. This voltage is affected by the voltage drop between the inverter output terminals and the main service disconnect or the meter. In locations where the utility voltage is on the high side of the nominal voltage (120, 208, 240, 277, 480), voltage drop (really voltage rise) to the inverter may cause the voltage at the inverter to be outside the anti-islanding range, causing the inverter to shut down. Keeping voltage drop below the typical 3–5% will minimize this problem. The larger conductors required by 690.64(B) will assist in minimizing the voltage drop/rise.

For some reason, some PV systems installers sometimes forget that theNECapplies to medium voltage (over 600 volts) premises wiring. Even at 12 kV, premises wiring circuits that may carry PV currents to the utility grid are subject to 690.64/705 requirements.

Something for Plan Reviewers

Plan reviewers can check the types of conductors being used and help the PV installer get code-compliance and added durability.

USE-2 conductors undergo a 350-hour accelerated UV exposure test. This length of test time is not sufficient to allow the USE-2 conductors to be marked "Sunlight Resistant” because that marking requires a conductor or product so marked to be tested for 720 hours.

Engineers at the cable manufacturers (not the sales staff) tell me that the PV industry is requesting USE-2 with colored insulation in addition to the basic black. Requested colors are green, red, and white and the manufacturers are making and selling those colors (photo 4). All of these cables that are marked USE-2 have passed the 350 hours of UV testing. However, black USE-2 has significantly more "carbon black” than the colored insulations have and the carbon black has different particle sizes. Carbon black gives the black-colored insulation significantly greater UV resistance than the cables with lesser amounts or no carbon black. While all USE-2 cables pass the 350-hour UV test, the black cable should last much longer in the PV environment than cable with colored insulation. Note that with an annual average of 6 hours of peak sun per day in the sunny Southwest, the exposed USE-2 conductors used in PV systems are subject to 2100 hours of sunshine each year.

InNECSection 200.6, an exception allows the grounded, exposed PV conductors to be marked with a white marking even though they are smaller than 4 AWG. With this marking allowance, there is no reason for anyone to use any colored insulation. Basic black is beautiful and suitable for all occasions—as any woman will tell you.

Now on to another issue facing the PV installer. UL Standard 4703 allows the new PV Cable/PV wire to be made with thermoset insulation (synthetic rubber, found on cables like USE-2 and XHHW) or with thermoplastic insulation (PVC, found on cables like THHN). Conductors with either insulation must pass the 720-hour UV tests and all will be marked "Sunlight Resistant.” In the hot and bright desert Southwest, cables with grey PVC, thermoplastic insulation marked "Sunlight Resistant” have failed in less than 10 years of exposure. Cables such as type UF have had the outer jackets disappear, and flexible nonmetallic conduits have fallen apart in periods much shorter than the warranted life of a PV module. On the other hand, USE-2 cables with black insulation made with thermoset insulation like cross-linked polyethylene (XLP or XLPE) have been in service, on hot roofs in full sun all day, for more than 30 years with no apparent signs of degradation.

The bottom line is: For exposed use in PV systems, single conductor cables/conductors with thermoset insulation (cross-linked polyethylene) in black are highly recommended. I would forego the use of colored insulation and PVC-insulated products in these exposed installations. See NEC Table 310.13(A) / 310,104(A) and the Cable and Wire Marking Guide in the UL White Book for more information


Plan reviewers and inspectors are a critical link in ensuring the long-term safety of the public where PV systems are involved. The highest levels ofCode-compliance and workmanship are required. The fully informed inspector and plan reviewer will ensure that this goal is achieved.

Errata.In the January-February 2011IAEI News"Perspectives on PV,” in Example 5, the equation should be 130 x 0.8 x 0.82 = 85.3, not 107. The answer and the rest of the example are correct.

For Additional Information

The US Department of Energy funding for providing Inspectors and the PV Industry with telephone and e-mail support from the author was terminated on March 1, 2011. Answers to your questions may be delayed or not answered at all depending on future funding. Consultation services are available on a contracted basis. E-mail: Phone: 575-646-6105

See the web site below for a schedule of presentations on PV and theCode.

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 (1.91) 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:

And, yes, it may be updated to the 2008 and 2011 Codes sometime this year.

Read more by John Wiles

Tags:  Featured  Perspectives on PV  September-October 2011 

Share |
PermalinkComments (0)

A Critical Look at Load Side Utility-Interactive PV Inverter Connections 690.64(B) / 705.12(D)

Posted By John Wiles, Friday, July 01, 2011
Updated: Wednesday, January 16, 2013
The NEC in sections 705.12(D) / 690.64(B) allows utility-interactive photovoltaic inverters to be connected on the load side of the service disconnect. This requirement has been in theCodesince the late 1980s when PV Article 690 first appeared. Except for a slight change in 2008, the requirement has been largely unchanged. A critical examination of the requirement and how it can be applied as well as various proposals that have been rejected over the years may yield insights on what is needed in the future.

Photo 1. 400-amp load center, 300-amp main. Internal supply side and load side PV connections are possible.

The Basic Requirement

This section of Code was written to address a general condition where any panelboard busbar or conductor might be fed by multiple sources of power that are connected to the busbar or conductor through overcurrent devices. Although the 2008 NEC 690.64(B) appears to restrict the connection point, in fact nearly any point on a load-side circuit (inside a panelboard or on the conductors of a feeder or branch circuit) may, and has, served as a connection point for either a PV inverter or for an additional load circuit. Of course, there are numerous restrictions and requirements on making such connections, but in general, all of the load-side wiring is up for grabs and the connections and circuits must be protected.

There are no restrictions in this code requirement as to the particulars of any specific installation. There are no restrictions on where the multiple power sources might be connected on the busbar or conductor nor are there any limits on the number of overcurrent devices. There are no restrictions on the loads connected to the busbar or conductor either in terms of their connection point or the rating of the overcurrent device and, in fact, loads are not specifically addressed in the section. When applying this requirement, no assumptions should be made as to the configuration of the circuit with respect the location of connections (taps) and the number, magnitude and locations of any sources or loads. Some people even feel that the code requirement was written to "Protect people from doing harm—in the future.”

Photo 2. Load side connection. Conductor size to the PV inverter too small to meet 690.64(B) requirements.

This is the manner in which many code requirements are formulated. The requirement is written in general terms and then the general requirement is modified by exceptions (restrictions or allowances) or additions to the requirement.

From an engineering point, the basic requirement is sound. A conductor or busbar will be prevented from being overloaded if the rating of that busbar or the ampacity of that conductor is greater than or equal to the sum of the ratings of all overcurrent devices supplying it [see 690.64(B)(2) in2005 NEC]. Note that the requirement refers only to theratingof thesupplyovercurrent devices, not to any calculated currents and it does not refer to any load overcurrent devices.

Because dwelling unit load centers are usually not fully loaded and the Chapter 2 load calculations usually result in light panel loadings, the 690.64(B) requirement up to the2008 NECallowed a dwelling exception to the extent that the sum of the ratings of the supply overcurrent devices could exceed the rating of the busbar or conductor up to 120%. This allowance for dwelling units would allow up to 20 amps of backfed PV breaker to be installed on a 100-amp rated panel that had a 100-amp main breaker. By calculation:

  • 120% of 100 amp busbar = 120 amps.
  • 120 amps allowance -100 amp main = 20 amps for a backfed PV breaker.
  • 20 (PV breaker) + 100 (main breaker) = 120 amps sum of supply breakers which is less thanor equal to 120% of busbar rating which is120% of 100 amps = 120 amps.

In a similar manner, a 200-amp rated panel with a 200-amp main breaker would be allowed to have up to a 40-amp backfed PV breaker.

Photo 3. Load-side connection on output main breaker. Conductors to PV disconnect/overcurrent protection should be as large as the main output conductors.

In the 2005 and earlier editions of this section, non-dwelling, commercial PV installations did not have the 120% exception and the basic requirement applied. That meant that in a commercial installation where a main breaker in a load center was rated the same as the busbar, no PV could be connected. Also when a conductor ampacity was the same as the OCPD for that conductor, no load-side connection for PV could be made. Supply-side connections, 690.64(A) / 705.12(A), were usually required.

In at least five code cycles, various changes and modifications have been proposed to change the basic requirement and wording. CMP-13 and now CMP-4 (2011 and subsequent editions of the Code) have ruled that theonlyway to protect this general busbar or conductor, that has no restrictions, is that the busbar or conductor must have a rating or an ampacity equal to or greater than the sum of the ratings of all overcurrent devices supplying that busbar or conductor.

Various Other Connections Can Be Safe

As the time progresses, we have seen various wiring configurations for that general, unrestricted, busbar or conductor that might allow exceptions to the basic requirements. These wiring configurations are discussed among inspectors, electricians, conductor and panelboard manufacturers and, as they are vetted to be safe, proposals are made to change theNEC. These are in the form of exceptions or modifications to the basic requirements.

This process is not unique to 690.64(B)(2) / 705.12(D)(2) and similar actions have been taken throughout the NEC.

With respect to 690.64(B)(2) / 705.12(D)(2), it has long been recognized that if there are only two supply overcurrent devices and that they are opposite ends of the busbar or conductor, then even if unrestricted loads or load taps are added between the two supply overcurrent devices, there is nowhere on the conductor or busbar where the currents may exceed the rating of the largest overcurrent device.

An internal CMP revision of 690.64(B) for the 2008 NEC recognizes this fact and requires that in a panelboard, if the two supply overcurrent devices are at opposite ends of the busbar (and possibly a conductor), the sum of the ratings of the busbar or conductor may exceed the current rating of the busbar by 20%. The assumption is made that actual load on the panel will not exceed the panel or conductor rating in most residential and commercial locations. Unfortunately, actual experience dictates that plug loads are essentially unrestricted and unmonitored and may result is loads higher than calculated by the installing electrician. But even if the actual loads on a busbar or conductor exceed numerically the rating of the busbar or the ampacity of the conductor, with the supply overcurrent devices at opposite ends, there is no place on that busbar or conductor where the currents will or can exceed the rating. This revision allows the 120% exception to be applied to both dwelling units and non-dwelling installations if the overcurrent device location requirement can be met.

Figure 1. The conductor has a 60-amp breaker at the utility feed end and it has a 40-amp backfed breaker at the PV inverter end. If the conductor is not tapped for loads or other sources, then the highest current that it could ever see under any normal or fault condition is 60 amps, the rating of the highest connected supply overcurrent.

Is the Code Too Conservative?

The information in the following paragraph is technical in nature and may be subject to further investigation. It gives some indication that theCodemay not be as overly conservative as many feel it is.

While this situation of connecting supply overcurrent devices at opposite ends may be safe for restricted conductors, it may not be suitable for busbars in panelboards (load centers), even though this allowance is in the 2008 and 2011NEC. Panelboards are subject to busbar current limitations and are also subject to thermal limitations due to the heating associated with the thermal trip elements in the common thermal/magnetic molded-case circuit breakers. For example, a 100-amp, 120/240V panelboard is tested during the listing process with a 100-amp main breaker (line 1 and line 2) and two 100-amp load breakers (one per phase) mounted directly below the main breaker. The ambient temperature is raised to 45 degrees Celsius, the input and output currents are set at 100 amps, the temperature is allowed to stabilize, and the panel must pass this test with no deformation of any parts that would result in external damages. The internal thermal load is related to the heat produced by 100 amps passing through four circuit breaker trip elements. This would be a thermal load equivalent to 400 amps. If we add a double-pole backfed PV breaker, for example 20 amps, at the bottom of the panel, and if the loads on the panel were increased to 120 amps (per phase), no breakers would trip, no busbars would be overloaded, but the thermal load in the panel would be that associated with 480 amps, not the 400 amps for which the panel was designed and listed. Panel manufacturers have stated that these panels may not be able to pass UL listing tests with those excessive thermal loads. Plastic insulators could deform and arcs and sparks could result.

How likely is it that increased loads would occur at the same time as high daytime PV outputs? No one knows, but the possibility exists and some inspectors report warm/hot load centers (without PV input) that may be operating already close to the rating of the main breaker. An extra copier, fax machine or large screen TV might tip the balance.

Code Requirements Do Not Always Make Sense

Consider Figure 1: The conductor has a 60-amp breaker at the utility feed end and it has a 40-amp backfed breaker at the PV inverter end. If that conductor is not tapped for loads or other sources, then the highest current that it could ever see under any normal or fault condition is 60 amps, the rating of the highest connected supply overcurrent. However, 690.64(B) / 705.12(D) require the ampacity of the cable to be not less than 120% of the sum of the ratings of the supply overcurrent devices. As a calculation:

  • 60 + 40 must be less than or equal to 120% of the ampacity of the conductor.
  • 60+40 <= 1.2 x A 100 <=1.2A
    A >= 100/1.2 = 83 amps the required cable ampacity

A proposal was made for the2011NECthat would apply to end-fed conductors that have a restriction (marking) that they not be tapped for either loads or supplies. If this proposal were accepted — it was rejected — then the conductor would need an ampacity only as high as the highest rating of one of the connected supply overcurrent devices.

In previous Code cycles, labels and placards that say, "Add no loads” have been proposed. Those proposals have been rejected. Proposals for dedicated ac inverter combining panels with no spaces for loads have been proposed. They have been rejected. Covering empty breaker positions with metal guards have been proposed — rejected. Marking conduits, "PV output circuits, multiple source, do not tap”— rejected.

Exceptions were proposed to 690.64(B) / 705.12(D) to allow more flexible installations. These exceptions place restrictions or allowances on the general conditions of an unrestricted busbar or conductor. The restrictions keep the various installations safe.

For example, the 2008 NEC 690.64(B)(2) requirement says to add the ratings of all breakers supplying current to the panel. This would include the main plus all backfed PV breakers. Assume that it is desired to combine the outputs of two inverters in a dedicated PV ac combining panel with two 40-A breakers. An 80-A main breaker would normally be needed. The sum of all breakers would be 160 amps, necessitating a 200-panel to meet 690.64(B)(2) / 705.12(D)(2). However, if an exception (restriction) were added that prevented any loads from being added to the panel, then the maximum current that the busbar would ever see would be limited to the sum of the PV breakers or the main breaker, if larger. The panel could then be rated at 80-A or 100-A — still safe, and less costly.

The AHJ Has the Final Say

An AHJ may certainly look at a specific installation consisting of a specific set of supply breakers, loads, and locations of the same and evaluate the ampacity requirements of the conductors or busbar. If an alternate methods and materials (AMM) approval is issued to allow a deviation from the wording of theNEC, then the AMM approval might also include instructions to the installer to modify the installation in a way to minimize the possibility of future changes to the installation that might violate the exceptions (restrictions). For example, a "No Loads Allowed” placard might be required on an ac PV inverter combining panel when an AMM approval has allowed the rating of the panel as either the main breaker rating or the sum of the PV breakers, whichever is greater. Another example (proposed for the2011 NEC but not accepted) is to allow a conductor fed from supply breakers at each end, to have an ampacity of the greater breaker rating, not the sum of the breakers, when the conductor is marked, "Multiple Power Sources — Do Not Tap” every ten feet where the conductor is accessible and inside any connected distribution equipment.


Well-substantiated proposals will again be submitted for the 2014NECto allow some exceptions to the basic requirements in 705.12(D). Hopefully, CMP-4 will carefully address these proposals and see that PV installations can be safe, durable and cost effective without overly restricting the installations.


In summary, 690.64(B)(2) / 705.12(D)(2) is written as an unrestricted requirement for sizing conductors and busbars fed from multiple sources. The conductor or busbar is protected for any combination of loads and/or multiple sources and locations of loads or sources connected to the busbar or conductor. It would appear that the existingCodemight be overly restrictive.

For Additional Information

The US Department of Energy funding for providing inspectors and the PV industry with telephone and e-mail support from the author was terminated on March 1, 2011. Answers to your questions may be delayed or not answered at all depending on future funding. Consultation services are available on a contracted basis. E-mail: jwiles@nmsu.eduPhone: 575-646-6105

See the web site below for a schedule of presentations on PV and theCode.

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 (1.91) 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:

And yes, it may be updated to the 2008 and 2011 Codes sometime this year.

Read more by John Wiles

Tags:  Featured  July-August 2011  Perspectives on PV 

Share |
PermalinkComments (0)

Changes and Challenges

Posted By John Wiles, Sunday, May 01, 2011
Updated: Wednesday, January 16, 2013
For nearly a century from about 1897 to 1997, premises wiring systems in residences and commercial buildings have largely been collections of passive conductors, disconnects and overcurrent devices. Certainly there have been incremental improvements in these systems and they can be quite complex with the addition of transformers, motor controllers, GFCIs and AFCIs, but much of that complexity is due to the connected loads that are not covered in inspections under the requirements of the National Electrical Code (NEC).

In 1997, interactive power sources such as photovoltaic (PV) power systems started being installed in large numbers due to financial incentives in California and elsewhere. PV systems were just the start of a parade of technology changes that will affect large segments of the electrical power distribution and premises wiring systems, and the inspection requirements for those systems.

Photo 1. Electric car

Photo1. Electric car

The Changes

Electric Vehicles
We now have both plug-in hybrid electrical cars (fueled engines plus electric motors and batteries) and pure electric cars (electric motors and batteries) that require charging stations at not only the home base of such cars, but also in locations throughout the area that these vehicles will roam. Like cell phone coverage, the charging stations will be concentrated in metropolitan areas and then spread to less populated areas as the demand for extended coverage grows. Owners of these electric vehicles will certainly have charging stations in their homes and probably at their job sites. At the very least, there will be a new type of receptacle outlet to deal with and probably relatively high current branch circuits.

Plans are also being made to have parked, fully charged electric vehicles feed some of the energy stored in the on-board battery bank back into the utility grid at peak demand times. To control this exchange of energy from grid to car and back, and to ensure that the car is ready and charged when needed, will require communication between the car, the owner, and the utility. Such communication links may be wireless, over the Internet or through a hardwired connection along with the power connections. Like utility-interactive PV systems, these vehicle storage systems will require new code changes and additional inspections to ensure the public safety.

Large Energy Storage Systems

The utilities will embrace the dispatchable energy storage and generation systems. They will be able to tap energy that has been stored or that is available throughout the distribution network for use to offset peak demand loads. This operation will avoid having to increase the size of already taxed power plants and transmission lines. Backup generators at hospitals and other locations are already being used in this mode of operation. These emergency power systems are leased, operated and maintained by third parties who run them when not needed for emergencies and sell the power to the utilities during peak load periods.

Photo 2. Backup generator

Photo 2. Backup generator

Flow batteries are coming to the market. These batteries use stored liquid chemicals in a process that yields a very long-lived battery that can be rapidly charged and deeply discharged virtually an unlimited number of times. The batteries will be charged and energy will be stored during off peak demand periods and released back into the grid during peak demand times. Of course, the process will require utility-interactive systems to interface with the utility grid and communications systems to control the process. These systems and fuel cells (NECArticle 692) operating from natural gas will probably first appear in commercial buildings that have the necessary space. These systems will either be leased or owned, but in most cases, these new technology systems will require permitting and inspections of the added mechanical systems, the utility-interactive electrical connections and the communication circuits.

The Smart Grid

Energy demands throughout the country, and the world, are increasing steadily and will necessitate some combination of increasing in the supply from new generation plants (coal, gas, oil or nuclear and renewable), reducing the demand through conservation, or restructuring of the existing distribution and consumption system. The infrastructure of utility generation and distribution systems is fairly robust, but very old, and somewhat inflexible in dealing with increased use of distributed energy sources and the issues associated with moving power from the sources to the consumers in other areas. The Smart Grid programs are designed to modernize the entire system from the generation plant to the end use-load.

Photo 3. Benefiting from renewable energy resources

Photo 3. Benefiting from renewable energy resources

Although many see the termSmart Gridand think that it will not impact the premises wiring, theNEC, or the inspection process, that would be a misconception. At the present time utilities are installing smart meters as rapidly as they can find funds to do so. These smart meters are computer (microprocessor) based and not only allow remote reading and power quality recording (real power, reactive power, power factor and more), but may also serve as the interface between the smart grid and the smart house. Some of the smart meters even have the ability to allow the power to be remotely disconnected when bills are not paid.

The smart house will soon become a reality. Appliance manufacturers are already making dishwashers, clothes washers and other appliances that communicate through either hardwired or wireless communication systems to the smart meter and then to the utility. When financially beneficial to the consumer, or possibly when legislated, these smart appliances will be remotely controlled (by the utility) so that they may be operated only during times of low demand on the utility system. Those appliances may have unique plugs and receptacles and possibly communication connections. All of those new load connections must be inspected of course.

Photo 4. Small wind system combined with solar panels

Photo 4. Small wind system combined with solar panels

What will be theCodeimpact of a house that has load circuits and loads that may be remotely controlled or managed? How will service, feeder, and branch circuit sizes be determined? Copper conductor prices may rise so high that we are forced to control power flow so that smaller conductors can be used. Eventually, the use of electricity on the premises may be scheduled so that the maximum current ever drawn may be significantly less than that requiring a 100- or 200-amp service today. Smaller conductors and circuit sizes may reduce the ever-increasing costs of electrical installations, butCoderevisions would be needed. With the demise of the incandescent light bulb do we really need three volt-amps per square foot for general-purpose circuits? Oh yes, there will be those 100+ inch flat panel displays on all four walls to deal with.

Is DC coming back?
Then we have the new trend of going back to direct current end-use appliances. Most electronic appliances such as cell phone chargers, radios, TVs, DVRs, DVD players, cableboxes, satellite receivers, track lighting and the like, while being plugged into a 120-volt ac receptacle outlet actually run on low-voltage direct current (dc). Fluorescent and LED lighting bulbs and fixtures also operate on direct current. Significant losses are incurred in transforming the 120-volt ac line voltage into low voltage dc.

At the present time dc lighting fixtures are being installed in commercial buildings and are being powered during the day directly from photovoltaic (PV) power systems with no conversion to ac until the electronic ballasts are reached. Solar lighting power is supplemented with utility power when necessary.

With the demise of the incandescent light bulb over the next few years, the return of low voltage dc power distribution systems for lighting and electronics is almost a certainty. Shades of the 1970s and 1980s! Maybe those off-grid long-haired solar hippies who insisted on staying with the 12-volt dc PV systems and electrical systems in their homes where far ahead of their time! Of course, appliances needing significant power for heat or mechanical motion like ranges, clothes washers, toasters, water heaters and the like will usually need higher voltages to keep the current and hence the conductor sizes to reasonable sizes. But then we do have heat pump water heaters, induction ranges, and ultrasonic washers that operate more efficiently than conventional appliances.

Renewable Energy Systems
Large wind power systems have been installed for many years and many of those systems are not owned and operated by utilities on utility property. They therefore come under the requirements of the NEC and should be inspected for safety even though the NEC does not have a large wind system article. Now that Article 694 has been added to the Code for small wind systems, and UL has standards for large and small wind turbines, can a large wind turbine article in the Code be far behind? Photovoltaic power systems for residential and commercial use have been around since the mid 1970s with substantial growth starting in the late 1990s.

While ever-increasing numbers of residential and small PV systems are being installed throughout the country, real power production will come from the numerous megawatt commercial systems being installed and planned. Systems as large as 300 megawatts are being panned and installed and some of these will be solar thermal systems along with the PV systems. In many cases, these large systems are said to be "Behind the Fence” and not subject to the requirements of theNECand inspections, but in reality, they are mainly owned and operated by private companies under power purchase agreements (PPA) and should be fully NEC compliant.

True AC PV modules with microinverters bonded to the back of the PV module with no dc wiring are appearing on the market in catalogs and in big box stores at impulse-buying prices. Will these products be listed? Will these be permitted? Will they be installed by qualified persons (690.4(E) NEC-2011)? Will they be installed on dedicated circuits that will ensure public safety? Or will they be plugged into the nearest GFCI outlet and abuse theCodeand safety in many ways?

Photo 5. Digital cameras

Photo 5. Digital cameras

DC-to-DC converters attached to or connected to PV modules are appearing on the market along with matching inverters in some cases. TheNECdoes not specifically give guidance on how to deal with them and future editions of the Code may show a similar trend.

All of these changing and emerging technologies will create challenges for the inspectors and plan reviewers and also an opportunity to excel.

The Challenges

Electrical inspectors, plan reviewers and combination inspectors are being challenged today and for the foreseeable future with all of these new and evolving energy production and storage sources that will be in use throughout the country. Many of them will appear connected to premises wiring and they will come under the requirements of the NEC.Many of those multi-megawatt PV, wind, and solar electric farms will fall under the Code.

The Code
Each edition of the NEC is developed over a three-year period through the code-making process that is well established. Competent, experienced volunteers make up the code-making panels (CMP) and with the NFPA/NEC Technical Correlating Committee (volunteers and professional staff) review and evaluate thousands of proposals and comments on proposals submitted from numerous sources. There are only two weeklong (or less) meetings over that three-year cycle where the CMPs develop and write the Code.

With electrical and electronic technologies changing at a rapid pace, it is unreasonable to expect the NEC to keep abreast of all of the newest technologies that appear in the marketplace, even though the volunteers and staff make a valiant effort to do so. Many of those technologies are changing in form and function on a monthly basis and are not addressed by theCode, even though they are listed and certified under appropriate standards and are in the marketplace.

The Standards
Underwriters Laboratories and other organizations are developing safety standards as rapidly as possible. However, the development and revision process for standards and the harmonization of U. S. standards with those from Europe can and does take long periods of time. Those periods can even exceed the three-year cycle of the NEC.

Although the NEC and the UL Standards are intended to be used together to achieve an essentially hazard-free electrical installation, there are sometimes gaps between the two due to the lengthy revision processes and the emergence of new technologies. For example, the2011 NEC, adopted by some jurisdictions on January 1, 2011, has a requirement for a DC PV Arc-Fault Detection and Interruption System in Section 690.11, but there was no current UL Standard as of January 2011 that covers the safety evaluation of such a device. And the DC PV AFCI devices are already in the market.

Continuing Education and Information
The challenge for every electrical inspector and plan reviewer is to keep abreast of these new developments as they start to appear in residential, commercial, and industrial electrical systems. The inspectors and the plan reviewers need to know as much, or more, about these new devices and systems as the people installing them. That has been true in the past and it needs to be the standard of performance in the future if the inspector community is to ensure the safety of the public.

Where the Code and the standards cannot keep up with these new systems and devices, the inspector and plan reviewer must devote time to educate themselves on the systems that they are and will be inspecting. Strong continuing education programs for the inspectors and plan reviewers must be a part of the planning in every jurisdiction. Time and funding must be budgeted for classes, for webinars, for technical documents, and for the equipment needed to efficiently and proficiently review and inspect these ever-changing electrical power systems.

The inspectors and plan reviewers should have electronic copies of all codes, handbooks for those codes, and technical data (including manuals and specification sheets) for all types of systems being inspected and equipment that may be installed on those systems. Laptop computers (with screens that can be read outdoors) with this information (updated as necessary) should accompany each inspector as the field inspections are conducted. Communication between the inspectors and the plan reviewers on a real time basis via cell phone and wireless computer link will be required. Digital cameras, downloads and transmission of on-site pictures will become necessary.

Inspectors and plan reviewers are professionals today and will remain professionals in the eyes of the public as they rise to the challenges presented by the changes in the electrical power system today and tomorrow.

This article is intended to help inspectors and plan reviewers keep abreast of the rapidly changingNECrequirements for the installation of photovoltaic power systems.


There are many resources available to the inspector and plan reviewer. Most equipment manufacturers have electronic downloadable PDF files of all manuals that will be useful. Here are a few magazines (available in print and on the web) that will enable inspectors and plan reviewers to keep abreast of the changing technologies.


For Additional Information

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

See the web site below for a schedule of presentations on PV and the Code.

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 (1.91) 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:

This work was supported by the United States Department of Energy under Contract DE-FC 36-05-G015149

Read more by John Wiles

Tags:  Featured  May-June 2011  Perspectives on PV 

Share |
PermalinkComments (0)

What Hath the 2011 NEC Wrought for PV?

Posted By John Wiles, Tuesday, March 01, 2011
Updated: Wednesday, January 16, 2013

The 2011 National Electrical Code (NEC) has been published by the National Fire Protection Association (NFPA) and is now available from numerous sources. It was adopted by some jurisdictions automatically on 1 January 2011, and will be adopted throughout the country over the next three years or even longer in some areas that are slow to change.

Anyone working with PV systems and equipment in either manufacturing, design, installation, or inspection arenas should get a copy of the 2011 NEC and the 2011NEC Handbook. TheNECindicates the code changes (which will not be repeated verbatim here) by highlighting and the Handbook provides additional explanations.

I hope that the following information is reviewed with the 2011NECin hand or at least it whets the appetite for getting the Code ASAP. Inspectors will usually start reading the new Code as soon as it becomes available for clarifications of the existing code, even though their jurisdiction may not adopt the newest Code for several years. In many cases where safety enhancements are involved, AHJs will permit or even enforce the requirements of the new Code before it is officially adopted by the jurisdiction.


Code-making panel (CMP) 4 processed Articles 690, Solar Photovoltaic (PV) Systems, and 705, Interconnected Electrical Power Production Sources, for the 2011 NEC. Those articles had previously been handled in CMP-13 for many years. CMP-4 did not have the long-term exposure to PV systems and the unique PV characteristics of current-limited dc generators and utility-interactive ac sources. Many of the carefully thought out and substantiated proposals were rejected for obscure reasons.

In general, we have many areas of Article 690 that were clarified, some that were not, and some added requirements, plus a move of several sections from Article 690 to Article 705. Minor clarifications and grammatical corrections will not be addressed in the following.

690.2 Definitions. Definitions ofsubarrayandmonopole subarraywere added so that they can be used in requirements dealing with the return of bipolar arrays. They have not been in evidence since the mid-1990s and at that time the safety issues resulted inCodechanges.

690.4(A) Installation. Clarification

690.4(B) Installation.Extensive marking requirements were added for all circuits in a PV system. Safe maintenance was the justification. When you open a junction box or combiner, circuit identification should be easy.

690.4(E) Installation. Qualified persons shall install all PV equipment and systems. See the definition ofqualified personin Article 100. Specific skills and training including safety training are mentioned in the definition.

690.4(F) Installation. Circuit routing requirements were added to reduce the likelihood that fire fighters will come into contact with energized circuits. PV circuits inside and outside the building are affected.

690.4(G) Installation. More stringent requirements for bipolar arrays were added to avoid exceeding the voltage rating on equipment. Inspectors will have to look closely at these new systems since the UL Standard 1741 does not specifically address these types of inverters.

690.4(H) Installation. Directory requirements were established for multiple inverters on a single building.

690.7(A) Maximum Photovoltaic System Voltage. An Informational Note (previously a Fine Print Note) gives a source of temperature data that could be used to calculate cold weather open-circuit voltage.

690.7(E) Bipolar Source and Output Circuits. A clarification of ground-fault actions on a bipolar array was added.

690.8(B) Ampacity and Overcurrent Device Ratings. An extensive revision was made to clarify and align PV overcurrent device rating and conductor size calculations with basic requirements found elsewhere in the Code. See January-February 2011 IAEI News, "Perspectives on PV” for details. DC PV conductor ampacity calculations do not always involve 1.56 Isc.

690.9(A) Circuits and Equipment. Exception: Clarification.

690.9(B) Power Transformers. Clarification

690.9(E) Series Overcurrent Protection. Clarification

690.10(E) Backfed Circuit Breakers. Clamping requirements for backfed circuit breakers in stand-alone system were modified to include requirements for multi-mode inverters in battery backed up utility-interactive PV systems.

690.11 Arc-Fault Circuit Protection (Direct Current). A new requirement was added for a dc PV arc-fault circuit interrupter. It must detect series arcs in the dc PV circuits, interrupt them, disable equipment, and annunciate. Equipment is in the market addressing this equipment, at least for off grid systems, and other equipment is coming.

690.13 All Conductors. Clarifications.

690.13 Exception No. 2. A disconnecting means will be permitted in the grounded conductor for maintenance actions and then when accessible only by qualified people.

690.14 PV Disconnecting Means. Unfortunately, no changes were approved.

690.16(A) Disconnecting Means. Clarification

690.16(B) Fuse Servicing. Disconnecting means from all sources of energy shall be located at the fuse location or a directory shall be provided to show disconnect location(s). This requirement is aimed at large inverters which have dc fuses bolted to an input bus bar with no way to de-energize those fuses without opening every single one of the possibly hundreds of fuse holders in the distant combiner boxes.

690.31(B) Informational Note. PV wire has a nonstandard outer diameter and conduit fill tables cannot be used.

690.31(E) Direct-Current Photovoltaic Source and Output Circuits. Corrects longstanding typo and indicates that only dc circuits must be in a metal raceway, not ac inverter output circuits. Allows type MC metal-clad cable to be used for DC circuits inside the structure. Four new paragraphs of requirements have been added on routing, protection, and marking of PV circuits inside the building. Addresses conductor protection, maintenance and fire fighter concerns.

Conductors under the roof shall be 10” below the roof decking. Small metallic raceways and cable assemblies shall be protected from physical abuse in accessible areas. All access points and exposed conduits will be marked as containing PV power sources.

690.43 Equipment Grounding. Clarifications in (A) through (F).

690.43(C) Mounting structures for PV modules shall be identified as equipment grounding conductors or shall have all parts bonded together and to the equipment-grounding system.

690.43(D) Mounting devices used for grounding modules shall also be identified as grounding devices.

690.47 Grounding Electrode System. Substantially revised and clarified. The requirements 690.47(C) in the 2005NECwere merged with the requirements of 690.47(C) in the 2008NEC. See September-October 2009IAEINews"Perspectives on PV” for details.

690.47(D) was deleted.

690.62 Ampacity of Neutral Conductor. Deleted and moved with califications to 705.95.

690.63 Unbalanced Interconnections. Referred to 705.100 without changes.

690.64 Point of Connection. Referred to 705.12 with only two changes; 690.64(A) becomes 705.12(A), and 690.64(B) becomes 705.12(D).

Both sections have needed substantial revisions since 1984.

690.72(C) Buck/Boost Direct-Current Converters. A new section has been added to establish how ampacity and voltage requirements are to be calculated for these devices. Although in Part VIII, Storage Batteries, these requirements may also be used for module circuit dc-to-dc converters.

705.6 System Installation. Qualified persons must do installations of parallel power sources.Qualified personsis defined in Article 100.

705.12(A) Supply Side. The sum of the ratings of power production sources shall not exceed the rating of the service.

705.12(D)(2)Exception. Describes a method of sizing ac output circuits for battery-sourced, multi-mode inverters operating in utility-interactive systems. The 120% equation, where allowed, may use 125% of the rated inverter utility-interactive current instead of the rating of the backfed circuit breaker.

705.60, 65, 70, 80, 82, 95, and 100 contain requirements that duplicate information in various sections of 690.

The Future

We are already working on proposals for the 2014NEC, which are due to NFPA by November 4, 2011. Sections that are being examined for revisions include 250.32, Figure 690.1(A), 690.2, 690.4(D), 690.6, 690.x (microinverters), 690.y (dc-to-dc converters), 690.7(E), 690.14, 705.12 and others. If you see a section of the Code in 690 that is not abundantly clear, send me e-mail with your proposed changes and substantiations.

Please visit the Solar America Board of Codes and Standards web site for updates on proposals being developed by the PV Industry Forum.

For Additional Information

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

See the web site below for a schedule of presentations on PV and the Code.

A color copy of the latest version (1.91) 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:

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 15 years are also available on this web site:

This work was supported by the United States Department of Energy under Contract DE-FC 36-05-G015149

 Read more by John Wiles

Tags:  Featured  March-April 2011  Perspectives on PV 

Share |
PermalinkComments (0)

Conductor Sizing and Overcurrent Device Ratings

Posted By John Wiles, Saturday, January 01, 2011
Updated: Wednesday, January 16, 2013
Conductor sizes and overcurrent device ratings are critical to the safe, long-term operation of any electrical system, but are of particular importance in PV systems where the outdoor environment can be extreme and the PV modules will be sourcing current for 40 years or more. Historically, most residential and light commercial electrical wiring and inspections of these systems have involved indoor wiring at room temperatures [30°C (86°F) or less]. The ampacity tables in the NEC Section 310.15 and Table 310.16 were developed with those conditions in mind. The commonly used molded-case circuit breaker has a terminal temperature limit of 75°C and is rated for use with conductors with 75°C insulation. They have a rated maximum operating temperature of 40°C.

Photo 1

With these conditions and equipment characteristics in mind, the typical electrician doing indoor wiring has generally used the 75°C insulated conductor ampacity tables in Table 310.16 and not bothered too much with temperature corrections (310.15) and terminal temperature limits [110.14(C)] since they were not necessary or were included in the tables being used.

However, direct current (dc) PV conductors normally operate in an environment that is too hot for conductors with 75°C insulation. Conductors with 90°C insulation must be used and appropriate temperature and conduit fill corrections must be applied along with verifying that connected equipment terminal temperatures (60° or 75°C) are not exceeded. To do otherwise and use the short-cuts of the old days will result in conductors that may be not suited for the application and that may be larger than code requirements resulting in unnecessary costs.

Photo 2. Fused PV combiner with large and small cables

Throughout the Code, circuits are sized based on 125% of the continuous load plus the noncontinuous load. See 210.19(A)(1) and 215.2(A)(1). This requirement establishes a situation where conductors and overcurrent devices are not subjected to continuous loads (currents) more than 80% of rating. (Note: 1/1.25 = 0.80 and we can either divide or multiply depending on how the calculations are being accomplished).

Electricians and PV installers typically use the 125% factor and then apply the conditions of use factors (temperature and conduit fill)sequentially. The NEC,in a careful reading of the two referencedsections, does not require that both factors be applied at the same time. See the 125% requirement below.

In the Code, we have at least two or three requirements that must be met in sizing conductors.

First is the definition ofampacityfound in Article 100. Ampacity is "The current in amperes that a conductor can carry continuously under the conditions of use without exceeding its temperature ratings.”

Next is the 125% requirement in 210.19(A)(1) and 215.2(A)(1): "The minimum feeder circuit conductor size,before the application of any adjustment or correction factor,shall have an allowable ampacity not less than the noncontinuous loads plus 125 percent of the continuous loads” (emphasis added). This requirement ensures that conductors and overcurrent devices are not operated continuously at over 80% of rating.

Photo 3. Fuse AC and DC ratings will be different.

Then, Section 110.14(C) requires that the temperature of the conductor in actual operation not exceed the temperature rating of terminals on the connected equipment.

An added requirement for any listed equipment such as overcurrent devices is that they not be used in a manner that deviates from the listing or labeling on the product [110.3(B)]. Most PV source-circuit combiners operating outdoors in the sunlight will have internal temperatures that exceed the 40°C rated operating temperatures of commonly used fuses and circuit breakers.

The following method of determining ampacity meets the three code requirements above and finds the smallest conductor that can be used to meet these requirements. It also determines the rating of the overcurrent device where required.

Step 1. Determine the continuous current in the circuit.

PV dc circuits and PV ac circuits are not "load” circuits so we will use the termcurrentinstead ofload. For code calculations, all dc and ac PV currents are considered continuous and are based on worst-case outputs or are based on safety factors applied to rated outputs.

A. PV DC Circuits.In the dc PV source and dc PV output circuits, the continuous currents are defined as 1.25 times the rated short-circuit current Isc(marked on the back of the module). If a module had an Iscof 7.5 amps, the continuous current would be 1.25 x 7.5 = 9.4 amps [690.8(A)(1)].

If three strings of modules (module Isc= 8.1 amps) were connected in parallel through a fused source circuit combiner, the PV output circuit of the combiner would have an Iscof 3 x 8.1 = 24.3 amps. The continuous current is this circuit would be 1.25 x 24.3 = 30.4 amps [690.8(A)(2)].

B. AC Inverter Output Circuits.In the ac output circuits of a utility-interactive inverter or the ac output circuit of a stand-alone inverter,

Photo 4. Copper busbars are used instead of cables.

Photo 4. Copper busbars are used instead of cables.

the continuous current is taken at the full power rated output of the inverter. Itis notmeasured at the actual operating current (which may be a small fraction of the rated current due to a small PV array connected to a large inverter) of the inverter. Usually the rated current is at the nominal output voltage (120, 208, 240, 277, or 480 volts). The rated output current is usually specified in the manual, but may be calculated by dividing the rated power by the nominal voltage. For stand-alone inverters, which can provide some degree of surge current, it is the rated power that can be delivered continuously for three hours or more [690.8(A)(3)].

In some cases, the inverter specifications will give a rated current that is higher than the rated power divided by the nominal voltage. In that situation, the higher current should be used.

For a utility-interactive inverter operating at a nominal voltage of 240 volts and a rated power of 2500 watts, the continuous current would be:

2500 W/240 V=10.4 A.

A stand-alone inverter with a model number of 3500XPLUS operates at 120 volts and can surge to 3500 watts for 60 minutes. However, it can only deliver 3000 watts continuously for three hours or more. The rated ac output current would be:

3000 W/120 V = 25 A.

C. Battery Currents.The currents between a battery and an inverter in either a stand-alone system or a battery-backed up utility-interactive

Photo 5. Improperly specified and sized cable

Photo 5. Improperly specified and sized cable

system must be based on the rated output power of the inverter (continuous for three hours or more) at the lowest input battery voltage that can provide that output power [690.8(A)(4)]. Normally the output current from the battery in the inverting mode is greater than the current to the battery in the charging mode. This current is usually marked on the inverter or found in the specifications.

The battery discharge current can be calculated by taking the rated output power, dividing it by the lowest battery voltage that can sustain that power, and also by dividing by the inverter dc-to-ac conversion efficiency at that battery voltage and power level. For example:

A 4000-watt inverter can operate at that power with a 44-volt battery input voltage and has a dc-to-ac conversion efficiency (inverting mode) of 85 percent. The dc continuous current will be:

4000 W/44 V/0.85 = 107 A.

On single-phase inverters, the dc input current is rarely smooth and may have 120 Hz ripple current that is larger in root mean square (RMS) value than the calculated continuous current. The inverter technical specifications should list the greatest continuous current.

Step 2. Calculate the rating of the overcurrent device, where required.

Since PV modules are current limited, overcurrent devices are frequently not needed for one or two strings of PV modules connected in parallel. In systems with three or more strings of modules connected in parallel, overcurrent devices are usually required in each string to protect not only the conductors, but also the module internal connections.

A. Rating Determined from Continuous Currents.The overcurrent device rating is determined by taking the continuous current for any of the circuits listed in Step 1 and increasing the continuous current by 125% (or by multiplying by 1.25). Non-standard overcurrent device values should be rounded up to the next standard rating in most cases.

In a very few rare cases, an overcurrent deviceinstalled in an enclosureor an assembly may be tested, certified and listedas an assemblyfor operation at 100% of rating. In these cases, the overcurrent device rating is the same as the continuous current and no 125% factor is used. The author knows of no overcurrent devices installed in an enclosure for PV systems that have such a rating.

B. Operating Temperature Affects Rating.Overcurrent devices are listed for a maximum operating temperature of 40°C (104°F). PV combiner boxes operating in outdoor environments may experience ambient temperatures as high as 50°C. Exposed to sunlight, the internal temperatures may reach or exceed 55–60°C. Any time, the operating temperature of the overcurrent device exceeds 40°C, it may be subject to nuisance trips at current values lower than its rating. In this situation, the manufacturer must be consulted to determine an appropriate derating. At high operating temperatures an overcurrent device with a higher rating will activate at the desired current. In PV source circuits, the new rating of the revised overcurrent device (under cold weather conditions) must not exceed the ampacity of the conductors or the maximum series fuse value marked on the back of the module.

Step 3. Select a conductor size.

The conductor selected for any circuit must meet both the ampacity requirement and the 125% requirement. The correctly sized cable is the larger of A or B below.

A. Ampacity Requirement.The conductor, after corrections for conditions of use must have an ampacity equal to or greater than the continuous current found in Step 1. See Article 100, Definition of ampacity.

B. 125% Requirement.The cable must have an ampacity of 125% of the continuous current established in Step 1. See 215.2(A)(1).

Example 1.Three (3) conductors are in a conduit in a boiler room where the temperature is 40°C. The continuous current in all four conductors is 50 amps. A copper, 90°Cinsulated cable is specified.

Photo 6. PV containers may operate above 40°C.

Photo 6. PV containers may operate above 40°C.

Temperature correction factor = 0.91, Conduit fill correction factor = 1.0

Step A, Ampacity Rule: Required ampacity at 30°C is 50/0.91/1.0 = 54.9 amps and this would require an 8 AWG cable.

Step B, 125% Rule: 1.25 x 50 = 62.5 amps and this would indicate a 6 AWG cable.

The 6 AWG cable is the larger of the two and is required.

Example 2.Now there are six (6) conductors in the conduit and the temperature has increased to 50°C. The continuous current is still 50 amps.

Temperature correction factor = 0.82, Conduit fill factor = 0.8

Step A, Ampacity Rule: 50/0.8/0.82 = 76.2 amps and a 4 AWG cable is needed

Step B, 125% Rule: 1.25 x 50 = 62.5 amps calling for a 6 AWG cable.

The 4 AWG cable is the larger of the two and must be used.

Step 4. Terminal temperature limits

A. The terminal temperature limits marked on the equipment must be used. If no temperatures are marked, then a 60°C limit is used for circuits rated at 100 amps or less or cables 14–1 AWG. For circuits rated greater than 100 amps and for conductors greater than 1 AWG, a 75°C terminal temperature limit will be used. See 110.14(C).

The following method is a terminaltemperature estimationmethod and is not an ampacity calculation method. It is used after the conductor size has been selected based on the ampacity calculation.

Take the conductor size in Step 3 above. Find the lowest terminal temperature limit for this conductor at any termination. Use that terminal temperature limit (either 60°C or 75°C) to enter the ampacity Table 310.16. For the conductor size selected, read out the current in the correct column, either the 60°C column or the 75°C column. There are no temperature adjustments or conduit fill adjustments to this current.

The current from the table must be equal to or greater than 125% of the continuous current. And, if the conductor meets this requirement, then the terminal temperatures are going to be less than the 60°C or 75°C limit for that conductor and that continuous current. The 125% factor is a fudge that accounts for many items not calculated in this simplified temperatureestimationprocess.

Example 3.Take the 8 AWG conductor and 50 amps of continuous current used in Example 1 above. This conductor is connected to a terminal with a 60°C marking.

From Table 310.16, an 8 AWG conductor in the 60°C column can carry a current of 40 amps.

We take 125% of the continuous currents of 50 amps.

1.25 x 50 = 62.5 amps.

This is larger than the 40 amps from the table, and this terminal will be heated above 60°C.
If we increase the conductor size to 6 AWG, the table gives us 55 amps, still less than 62.5 and too hot.

Increasing the conductor size to 4 AWG will give 70 amps from the table; and since this is greater than the 62.5 amps, we will be assured that the terminal will stay below its 60°Ctemperature limit.

Example 4.Use the 4 AWG conductor selected in Example 2 connected to a terminal with a 75°Ctemperature limit. The continuous current is 50 amps. Taking 125% of that continuous current yields:

1.25 x 50 = 62.5 A.

A 4 AWG conductor in the 75°C column of Table 310.16 shows a current of 85 amps. Since this is greater than the 62.5 amps, the conductor will operate cooler than the 75°C terminal temperature limit. No increase in conductor size is necessary.

Step 5. Verify that the overcurrent device protects the conductor selected under the conditions of use.

Where an overcurrent device is required, it must protect the conductor under operating conditions (conditions of use). Conductors may be protected using the round up allowance found in 240.2(B).

Example 5.A circuit has a continuous current of 70 amps. After conditions of use (4 conductors in the conduit, 48°C) are applied, a 3 AWG, 90°C conductor is selected to meet all ampacity and 75°C terminal temperature requirements.

The ampacity after conditions of use have been applied is:

110 x 0.8 x 0.82 =72.2 A.

The required minimum overcurrent device for this level of continuous current is

70 x 1.25 = 87.5 A.

A 90-amp overcurrent device would typically be used. A few people have suggested using an 80-amp overcurrent device, but that would result in running it at more than 80% of rating and in dc PV circuits could result in nuisance trips during short periods of cloud enhanced irradiance.

However, the largest overcurrent device that could be used to protect the 3 AWG conductor with an ampacity of 72.2 amps is an 80-amp overcurrent device and a 90-amp overcurrent device is the smallest allowed in this circuit.

The conductor size would have to be increased to 2 AWG for full compliance with NEC requirements.
The ampacity of a 2 AWG, 90°C conductor under the conditions of use is:

130 x 0.8 x 0.82 = 107 A.

The required 90-amp overcurrent device can protect the 2 AWG conductor.


PV installers, plan reviewers, and inspectors need to know how to do conductor sizing and overcurrent device ratings properly to get safe, reliable, and cost effective PV systems. This procedure for sizing conductors and overcurrent devices meets NEC requirements. In general, it can be used for any type of electrical circuit except possibly HVAC and other motor protection circuits. A part of this procedure is in Section 690.8(B) of the 2011NEC.

For Additional Information

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

See the web site below for a schedule of presentations on PV and the Code.

A color copy of the latest version (1.91) 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:

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 15 years are also available on this web site:

Read more by John Wiles

Tags:  Featured  January-February 2011  Perspectives on PV 

Share |
PermalinkComments (0)

Utility Interconnections and Code Requirements

Posted By John Wiles, Monday, November 01, 2010
Updated: Friday, January 18, 2013

Inspectors and installers continue to puzzle over the requirements in Section 690.64 of the National Electrical Code (NEC) that apply to the connection of utility-interactive inverters to the premises wiring and finally to the utility. This article, using the simplified block diagram (figure 1), will attempt to clarify some of those requirements. Please refer to previous Perspectives on PV articles over the last two years for more detailed information.

One Diagram Is Worth a Thousand Words

Many people do better with diagrams than they do with words, so the diagram shown should be just up their alley. This diagram works with many types of utility interactive PV systems. These systems all start with a meter connected to the utility as shown on the left. After that, we may be dealing with an existing service disconnect and the connected existing load center or with a PV supply side connection, which is just a second service entrance on the existing premises wiring system. In either case, the NEC requirements of Article 230 apply as noted at the bottom. In most jurisdictions, the local utility will require a PV disconnect on the ac output of the PV system and many areas will use a Renewable Energy Credit (REC) meter to measure the PV system output. As shown, one or more single inverters may be connected or even one or more "strings” of microinverters or AC PV modules may be connected to the added combining panel (blue blocks). Or, a single inverter could be connected to an existing load center (red blocks). In some cases multiple inverters might be connected through an ac combining panel and then backfeed an existing load center. Let’s start our examination of the requirements at the inverter end of the circuit.

Inverter Output Circuit

All utility-interactive inverters have a rated output current that cannot be exceeded. There are no surge currents in these output circuits andNEC690.8 requires that the circuit and the overcurrent protective device (OCPD) be rated at 125% of that rated output current. When the calculated OCPD value is a nonstandard value, the next standard higher value should be used, but not to exceed the maximum overcurrent value given in the technical specifications for the inverter. Conductor size should be selected so that it is protected by the OCPD rating.

Figure 1. Utility interconnections and NEC requirements

Figure 1. Utility interconnections and NEC requirements

The asterisk (*) by the 690.8 in the diagram indicates thatifthere is an overcurrent device mounted at the inverter, then the requirements of 690.64(B),and not 690.8, will apply. Some installers and manufacturers use a circuit breaker or fused disconnect at the inverter to meet the requirements of 690.15 to have a maintenance disconnect at the inverter. The inclusion of an overcurrent device at this location generally forces the output conductors from the inverter to be larger [as required by 690.64(B)] than would otherwise be required by 690.8.

After the First Inverter Overcurrent Device

Anyconductor or busbarthat can have power flowing from more than one source of supply (under normal or fault conditions) such as the utility and a PV inverter, and where the conductor is protected by an overcurrent device on each supply source must meet 690.64(B) requirements. This is the long-standing 120% allowance [when 690.64(B)(7) conditions can be met]. Section 690.64(B) is going to apply to allconductors and busbarsfrom the first overcurrent device connected to the inverter output all the way to the service disconnect.

These busbars and conductors would include the busbars of any backfed main panelboards connected to one or two inverters or sets of microinverters, and any busbars in PV ac inverter combiner panels. The conductors or feeders between the panelboards or load centers and the main service disconnects are also subjected to the requirements of 690.64(B)(2) as noted on the diagram.

In general the ratings of all of the breakerssupplyinga busbar or conductor areaddedtogether and the sum is divided by 1.2 (for the 120% allowance). If the location requirements of 690.64(B)(7) cannot be met (PV breaker located at the opposite end of busbar or conductor from the

Photo 2. PV on south-facing roof at a good slope

utility breaker), then the sum may be divided by only 1, and the busbar rating or cable ampacity goes even higher. For example:

Two inverters each require a 50-amp backfed breaker in a main lug PV ac inverter combining load center to meet 690.8 requirements. A supply-side connection is going to be made with a 100-amp fused disconnect. The rating of the combining load center and the ampacity of the conductor to the 100-amp fused disconnect must follow the 690.64(B)(2) requirements.

(50 + 50 +100)/1.2 = 200/1.2 = 166.7 amps

The numbers indicate that a 200-amp PV ac inverter load center/panelboard would be needed and a 2/0 AWG conductor should be used between that panel and the 100-amp fused disconnect.

Now suppose that the two inverters are being backfed into an existing panelboard (switchgear) and it is not possible to position the two backfed PV breakers at the opposite end of the switchgear busbar from the main breaker. The requirements of 690.64(B)(7) are not met and the 120% allowance cannot be used. The equation becomes:

(50 + 50 + main breaker) must be less than or equal to the busbar rating.

If the main breaker were rated at 200 amps, then the busbar would have to be rated at 300 amps.

As the diagram shows, 690.64(B) applies to any panel or load center that has connections to the utility and to the PV inverter. It can be an existing load center or an added PV ac inverter combining panel.

The Main Disconnect and on to the Meter

Any circuit between the meter and the service disconnect would be considered a service-entrance circuit and be governed by the requirements of Article 230. This would be true if the circuit was an existing service- entrance conductor or a new 690.64(A) supply-side connection. The conductor size, type, and routing as well as the size and location of the service disconnect would have to meet 230 requirements. However, after passing through the overcurrent device on either an existing service disconnect or through the overcurrent device on an added PV supply-side connection, the requirements of 690.64(B) apply all the way to the first overcurrent device connected to the inverter output.


A diagram can simplify understanding of the requirements ofNEC690.64. While the PV industry had hopes of getting additional clarity into this section of the Code, those hopes were not realized for the 2011NEC, and we must continue to work with the existingCodelanguage.

For Additional Information

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

See the web site below for a schedule of presentations on PV and the Code.

A color copy of the latest version (1.91) 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:

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:

Read more by John Wiles

Tags:  Featured  November-December 2010  Perspectives on PV 

Share |
PermalinkComments (0)
Page 2 of 6
1  |  2  |  3  |  4  |  5  |  6