| ||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||
On Tue, 01 Apr 1997 Redwood Kardon wrote:
One of these days there's going to appear on my daily activity list a call to inspect a photovoltaic installation. PV's are not part of my 20 years in the electrical business. To help prepare myself for this inevitable day I'm putting out a call to this news group to help me prepare an inspection punch list. What follows is a list I've started based directly on my readings from the NEC. It's perfunctory, and open to change. Please lend a hand and let me know what you know about this subject.
- Equipment shall be rated for current & voltage imposed on it [110-9]
- PV output conductors must be isolated from other systems(AC) [690-3b]
- Roof modules shall have ground fault protection [690-690-5]
- Two wire systems >50 volts (open circuit) requires a GEC [690-7]
- Three wire systems (any voltage) requires GEC [690-7]
- GEC connection should be as close to modules as practical [690-42,250-22]
- Disconnect means req'd for inverters, batteries, charge controllers [690-15]
- Array must have a way to be disabled [690-18] [n/a]
Let me introduce myself: I am a research engineer working at PVUSA in Davis, CA. My involvement with PV began sixteen years ago performing the acceptance testing of the Lovington Shopping Center in New Mexico, one of the PRDA systems. Since then I have worked on both small and large, UI and stand alone in New Mexico and California.
rk>I would like to add a PV inspection "punch list" at my site. As with my book, I'd like to start from an understanding of the core safety issues. Does anyone subscribing to this newsgroup have any first hand knowledge of PV related hazards?
Yes, PVUSA has several staff members, including myself, with the physical, personal, and electrical hazards of terrestrial PV.
rk>To prepare myself, I've been reading through the John Wiles' book Photovoltaic Power Systems and the National Electrical Code. It's raising a number of questions for me i.e. are people installing arrays that put out more than 600 volts open circuit voltage?
Yes, Omnion's 3200 series inverter typically needed a 600 Volt input. However, this was a larger power unit (>100 kW).
rk>So is it safe to say on a residential job I won't be seeing any 600Volt required components. I'm most concerned with non-engineered residential jobs. I assume larger commercial jobs will have wet stamped engineered drawings that will make it easier to plan check for code compliance.
top of page Roof modules shall have Ground Fault Protection [690-690-5]
As mentioned in the latest issue of Home Power, inspectors in states that have adopted the 1996 NEC are requiring installation of ground fault protection (GFP) devices. John Wiles mentions that only two companies build listed devices: Trace and Ananda Power. I talked with an Ananda Power rep this morning. Ananda builds a UL listed power center primarily as an interface between batteries and PV (load control center). They will build any type of control center, a GFP only for example, if you supply the necessary funds and configuration data: voltage and current.
In my opinion the most difficult item in the system to install and maintain is the GFP. Where do you put it? How do you test it? How do you disable it to work on the array? Remember that this requirement is to reduce fire hazard on residential roof mounted arrays. How often does the home owner go up on their roof. As an inspector you will be looking for type accepted equipment. Ananda has a UL rating for their equipment. Not to say it functions properly, just that it is UL listed. I don't know about Trace. BTW even though I have been in this field for some time I continue to be stumped by the nuances of the various codes. PV is simple, installing PV to code (NEC, IEEE, IEC, or UL) is permuted at least.
If you don't isolate conductors, aside from obvious ground fault potential, the cross talk between conductors may hazardous to your sensitive electronics, whether it is VCR or inverter.
John mentions that article 300-3b mentions the array wires must be installed in the same conduit to reduce inductive stresses on protection equipment.
rk> Are you talking about an arcing fault from the roof array that causes damage to the electronic equipment.
No, I was talking about switching noise that is present on the dc (array) bus that is induced or passed through to the ac side of the system. The array experiences ringing and emits some EMI and RFI. Most of the inverters on the market will wipe out AM radio reception. I heard of a homeowner with a satellite receiver in the RV that can't keep the bloody thing programmed because of the inverters.
By ground fault I do mean a fault to earth, usually through the array frame, but often through conduit or adjacent conductors. Of course the more recent versions of the code which require ground bonds improve the situation. However, installers today still use painted surfaces to establish a ground and the resistance may be higher than other locations within the array.
rk> I think we're having some problem with our terms. For me, a ground fault is a fault to earth or non-current carrying components of an electrical system. A phase to phase fault presents a different set of conditions. Is that what you > are talking about, an unintentional cross connection between distinct voltages? Or are you talking about inducing a current in the ac system when the fault opens and closes?
See above, PV arrays are only dc until connected to an inverter.
PV output conductors must be isolated from other systems(AC) [690-3b]
JPL Block V specifications were such to provide at least a minimum of 1000 Volts isolation. Current standards under development are derived from the JPL module tests and are more like twice the system voltage plus 1000 Volts.
rk> By "isolation" I assumed the NEC was referring to physical isolation between conductors of different voltages i.e. don't share ac with PV conductors in a conduit.
Yes, I think that is true.
With respect to the NEC: You know better than I the intent and good sense of separating the conductors of different voltages or types. In general wiring of different types must be run in separate conduit or wireways. I am most familiar with data cabling which can be quite interesting when run a a 480 Vac 3 ph. cabinet. We have an inverter at PVUSA that switches at 4 kHz. All of the data cabling is isolated from the system electrically, but runs in the cabinet. The coupling is good and the noise levels induced are significant. Our data system is limited to 5 volts input and induced voltages are in excess of 20 Vac. Close physical wire runs between the ac and dc wire runs would result in some inductive coupling and may cause problems with the inverter or other equipment that is sensitive other than fundamental frequencies.
There is still an argument by Solarex that shorting the array is PV abuse and others contend that shorting the array on ungrounded systems is more dangerous than open circuit. The bottom line is that shorting the array (crowbar) is a function of the inverter and that appears to be acceptable to most PV experts (whether it is a good idea or not is arguable).
rk> Excuse me if I'm a little slow on the uptake. How would shorting the array to the inverter help. Isn't the array downstream of the array? If the array is going to ground through the roof(metal gutter?) how would a shunt downstream remove the fault? (Be patient, I'm new to this stuff).
It doesn't, but that was as good as it gets for awhile. John covered his bases by noting that the inspector may fall back to a previous version of the Code that doesn't required a GFI.
A phase to phase fault is relatively benign for a PV system. The characteristic IV curve shows that the power available for the system is small if the short is good (low resistance), however if the phase to phase fault results in a resistance of a few ohms it is possible to have the array at peak power into the faulted load. This is an interesting thought problem and bears some research. I believe that generally the phase to phase fault is less of a fire hazard than a fault to ground through the array frame or wiring/conduit.
You are correct of course. The availability of equipment that meets the code to short the array was limited (or not available). So previously installers just ignored the requirement and put the shorting circuit in the inverter. As a matter of interest the shorting/isolating part of 690-5 is subject to change in future versions of the code. It is very unpopular and controversial. The procedure is to: 1) detect the ground fault, 2) open the grounded conductor, 3) short the array to disable it. Finding an acceptable device that is type accepted is nearly impossible, although Trace and Ananda reportedly provide the equipment. There is a product called an arc detector made in Europe that we will be testing real soon now. I hope to have more opinions, facts, and equipment available at the workshop.
Word is the 1999 code will require fuses or blocking diode in the module junction boxes. The question remains will they be accessible when they clear? Do the fuses provide any value? PV modules are inherently current limited. The fuses or diodes provide protection against backfeed by a relatively infinite source (grid). It is possible so I wouldn't poo poo the idea, however fuses or circuit breakers in the dc connection device seem to be easier to monitor and service.
What's a GEC?
rk> GEC stands for grounding electrode conductor.
Where does it say as close to the modules as practical?
rk> In the fine print note for 690-42>
The FPN sez that this is for maximum lightning protection. And 'close as practical' is a wide latitude for either the designer or inspector. Besides the FPN is 'explanatory' and is not a mandate or requirement.
BTW I am putting together a workshop outline, do you have any specific interests or requests?
Sorry, I forget to fill in the gaps sometimes. Jet Propulsion Labs, Pasadena was in charge of the terrestrial low cost array program. They were charged with developing techniques for improving the state of the art for terrestrial PV in the early seventies. The program was complete sometime in the mid to late eighties. Out of the program comes many standards and techniques for installation and application of PV on earth including the Block program for terrestrial MODULE development. The Block program concentrated on improving the previously unreliable contacts, cell grids, junction boxes, etc. Block V (five) modules must withstand impact by 1 inch hail, high voltage isolation between the module and frame, withstand 125 mph wind loads, and more. The bottom line is that meeting the Block V standards resulted in and estimated twenty year module life when tested in the JPL environmental chambers with accelerated test programs.
I would like to extend an offer to all the reader of this newsgroup to make an appointment to come and tour the Photovoltaics for Utility Scale Applications (PVUSA) facility, 1 mile north of Davis, CA (fourteen miles from Sacramento). Call 916-753-0725 to arrange a tour. Tours are conducted daily, Monday through Friday, from 9 AM to 3:30 PM.
PVUSA is currently installing three 4 kW roof top PV systems on a simulated shingle roof. Come on out and see how Sacramento Municipal Utility District installs the PV Pioneer systems.
