Ground Mount grounding, NEC and local Qustions

Pgovetom

I'd like to hear about anyone who has experience with difficulties around "what exactly is required and why for ground mount systems". I have installed a 12KW ground mount system in California. We use the 2005 NEC so far. I tried to determine if ground rods were required at my four racks each carrying 14 210 watt panels with pairs of strings returning to Sunny Boy 7000 inverters.

The panels are mounted on ProSolar GroundTrac systems which use 1 1/2" galvanized steel pipe for posts sunk 4 feet into concrete spaced by 8 feet so there are 10 per 25 foot by 10 ft rack. Then aluminum rails carry 3 high panels by 5 wide with one spare for future use. The pairs of racks/strings wiring go to 2 midnight boxes to provide a ground bar, 15A fuses plus the underground transition. They are a little overkill but did the job nicely.

I planned to drive ground rods and bring the continuous #6 solid copper wire that hits all the metal with lay-in lugs and then to the Midnight ground bar. Then the midnight ground bar also has a #8 green THWN-2 along with the two strings back to the Sunnboy 7000 PV frame ground lug. The Sunnboy also has an Earth ground to my main panel Earth rod and an AC ground carried along with L1,L2, Neutral through a small inverter combining Square D load center with 40A breakers from the inverters and an 80A toward my AC disconnect. The AC disconnect has a direct #6 solid to the Earth ground as do the inverters and my existing main panel.

I was told by 3 experienced electricians in solar not to put ground rods at the panels as we have no lightning in Northern California where I am. When I got my final inspection, I was told I needed ground rods at the panels. The code section refers to "Structures" so I learned my county defines a ground mount structure a "Structure" for an NEC point of view. The NEC does require "Structures" to have ground rods. But I have 4 structures with pairs each bonded through #6 solid copper. So I checked with John Wiles and he said the NEC 2005 doesn't require it but any local is free to define the mount a structure bringing in those requirements. Ok so I put the rods in at the Midnight boxes and bonded them to the already bonded metal. I was worried they might want 4 rods since I have 4 structures albeit 4 feet apart and pairs bonded. I asked the head of the building inspection who didn't know and he asked me to just have it inspected. I asked if he had someone who knew so he gave me the number of one of his people.

This guy went around a bit saying things like "you need one rod per string" which told me he was not that knowledgeable. Some systems have many "Strings" and a rod per string makes no sense. He finally said I should put one per rack so I went out and drove 2 more rods and bonded them to the racks that didn't have a Midnight.

So now I have 5 ground rods in a no lightning strike area. I'm only doing it to pass the local inspection where they can't tell me "why". Should I be concerned about the Sunnboy GFI circuits seeing leakage through Earth back into my panel AC ground and into the inverter via Earth/AC path? The negative lead is ground referenced by the Sunnyboy but not connected to ground. The positive leads are hot out to the panels.

Does anyone know the real story on grounding and whether 1,2,3,4,5 rods in my example is adequate from a safety/lightning/NEC perspective? Does anyone know why one locale would call a ground mount a "Structure" bring in those requirements, If others don't. Is it superfluous? Since Sonoma County where I live is the #1 county in the US leading county energy independence with its own programs including property tax incentives but their building department doesn't seem tuned in. Is this typical? Why would they also impose the harsh code when its not really needed. Seems crazy!

thanks

Kamala

Re: Ground Mount grounding, NEC and local Qustions

Experience? I have little. What I have learned is that grounds should be single point. On the other hand, if the purpose is to dissipate HV, HA, TERRA POWER, (as in lightning strikes,) a single point collector (aerial) with heavy gauge to a star pattern, deep ground connection ( 8 feet deep with salt) could achieve the desired result. E. G. dissapation of energy.

niel

Re: Ground Mount grounding, NEC and local Qustions

yup, they can make you do that, but you can optionally impress them and tie each rack rod with bare #6 copper wire buried 1ft or so underground making a boxed pattern and don't bury it so they can see it. not too sure if by code you have to do that, but it is good practice to stop voltage differentials from appearing on each rod. i do admit i had some difficulty in following you with all of this, but i assume they are making you have this "ground wire" go back into the house where it has to meet with the main service and then out to the main service ground. this is a hazard in my opinion and my preference is to bury a bare copper wire from the external rods to the main rod as per the underground bare wire like each rack rod would be tied to each other with no "ground wire" entering the house. my way would be better imho, but it would be more work for you and may be overkill in an area that rarely gets lightning.

Pgovetom

Re: Ground Mount grounding, NEC and local Qustions

Thanks

No thats not quite what they are doing. They are making me add 4 rods, with pairs of rods spaced 4 feet apart ( the racks are 4 foot spaced) and bonded via an underground conduit #6 ground/bond wire. The 2 pairs of continuous #6 wires that hit all the metal ( panels, aluminum rails and 1 1/2" pipe) via lay-in lugs and returns to the midnight combiner/fuse box on one rack were already installed along with the #8 green inverter 150' frame ground from racks to inverter with the DC pairs. That originally put the rack frame ground Earth connection to a rod back at my original service panel rod about 150 feet away. They now asked for 4 new rods at each rack in addition.

The electrical design requirements due to the inverter and its GFI system requires 3 ground connections per inverter.

1) A #6 directly to the house service panel ground rod that is about 6 feet away.

2) An AC ground that follows its AC output path with the L1,L2,Neutral back to the grid tie in connecting all the "AC boxes" ( inverter, load center, AC disconnect, conduit and main panel) for safety reasons. The main panel then connects that via its #6 armored ground wire to the same existing 8 foot rod driven when the panel was installed years ago.

3) The PV frame ground lug on the inverter which is a wire that a green frame ground #8 that goes along with the string DC #8 wires and ties to the Midnight box ground bar which is mounted to two of the 4 panel racks. Then the Midnight box ground bar is also tied to 2 #6 solid wires, one continuous run going to each rack associated with that inverter to ALL metal pieces. Those #6 wires are used to bond every piece of metal on the rack ( 1 1/2" pipes not already bonded, 10 aluminum panel mount rails, 14 solar panels) and used stainless steel bolts and star washers and nuts and Lay-in lugs on every metal piece.

Before my inspection, the racks were Earth grounded via the #6 wire to the midnight boxes ground bar and through the green 150' underground in conduit #8 THWN-2 to the inverter's frame ground lug which in turn ties it internal to the inverter to the original Earth ground rod wire at the panel. That Earth grounded all the metal, Midnight boxes and inverter all together via the single Inverter Earth ground to my old rod 6 feet away. Since there are two ( one per inverter) green #8 frame grounds from the midnight boxes plus their two Earth connections with the AC ground via the boxes and Main Service Panel a backup ground. It does however suck for a lightning strike out at the panels. That 150" of stranded #8 will look like an inductor to that lightning current. The AC impedance would be too high to protect the panels.

But each rack has 10 1-1/2 rigid galvanized steel pipe posts sunk 4 feet into the ground with concrete. The lightning is going to look for the lowest impedance and most direct path to Earth. It won't even see the #6 solid wire when there are 10 solid pipes 4 feet into the ground and we only have lightning here when its winter and the ground is wet. My ground is moist even in the summer at 4 feet. With 10 inch and a half steel pipes into moist concrete and soil at the bottom, 99.999% of the current will go directly down those posts into Earth. The #6 solid wire to the rod will carry some current but is only useful if my soil was dead dry at 4 feet which it never is. It would also require the lower 4 feet of the ground rod be in moist soil or it does no good. Totally dry soil or rock is a good insulator. Even if a resistance test is done on the rods, it will pass when there is some moisture to 8 feet but if it dries out, the rod doesn't do anything.

So if there was a lightning strike on the panels, it would have two paths to ground. It would follow the 150" #8, through the inverter and down the #6 inverter earth ground to the rod and discharge into the Earth. That is a lousy lightning protection scheme but here in Sonoma in my 50+ years, I have never seen a lightning strike hit the ground and only seen lightning maybe 20 times, every 3 years or so.

But if a string wire with about +400V somehow got shorted to any piece of metal on the frame of any rack anywhere it would be safe. It would short the panels dropping the voltage way down or trip the GFI or both. The GFI on inverter would trip if a person touched it while standing on Earth and trip so quickly due to the current returning on the ground leg of the inverter, it would not be harmful. The GFI protects from high voltage on the panel rack in the long #8 case or the rods at the panel case. If the ground path had poor current carrying capacity ( high resistance between rack located rod(s) and the main service panel rod, then the 14 panels would go to short circuit current but not zero volts but not as harmful as 400VDC.

In both cases, there is a reliance on the ground path not failing or the GFI detection scheme not being defeated or fooled by leakage currents. Having the 5 grounds is essentially 5 way redundancy since everything is bonded a couple different ways. If any one rod connect fails, the duplicated bonding and duplicate rods will provide the ground for safety and GFI tripping.

I am concerned because 2 of the electricians told me that the SunnyBoy GFI could be fooled by the multiple rods and return current paths via Earth to the AC ground path and the #8 green wire path which is required by Sunnyboy and the NEC. The Sunnboy GFI is looking at the currents in both the positive leg and negative leg and ground. It expects 100% of the current on the positive to return to the negative and none via the Earth. If it sees some current ( maybe through your body) coming in on its AC path via the Earth, it depends on the design as to whether that gets caught. The Sunnyboy manual is very specific about wiring the DC leads, frame ground to the racks/panels and AC ground and its own Earth and they are different lugs.

Whoever designed the GFI detectors assumed the wiring was done correctly and may be monitoring currents on all the leads. It would like to see 100% of the current on the positive leg returning on the negative leg within some tolerance and with some time delay/constant. With 5 rods plus 2 direct frame grounds and AC, frame and Earth all possible paths, I'm hoping what the inspector called overkill but redundant might also impact the GFI detection circuits and cross couple them.

Low cost solar will never arrive if the code keeps adding redundancy that is not even needed. Its not the space shuttle. My whole house has one rod 50 feet away DUH!

niel

Re: Ground Mount grounding, NEC and local Qustions

i hear ya, but they can still make you do that stuff. be glad they didn't specify a 120ft ground rod to reach moist earth below. best that you can do is go along with the inspector and get it passed.

Solar Guppy

Re: Ground Mount grounding, NEC and local Qustions

simple answer is ground mount has no NEC requirement for ground rods or even grounding the PV frames for that matter.

Reality is your inspectors are basically GOD and can require anything the they deem necessary ... so the right answer is do what the person requires to get the permit signed of OR protest the request up the supervisor chain ( never pleasant )

Cariboocoot

Re: Ground Mount grounding, NEC and local Qustions

I'll just say never tie a lightning ground to a system ground. Especially do not run it inside a house, tie to the box, and back out.

Problem with inspectors is, they don't always know what they're doing. They're civil servants with a book of rules which they may/may not understand. Trouble is, they have the final say-so.

I can't imagine what your soil conditions must be like to require a grounding system like that. Desert? Most of the time a single rod driven 2' below frost level is more than enough for either lighting (per strike point) or system grounding. If the terrain is rocky and you can't go deep you'd also have to 'spread it around'.

How do I know? Lightning burns out lots of motors in agriculture (former Emerson Electric employee).

niel

Re: Ground Mount grounding, NEC and local Qustions

"Most of the time a single rod driven 2' below frost level is more than enough for either lighting (per strike point) or system grounding."

now that definitely isn't true on either count and an 8ft copper rod is a compromise as it too is not enough for the average soil. i wish i still had a link to an article i read and commented on here many years ago as it was quite enlightening. i guess i had too many crashes and os changes while seeing many bbs changes here on the forum too. i may try to search the net for it again, but i don't hold high hopes of finding it. before i do i'll research the archives again as it may show up.

found something similar. the original one i was referring to was much longer and more in depth and even addressed such things as skyscraper grounding. of course this concerns nec rules and local inspectors can lower or up the requirements if they please. here is a quote from the article.
"[FONT=Verdana, Arial, Helvetica]As shown by the data presented, the average 8 to 10 foot ground rod will not meet minimum NEC code requirements for earth resistance."[/FONT]
happy reading.
http://www.cpccorp.com/deep.htm

Cariboocoot

Re: Ground Mount grounding, NEC and local Qustions

I never said it met code. But it is what is done in northern climes on loam. Although it's often done WRONG because idiots have no idea how deep frost is and they forget things like 'per point' or to multiply it out for the Amperage rating of the service.

I'm probably just not explaining it right. Brain shut-down half-way through post again!

And in some other thread people are recommending grounding to concrete foundation. Tell me how that's allowed? If the concrete is wet, it conducts. But any civil engineer will tell you you don't want your concrete saturated with water.

I suspect a lot of people do things horribly wrong and just 'luck out' for fifty years or so.

mike95490

Re: Ground Mount grounding, NEC and local Qustions

Cariboocoot wrote: »

And in some other thread people are recommending grounding to concrete foundation. Tell me how that's allowed? If the concrete is wet, it conducts.

Concrete stays damp/moist, and releases lots of conductive ions into the surrounding soil, making a very good ground, except the the driest desert. The rebar should be well joined in the pour. Read the article in the thread, it surprised me.

http://forum.solar-electric.com/showthread.php?t=3583&highlight=ufer

BB.

Re: Ground Mount grounding, NEC and local Qustions

Ufer Ground History:

The term "Ufer" grounding is named after a consultant working for the US Army during World War II. The technique Mr. Ufer came up with was necessary because the site needing grounding had no underground water table and little rainfall. The desert site was a series of bomb storage vaults in the area of Flagstaff, Arizona.

The principle of the Ufer ground is simple, it is very effective and inexpensive to install during new construction. The Ufer ground takes advantage of concrete’s properties to good advantage. Concrete absorbs moisture quickly and looses moisture very slowly. The mineral properties of concrete (lime and others) and their inherent pH means concrete has a supply of ions to conduct current. The soil around concrete becomes "doped" by the concrete, as a result, the pH of the soil rises and reduces what would normally be 1000 ohm meter soil conditions (hard to get a good ground). The moisture present, (concrete gives up moisture very slowly), in combination with the "doped" soil, make a good conductor for electrical energy or lightning currents.

Ufer techniques are used in building footers, concrete floors, radio and television towers, tower guy wire anchors, light poles, etc. Copper wire does not function well as a "Ufer" ground due to the pH factor of concrete (+7pH is common). The use of steel reinforcement as a "Ufer" ground works well and concrete does not chip or flake as has been found with copper. The use of copper wire tied to the reinforcement rods outside the concrete shows none of these problems...

-Bill

solartek

Re: Ground Mount grounding, NEC and local Qustions

Solar Guppy wrote: »

simple answer is ground mount has no NEC requirement for ground rods or even grounding the PV frames for that matter.

Under NEC 2008, this is no longer true. See 690.43 and 690.47(D)

Scott.