The ground is not "The Ground" and the pathology of NEC code.

[glandpuck]

For what it's worth, when we built our ground mounted arrays in 2017, they were anchored either by 12 inches x 42 inches of concrete into schedule 40 steel posts or the posts were anchored into 1,000 pound 2x2x2 concrete poured blocks. The county inspectors still required that each panel have an approved grounding screw with a proper bond to the panel frame attached to copper wire, 6 or 8 gauge. One end of the copper wire was attached to a 10 foot grounding rod at the array and the other end terminated in the ground buss bar of the combiner box. And we still connected a DC lightening arrestor to the combiner box. In 7 years now no issues at all. However, in San Diego county maybe a handful of lightening strikes in out area over that time. The idea is ground will direct any charges into the ground itself via the concrete which is a good ground conductor as well as the rod and the ground will have zero volts relative to the hot.

With our inverter and charge controllers and sub pannel, we were required to place 2 10 foot grounding rods 10 feet apart and run the same solid copper wire between them and into the inverter, battery bank and sub panel. Well it works and passed inspections.

Where I think proper grounding is most important is a case where somebody connects a neutral that is damaged (like a staple through the Romex insulation) to the ground conductor to get a circuit to work, but now risks an electrocution this way.

THE SMART INSPECTOR WILL CARRY A METER AND BE ABLE TO TELL IF YOU CUT ANY LENGTH OFF OF THE GROUNDING ROD TO TRY AND FOOL THEM.

 

[ElectricIslander]

Is your resolution good enough to map small volumes or areas? If it is, it offers a method to actually test grounding/earthing systems.
With that kind of data available, NEC could produce code that actually makes sense.
Code inspectors could approve or deny grounding systems based on actual data.

Magneto-Tellurics are usually used on a large scale (multi kms in extent and depth) as there are few other techniques available (other than very expensive crustal scale seismic).

For smaller scale (metre scale) we often use galvanic resistivity (based on arrays of from three or four electrodes up to hundreds of multiplexed electrodes. Electrodes are interrogated in two pairs usually (two to inject current, two to measure voltage). Often industrial grounding designs for large facilities like substations or windfarms etc. are based on these surveys in order to meet the required ground resistance. The resolution can be as fine as desired however it drops off rather rapidly with depth due to the physics of electrical flow paths. Interestingly, current flow refracts at a resistivity interface in a similar way that light refracts and changes angle between water and air.

There are IEEE and ASTM standards online for how to perform grounding studies with geophysical methods. For large projects the effectiveness of a grounding system is indeed tested based on measurements using these standards. For home-owners a prescriptive approach such as XX feet of galvanized rod driven into the earth is usually sufficient.

There are lots of publicly available examples of data sets - Google "Electrical Resistivity Tomography" or "Magneto-Tellurics".

We often use rather exotic means to get a good temporary reference connection to the earth. One popular method is to bury a disposable baby nappy saturated with salt water. The polymer absorbent in the nappies holds the salt water . Despite the manufactures claims of "no Leaks" just enough seeps out to make a good ground! It's all about getting enough surface area in contact with a moist subsurface strata that has enough fines content (large surface area material such as clay or silt holding bound water) and enough free interstitial moisture to conduct the current into the ground.

 

[McKravitts]

Magneto-Tellurics are usually used on a large scale (multi kms in extent and depth) as there are few other techniques available (other than very expensive crustal scale seismic).

For smaller scale (metre scale) we often use galvanic resistivity (based on arrays of from three or four electrodes up to hundreds of multiplexed electrodes. Electrodes are interrogated in two pairs usually (two to inject current, two to measure voltage). Often industrial grounding designs for large facilities like substations or windfarms etc. are based on these surveys in order to meet the required ground resistance. The resolution can be as fine as desired however it drops off rather rapidly with depth due to the physics of electrical flow paths. Interestingly, current flow refracts at a resistivity interface in a similar way that light refracts and changes angle between water and air.

There are IEEE and ASTM standards online for how to perform grounding studies with geophysical methods. For large projects the effectiveness of a grounding system is indeed tested based on measurements using these standards. For home-owners a prescriptive approach such as XX feet of galvanized rod driven into the earth is usually sufficient.

There are lots of publicly available examples of data sets - Google "Electrical Resistivity Tomography" or "Magneto-Tellurics".

We often use rather exotic means to get a good temporary reference connection to the earth. One popular method is to bury a disposable baby nappy saturated with salt water. The polymer absorbent in the nappies holds the salt water . Despite the manufactures claims of "no Leaks" just enough seeps out to make a good ground! It's all about getting enough surface area in contact with a moist subsurface strata that has enough fines content (large surface area material such as clay or silt holding bound water) and enough free interstitial moisture to conduct the current into the ground.

That's very interesting and I thank you for sharing the info.

I have to wonder if there are any software programs available to us hobbyist where one can input the conductivity/resistivity of the ground, design a grounding system and then check it by inputting lighting strikes at various points.

I know from many of my encounters with code inspectors, having pre-engineered your work goes a long way toward getting it approved.

 

[McKravitts]

I thought the conventional wisdom in electrician land is that actual data is harder to get than to just overkill on ground rod depth and dig/drive a little more. Because you would either need to get some precision measuring equipment or get some geology and engineering done.

Unless you can use modeling to justify some simpler rules of thumb.

I think that "conventional wisdom" when it comes to dsolar arrays is still being debated, even among the authors of NEC code.

 

[ElectricIslander]

I think that "conventional wisdom" when it comes to dsolar arrays is still being debated, even among the authors of NEC code.

The trouble is that design of anything, even grounding systems always involves compromises and competing objectives.

The code is developed to make the most common situations safest. This means that the code might not provide the optimum design solution in certain oddball configurations.

Grounding a solar array can involve competing objectives. For example, on a really long PV run to a house with a ground array one may have to choose between designing protection against lightning ground gradient effects (i.e. multiple ground points to keep all components as close as possible to local earth potential) versus avoiding ground loop effects (ie use of single system ground point).

I haven't heard of any publicly available lightning protection software. There are so many variables that it would be difficult to predict much of anything.

 

[McKravitts]

The trouble is that design of anything, even grounding systems always involves compromises and competing objectives.

The code is developed to make the most common situations safest. This means that the code might not provide the optimum design solution in certain oddball configurations.

Grounding a solar array can involve competing objectives. For example, on a really long PV run to a house with a ground array one may have to choose between designing protection against lightning ground gradient effects (i.e. multiple ground points to keep all components as close as possible to local earth potential) versus avoiding ground loop effects (ie use of single system ground point).

I haven't heard of any publicly available lightning protection software. There are so many variables that it would be difficult to predict much of anything.

Would something like the following simplified model be of any use?

Since lightning is so powerful relative to the base line voltages/currents in the earth, the ground could be modeled as a three dimensional resistor array with no electrical activity. As a first approximation, the resistors could be set to the same value. The edges of the array could be modeled by resistive connections to an ideal current sink. Any errors this may cause can be minimized by simply making the array much larger than the area under coideration. The eathing array under consideration could be modeled as zero ohm resistors added/connected into the homogenous array of fixed ohm resistors. Ground rods (points) modeled as connections to an ideal current sinks.The facility/house could be modeled as a resistor connected to an ideal current sink
To test the effects of a lightning strike at any one point, you/the computer places a source of current at a selected location on the surface of the resistor array and determines the current across the facility/house. The computer could test many points on the surface depending on the resolution the user wants.
The earthing array (placement of zero ohm resistors ) could be modified by the user until a satisfactory result is obtained.

It wouldn't be exact science but it may be an improvement to the guessing game many solar installer are playing now.

 

[ElectricIslander]

Current flow into the earth is indeed three dimensional and easy to model without computers in simple cases. The only wrinkle is that resistivity often varies in 3d which can produce complications. Resistivity (unit of Ohm-metres) is the bulk measure of resistance in 3D, and is defined as the resistance across the face of a 1 metre cube. Ohms law still works but you have to add geometric correction factors depending on where you measure voltage and current. (which becomes current density in 3D)

There is an excellent explanation here of how electrical currents spread out into the earth and create concentric shells of equal voltage.

 

[Batvette]

I have a question. I used to live in a 70 year old duplex near the beach with 2 prong outlets and really old wiring. The spare bedroom was a workshop with a large metal frame workbench, the floor was asbestos tile over concrete.
If I got out of the shower in my bare damp feet and touched any metal on the workbench I got a mild shock. One day I was sitting at the bench with a multimeter, put a lead on the bench and another to the floor and it measured 85 volts. I never did find any faults in the bench wiring(just a 2 tube fluorescent light attached) or the power strip and appliances attached to that. (o'scope, power supply, a marantz receiver, a lot of wall warts)
Anyone know what was going on there?
This was right at the beach and there was salt water in the ground in places that touched the slab.

 

[Warpspeed]

The concrete floor will be pretty well guaranteed to be at local ground potential.
Your workbench will be at 85v above ground, due to electrical leakage between the ac wiring and the metal of the bench.
It takes very little leakage to do that, and its not uncommon.

The way to fix it is to connect the metal frame of your bench, and any other exposed metal that is in close proximity to wiring, like metal conduit, to a point that is solidly grounded.
The leakage current then goes through that ground wire to ground, instead if through YOU to ground.

The ideal ground connection would be the usual rebar reinforcing steel cast into your concrete floor. Almost as good would be one or more metal stakes driven deep into the ground, or a significantly buried long metal water pipe if available.
Any appliance that has exposed metal, should (must!) be grounded, usually through a third pin on the mains plug.

 

[Checkthisout]

The trouble is that design of anything, even grounding systems always involves compromises and competing objectives.

The code is developed to make the most common situations safest. This means that the code might not provide the optimum design solution in certain oddball configurations.

Grounding a solar array can involve competing objectives. For example, on a really long PV run to a house with a ground array one may have to choose between designing protection against lightning ground gradient effects (i.e. multiple ground points to keep all components as close as possible to local earth potential) versus avoiding ground loop effects (ie use of single system ground point).

I haven't heard of any publicly available lightning protection software. There are so many variables that it would be difficult to predict much of anything.

I would think the best way to protect the array and equipment from lightning would be to have lightning rods taller than the array that are attached to their own ground system.

 

[mxbadboy]

So. let's take a real world situation, one that is a common topic on this forum.
A lot of people have a solar array on a shed, barn, or ground based array. Say the remote array is 100 feet from their house and that the remaining solar equipment is in the house.
The ground potential could be large between the house and the array.
How should the array be grounded, or should it just not be grounded?

The shed, barn, solar array metal should all be "bonded" together by one continuous bonding wire and that wire to one grounding rod.

 

[glandpuck]

Our ground mounted arrays are grounded as follows. They were designed by a solar engineer and passed inspection and permitting. The array structure is schedule 40 2 inch galvanized steel pipe in the ground 36 inches in 12 inch concrete tubes. Above ground the structure is connected by Hollaender connectors into one big frame for the panel rails and panels. Each panel has an approved ground contact connection to it. The entire array has a single solid copper wire connecting from a buried 10 foot ground rod at the array, across all of the panels, into the combiner box (Midnite Solar MNPV6 in our case) where it terminates on a ground buss bar. At the combiner box, a Delta lightening arrestor is installed. Then the PV +, - and ground wire travel into the DC disconnect switch. At the DC disconnect switch, the switch and all of the arrays grounding wires land on a ground buss. Then this ground travels with the PV =/- of the 3 arrays into the garage and lands on the inverter ground buss bar. Just outside the garage are placed 2 10 foot grounding rods 10 feet apart. They are connected by a single ground wire which then travels into the the inverter and lands on the ground buss bar. The inverter is grounds, the grid in is grounded and the subpanel out is grounded. The batteries are all grounded as well as the battery cabinet. All connections in the garage that carry current travel in metal conduit which also provides grounding points. The subpanel has installed a Delta brand AC lightening arrestor.
We are in a very low lightening risk location, maybe 5 lightening strikes in the area in a year. But the system is grounded to dissipate electricity and protect equipment and people.

 

[McKravitts]

The shed, barn, solar array metal should all be "bonded" together by one continuous bonding wire and that wire to one grounding rod.

Agree, but that grounding rod should be no more than 6 feet away from (codes may vary) and bonded to your house's ground rod.
That's not always easy of even possible in many situations.
I've seen/read that a long ground/earthing run from the array to the facility (house) should use no higher than 6 gauge bare copper and additional ground rod attached every 6 feet.

The purpose is to flatten any natural voltage gradients from the house to the remote array.

 

[gotbeans]

Confused, confounded and puzzled are terms found throughout this forum about NEC code regarding “grounding”.

We are conditioned to believe that there is object; “the ground” that is electrically quiet, safe and always zero volts and could never be a party to harming anyone or damaging our equipment.

The authors of NEC code certainly want us to believe that, but their frequent updating, changing and sometimes seemingly contradictory requirements can only lead one to wonder if is there more to the ground than they care to share.

The ground is zero volts right! But relative to what? The ground.

If you had very long meter leads and placed one on your ground rod and the other on someone’s a mile away would your meter read zero?



I am starting this discussion in hope of clearing up some of the mystery regarding grounding and possibly help people develop grounding systems that meet code and protect their equipment as best as possible.

If you had very long meter leads and placed one on your ground rod and the other on someone’s a mile away would your meter read zero?

as far as I have read, no. Even at 6ft apart they can be different. (They require a wire connecting multiple ground rods I'm pretty sure? due to multiple reasons probably)

Soil material matters a lot though, that's why those no battery moisture sensors that are anode/cathode with no battery work really well in some places and not at all in others

 

[ElectricIslander]

as far as I have read, no. Even at 6ft apart they can be different. (They require a wire connecting multiple ground rods I'm pretty sure? due to multiple reasons probably)

Soil material matters a lot though, that's why those no battery moisture sensors that are anode/cathode with no battery work really well in some places and not at all in others

As others have said - our idea of "the ground" as a uniform grounding reference isn't really the real world situation.

For our geophysical surveys we often have to monitor the variations in the background reading between two electrodes. For surveys like self potential and other sensitive work, we log the time varying voltage between two reference points and subtract that time varying background from our signals as the surveys progress.

Voltages induced in the ground vary all the time from things like daily changes in ionospheric and telluric earth currents that writhe around the earth as the sun comes up, as the auroras form and lightning storms flash in the distance and even when rain falls etc. etc.

Our earth is a very dynamic beast electrically. It's what you'd expect of a conductive ball of dirt hanging in space right next to a star brimming with seething hot plasma ....?