Grounding Questions (Fuse panel, grounding rod, frame of panels)

[max2k]

You are not showing a bond to N at the Inverter.

Now I hope I got it right. Green EGCs are meant to be local to each 'subsystem' if I understand you correctly. So in this setup in case of either Inverter or Array site becomes subject of nearby lightning strike its ground potential will jump relatively to the other 2 sites. What is the recommended way to protect from damage? Left like drawn L1/N/L2 for example could be subject to tens of thousands of volt difference and that will create currents enough to damage equipment. Shall we utilize TVSS on both input and output lines to 'clamp' voltage spikes to local ground potentials? The wires between sites would still be subject to high current spikes but at least those currents would bypass equipment.Grounding_2.png

[bcroe]

Now I hope I got it right. Green EGCs are meant to be local to each 'subsystem' if I understand you correctly. So in this setup in case of either Inverter or Array site becomes subject of nearby lightning strike its ground potential will jump relatively to the other 2 sites. What is the recommended way to protect from damage? Left like drawn L1/N/L2 for example could be subject to tens of thousands of volt difference and that will create currents enough to damage equipment. Shall we utilize TVSS on both input and output lines to 'clamp' voltage spikes to local ground potentials? The wires between sites would still be subject to high current spikes but at least those currents would bypass equipment.Grounding_2.png

I went out and checked some details here, most from before I arrived. That drawing looks essentially like
mine (ignoring disconnect boxes), EXCEPT a 6 gauge ground runs from the inverter ground bar, out to the
combiner box and then to the frames of all panels. Bruce Roe

[max2k]

I went out and checked some details here, most from before I arrived. That drawing looks essentially like
mine (ignoring disconnect boxes), EXCEPT a 6 gauge ground runs from the inverter ground bar, out to the
combiner box and then to the frames of all panels. Bruce Roe

I know, I was trying to reflect SK suggestions otherwise it is all yours- I have pure academic interest in this as my own setup is much more common: just single house with panels on the roof.

[Sunking]

Now I hope I got it right. Green EGCs are meant to be local to each 'subsystem' if I understand you correctly.

Graphically you got it right. Only thing that looks out of place is that short green stub at the Inverrter that goes no where. Delete that and you got the right idea.

I do not know what you know about utilities and their regulations and standards. They are controlled by NESC which has nothing to do with NEC. If you look at any utility supply like the common single phase 240/120 almost all homes in the USA have is exactly what you are showing in your drawing, three circuit conductors of L1, L2, and N. There is no ground from the utility.

When it hits your Meter Box aka Service Disconnect location the N Conductor is boned to Earth. From that connection point you have a Green EGC and White Neutral conductor to your main breaker panel.

What is the recommended way to protect from damage? Left like drawn L1/N/L2 for example could be subject to tens of thousands of volt difference and that will create currents enough to damage equipment.

OK you are over looking something but on the right track. What you are not seeing is the Neutral Wire between Inverter and you home is the Bonding Conductor.. Yes current will flow in that wire if there is a Strike, but it is only flowing between the Ground Electrodes, not if the equipment is wired properly.

Where you are on the right track is you can use TVSS. You only need 2-Modes installed between L1-N and L2-N, no ground required except to the chassis via an EGC. A TVSS is nothing more than a Voltage Clamp. It does nothing under normal operating conditions. The best ones use Silicon Avalanche Diodes and MOV's together. For a 240/120 Single Phase circuit they do nothing until the Voltage Potential reaches roughly 300 volts. At that point they Turn On and CLAMP the voltage, or Short Circuit until the voltage falls to a Release V, or they die trying. That limit the voltage to tolerable levels.

There is weak spot in your drawing. It is the Inverter being located remotely. The protection is at your home, not the distant Inverter, Now do not walk away thiong the Inverter is not protected. That is not the case. All I am saying is it would be better protected at the house. If it were me doing the design, I would run the PV DC Voltage to the house where the Inverter would be located. You can run the DC voltage up to 400 or 500 volts. Single phase is 240 volts, so at higher voltage means smaller wire and less expense. Plus the Inverter is better protected.

[bcroe]

There is weak spot in your drawing. It is the Inverter being located remotely. The protection is at your home, not the distant Inverter, Now do not walk away thiong the Inverter is not protected. That is not the case. All I am saying is it would be better protected at the house. If it were me doing the design, I would run the PV DC Voltage to the house where the Inverter would be located. You can run the DC voltage up to 400 or 500 volts. Single phase is 240 volts, so at higher voltage means smaller wire and less expense. Plus the Inverter is better protected.

And what do you think of a 6 gauge ground running from the inverters to the end of all panel metal
supports? This I regard as for personal safety against equipment faults, not for lightning.

There were some reasons here for putting inverters at the remote location. There was already a
heavy AC run out there, so it saved half the trenching and wire (at some cost in efficiency). And I
consider the inverters expendable, didn't want them in the house. Bruce Roe

[Sunking]

And what do you think of a 6 gauge ground running from the inverters to the end of all panel metal
supports?

Bruce stop and think about it. What good would it do?

The Inverter has built-in GFCI. If there is a fault anywhere, the Inverter shuts down. Think of all the older homes with 2-wire receptacles. If a electrician makes any modifications or adds, does not tear out all the wiring to bring everything up to code. No he or she would just install a GFCI receptacle or breaker in the panel. The code allows them to do that and is SAFE, in fact safer than a standard 3-wire circuit. A GFCI can prevent you from being electrocuted.

Keep in mind breakers and fuses do not protect you from electrical shock. That is not possible. It is there to protect the wiring and nothing else. Inverters like Solar Panels are not capable of burning up wiring under normal conditions. Think about that. Say you have a 6 Kw GT Inverter running 24/120 single phase. At that distance what size wire are you using? Heck with that just say code minimum and screw voltage drop and you used 10 AWG wire. The absolute highest current the Inverter and panels can deliver is 25 to 30 amps. Not enough to even blow its 30 amp fuse if it had one right? So what happens to a 10 AWG wire with 30 amps of current? Nothing right?

Fusing Current for copper 10 AWG (the amount of current to melt the wire) is 340 amps. Depending on what code you use, how it is used, 10 AWG is good up to 60 amps. So 30 amps of currrent 24 x 7 365 is no problem for the wire. No let's say you upsized for voltage drop to 6 AWG. Who cares, 10 AWG was over kill from a Safety POV.

See where I am going? So what purpose would it serve? If you wanted to do anything, is just make the Neutral wire the same size as the L1 and L2 wire. At the end of the day, the real answer is all you are doing is putting another conductor in parallel with the Neutral Conductor. Not needed or does anything for you.

[bcroe]

The Inverter has built-in GFCI. If there is a fault anywhere, the Inverter shuts down. Think of all the older homes with 2-wire receptacles. If a electrician makes any modifications or adds, does not tear out all the wiring to bring everything up to code. No he or she would just install a GFCI receptacle or breaker in the panel. The code allows them to do that and is SAFE, in fact safer than a standard 3-wire circuit. A GFCI can prevent you from being electrocuted.

See where I am going? So what purpose would it serve? If you wanted to do anything, is just make the Neutral wire the same size as the L1 and L2 wire. At the end of the day, the real answer is all you are doing is putting another conductor in parallel with the Neutral Conductor. Not needed or does anything for you.

Thanks, but I think you are talking about between the house main and the remote location. I am
talking about the bare ground wire from the most remote panel frame, back 200' to the combiner
box, and then back another 230' to the inverters. If the framework were not grounded, a shorted
panel/wire might put up to 400VDC on the frame. So that wire protects people, but how does it
fit into the lightning situation (when nobody will be out there)?

These Fronius inverters have the DC inputs isolated, except there is a (very expensive) 10A
between DC Neg and ground. The idea is, an array fault will complete the circuit and blow
the fuse, which the inverter can detect. However a higher impedance fault (conducting less
than 10A) will not blow the fuse, but certainly would be a problem for people. AC GFIs are
set for .005A trip. So I see the DC GFI function as strictly for protecting equipment and wire,
people depend on good grounding. Bruce Roe

[max2k]

Thanks, but I think you are talking about between the house main and the remote location. I am
talking about the bare ground wire from the most remote panel frame, back 200' to the combiner
box, and then back another 230' to the inverters. If the framework were not grounded, a shorted
panel/wire might put up to 400VDC on the frame. So that wire protects people, but how does it
fit into the lightning situation (when nobody will be out there)?

These Fronius inverters have the DC inputs isolated, except there is a (very expensive) 10A
between DC Neg and ground. The idea is, an array fault will complete the circuit and blow
the fuse, which the inverter can detect. However a higher impedance fault (conducting less
than 10A) will not blow the fuse, but certainly would be a problem for people. AC GFIs are
set for .005A trip. So I see the DC GFI function as strictly for protecting equipment and wire,
people depend on good grounding. Bruce Roe

I think you might be talking about 2 different type of inverters- SK is referring to transformerless ungrounded ones where both DC wires are floated and those can actually detect disbalance in the currents and disconnect from DC very quickly breaking possible path to anything. 400V DC on the frame is not a problem on its own unless someone touches it with one hand and touches the opposite wire with another. Old CRT TVs had -20,000V ... -30,000V DC on their chassis and nobody thought twice connecting antenna connector to it as 'another' wire was well isolated inside TV.

You're referring to the transformer based inverter which has its DC part completely isolated. It has DC'-' already connected to the frames so any positive wire faulting to that would just close the circuit. If the isolation breaks on a positive wire and person touches that with one hand and the frame with another he would be shocked and DC GFCI wouldn't help as the shocking current would simply reduce DC output of the array without disturbing its balance. I think GFCI on AC side of inverter won't work either as any fault to the ground won't create 'alternate' path from inverter's point of view and no current disbalance will occur to trip GFCI.

Now in case of lightning at Array site your grounding wire would be in parallel to DC '-' wire you already have there. Since it has much bigger gauge it would proportionally take over higher fraction of the current between sites. Whether it will help or not I don't know but I don't see why would it hurt.

[max2k]

Graphically you got it right. Only thing that looks out of place is that short green stub at the Inverrter that goes no where. Delete that and you got the right idea.

I was just trying to draw 'local' EGC - in case outgoing AC wire develops short to surrounding boxes/conduits/etc that current needs good path back to inverter. The green wires at the house and at the inverter are not connected which implies the conduit to run AC wires between them must be non metallic.

I do not know what you know about utilities and their regulations and standards. They are controlled by NESC which has nothing to do with NEC. If you look at any utility supply like the common single phase 240/120 almost all homes in the USA have is exactly what you are showing in your drawing, three circuit conductors of L1, L2, and N. There is no ground from the utility.

When it hits your Meter Box aka Service Disconnect location the N Conductor is boned to Earth. From that connection point you have a Green EGC and White Neutral conductor to your main breaker panel.

I have no clue and thank you for this information. I spent EE part of my career designing low level signal measurement circuits, never dealt with power lines and such. While currents/voltages are all the same I used to work with nano- prefixes, not kilo- .

OK you are over looking something but on the right track. What you are not seeing is the Neutral Wire between Inverter and you home is the Bonding Conductor.. Yes current will flow in that wire if there is a Strike, but it is only flowing between the Ground Electrodes, not if the equipment is wired properly. ...

I actually meant to ask if the thick enough (#6) wire running between sites would be enough to bring voltage difference during lightning event to the acceptable level. I was under impression it wouldn't but it sounds like it actually would. Let's say resistance to the ground at both sites is 25 Ohm but #6 over 300' is only 0.13 Ohm so both grounds should 'jump' more or less together the Inverter site slightly 'behind' at 1/200 of the voltage at Array site. Now with voltage to the 'calm' ground of 50,000V during this event Inverter site will be at 49,750V - 250V lower. There's also inductance of that 300' wire involved which would briefly prevent charge current going to that #6 but DC wires also have inductance so it would probably won't be a problem for inverter's inputs except the charge would create spike in voltage between DC'+' and array frames at Array site and panel's isolation wouldn't be able to sustain that. I think using TVSS at both outgoing and incoming circuits would improve this whole design to the point that #6 won't be even needed as DC '-' + TVSS would do the job protecting equipment. I fully agree bringing inverter to the house would make things better but that probably was not desirable.

[Sunking]

I actually meant to ask if the thick enough (#6) wire running between sites would be enough to bring voltage difference during lightning event to the acceptable level. I was under impression it wouldn't but it sounds like it actually would. Let's say resistance to the ground at both sites is 25 Ohm but #6 over 300' is only 0.13 Ohm so both grounds should 'jump' more or less together the Inverter site slightly 'behind' at 1/200 of the voltage at Array site. .

OK I can see where you are getting off track. The size of the conductor used as grounding electrode conductor need only to be sized to safely handle the currently likely imposed upon it. Lightning can induce very large amounts of current, but the duration is extremely fast. That limits heating. The dirt also acts as a heat sink and allows the conductor to handle more current.

But lets get down to what is important and real. I have been doing this for over 40 years. In those 40 years I have never sen or even heard of a Ground Rod with a 25 Ohms. Additionally that measurement is at power frequencies, not high frequencies of lightning. Once you factor that into the equation, the impedance would be in the Kilo Ohms. Second point what we are trying to do is limit Step Potential, the distance between your feet to safe levels. Dirt is a horrible conductor, and copper is a very good conductor. So if the distance between your feet on dirt is 1000 Ohms, and copper .001 Ohms, we have reduced Step Potential by a magnitude of 1,000,000 to 1. So instead of 10,000 volts, we reduce it to .01 volts.

Try this some time while you are out and about. If you get a chance take a close look at a Electrical Substation, a large one of 69 to 1000Kv. Look around the yard at the base of equipment and take notice of steel platforms. Those platforms are for the Switchman to stand on when they perform switching operations. This is done because there is a chance a circuit they are switching is Faulted to earth . The steel platform forms a equipotential ground plane. A short circuit between their feet, thus no difference in step potential. All those platforms are connected to a massive Ground grid made up of conductors and ground rods. All the equipment is bonded to that ground system.

[max2k]

OK I can see where you are getting off track. The size of the conductor used as grounding electrode conductor need only to be sized to safely handle the currently likely imposed upon it. Lightning can induce very large amounts of current, but the duration is extremely fast. That limits heating. The dirt also acts as a heat sink and allows the conductor to handle more current.

But lets get down to what is important and real. I have been doing this for over 40 years. In those 40 years I have never sen or even heard of a Ground Rod with a 25 Ohms. Additionally that measurement is at power frequencies, not high frequencies of lightning. Once you factor that into the equation, the impedance would be in the Kilo Ohms. Second point what we are trying to do is limit Step Potential, the distance between your feet to safe levels. Dirt is a horrible conductor, and copper is a very good conductor. So if the distance between your feet on dirt is 1000 Ohms, and copper .001 Ohms, we have reduced Step Potential by a magnitude of 1,000,000 to 1. So instead of 10,000 volts, we reduce it to .01 volts.

Try this some time while you are out and about. If you get a chance take a close look at a Electrical Substation, a large one of 69 to 1000Kv. Look around the yard at the base of equipment and take notice of steel platforms. Those platforms are for the Switchman to stand on when they perform switching operations. This is done because there is a chance a circuit they are switching is Faulted to earth . The steel platform forms a equipotential ground plane. A short circuit between their feet, thus no difference in step potential. All those platforms are connected to a massive Ground grid made up of conductors and ground rods. All the equipment is bonded to that ground system.

OK, #6 buried in the dirt changes the picture- I was considering case when #6 was isolated however strange that might look as this is pure academic discussion. In case of buried wire the voltage gradient will be close to 0 by the moment that wire reaches Inverter site. Unfortunately, this would change things at the Array site as it will be truly 50,000V above potential of DC+ wire since both ground and DC- went up there as a result of the event but DC+ would still be at 400V DC potential which is negligible compare to 50kV. I think only clamping this with TVSS would improve the odds of the inverter survival.

In terms of Step Potential I think your analogy is off here- without that buried wire the charge would spread itself around grounding electrode in all directions. Since dirt is bad conductor most of it now would actually go through our buried wire creating those dangerous conditions along its path. As analogy I suggest bunch of 1 kOhm resistors connected between grounding rod and some 'ideal' grounding plane underneath imitating dirt conductivity while our buried wire is connected from the same point to the same 'ideal' grounding plane 50 ft off. Since its resistance is much lower than all those 1kOhm resistors and they all under the same voltage most of the current would pass through our buried wire. That would mean 50kV are dropped on that distance creating gradient of 1kV/ft for very short time. Am I correct here? It would still be much much better than 'cone' type of 1kV/ft gradient surrounding grounding electrode without that buried wire. I have 0 experience with lightning protection, just trying to analyze this from common electrical point of view.

After some further thinking it seems my theory has a problem with initial assumption of lightning strike as 'voltage source' while in reality it is more like 'charge source'- once those charges pass through they're gone for good. In this case currents distribution would still be like I described but the voltage will be much reduced compare to the standalone ground electrode. When fixed amount of charge passes through some point over fixed amount of time it creates fixed current, not voltage. So, instead of 50,000 V at the strike point we would have 50,000 / 200 = 250V given the same charge and time both of which are defined by lightning strike itself and don't depend on our grounding efforts. The voltage gradient over the same 50 ft distance would be 5V/ft which is nothing. 200 here is how many times resistance to that ideal ground plane of the buried wire is lower compare to the standalone ground electrode. May be a little too optimistic.

[bcroe]

I think you might be talking about 2 different type of inverters- SK is referring to transformerless ungrounded ones where both DC wires are floated and those can actually detect disbalance in the currents and disconnect from DC very quickly breaking possible path to anything. 400V DC on the frame is not a problem on its own unless someone touches it with one hand and touches the opposite wire with another. Old CRT TVs had -20,000V ... -30,000V DC on their chassis and nobody thought twice connecting antenna connector to it as 'another' wire was well isolated inside TV.

You're referring to the transformer based inverter which has its DC part completely isolated. It has DC'-' already connected to the frames so any positive wire faulting to that would just close the circuit. If the isolation breaks on a positive wire and person touches that with one hand and the frame with another he would be shocked and DC GFCI wouldn't help as the shocking current would simply reduce DC output of the array without disturbing its balance. I think GFCI on AC side of inverter won't work either as any fault to the ground won't create 'alternate' path from inverter's point of view and no current disbalance will occur to trip GFCI.

Here we are again discussing "transformer" and "transformerless" inverters. I don't
think this has been well defined or understood here. But here is what I perceive.

In ancient electronics times, we raised DC voltages by chopping them up and driving a big heavy
iron transformer. We could wind the secondary for a high AC voltage, or add rectifiers to get high
voltage DC. But the AC was a square wave, not at all compatible with the PoCo, induction motors,
etc. There were some magnetic tricks (ferro resonant) that helped. This "transformer" equipment
was noted to be quite heavy, with all that iron and copper operating at 60 HZ.

When good power transistors came along, we could chop our DC at frequencies like 100 KHZ; the weight
of copper and iron decreased dramatically. With more intelligent control at those frequencies, we could
even piece together an output that emulated the PoCo line frequency. This method has been the standard
in solar for quite a long time, part of the reason solar has become a practical power source. I see these as
the "transformerless" inverters (even though they have 100KHZ transformers).

So I see a transformer inverter as having heavy magnetic components operating internally at the PoCo line
frequency. I see "transformerless" inverters as operating internally at several orders of magnitude higher
frequency than the power line frequency.

I don't see any transformer inverters in solar use here, maybe there never were any. I don't see ANY
connection between the above, and how the system is grounded. Both provide input-output isolation. If
some ancient type did prefer a certain grounding plan, I haven't seen it.

Disagree, not what you meant, lets hear it, but PLEASE don't be vague. Bruce Roe

[bcroe]

You're referring to the transformer based inverter which has its DC part completely isolated. It has DC'-' already connected to the frames so any positive wire faulting to that would just close the circuit. If the isolation breaks on a positive wire and person touches that with one hand and the frame with another he would be shocked and DC GFCI wouldn't help as the shocking current would simply reduce DC output of the array without disturbing its balance. I think GFCI on AC side of inverter won't work either as any fault to the ground won't create 'alternate' path from inverter's point of view and no current disbalance will occur to trip GFCI.

Now in case of lightning at Array site your grounding wire would be in parallel to DC '-' wire you already have there. Since it has much bigger gauge it would proportionally take over higher fraction of the current between sites. Whether it will help or not I don't know but I don't see why would it hurt.

"completely isolated" and "DC'-' already connected to the frames" are in direct opposition.

"400V DC on the frame is not a problem on its own unless someone touches it with one hand and touches the
opposite wire with another." In the industry, voltage #s with the reference not specified IMPLY that the reference
is ground. If a frame is at 400V, it CERTAINLY IS a hazard to any person (on the ground).

For the record, none of my wiring is in direct contact with the ground (except rods). Bare #6 is mounted on
aluminum frames. I am concluding what is in place here is optimum for personal; equipment protection
could be improved. Bruce Roe

[Sunking]

OK, #6 buried in the dirt changes the picture- I was considering case when #6 was isolated however strange that might look as this is pure academic discussion. In case of buried wire the voltage gradient will be close to 0 by the moment that wire reaches Inverter site. Unfortunately, this would change things at the Array site as it will be truly 50,000V above potential of DC+ wire since both ground and DC- went up there as a result of the event but DC+ would still be at 400V DC potential which is negligible compare to 50kV. I think only clamping this with TVSS would improve the odds of the inverter survival.

OK Max you are overlooking a very important point, and perhaps this will turn the light on. It does not matter if the voltage rises on Ground. In fact there is a name for it. Ground Potential Rise (GPR) is a phenomenon that occurs when large amounts of electric energy enter earth via dirt. If you watched the videos is what Mike is talking about and what causes Step Potential. Go google Earth Potential Rise aka Ground Potential Rise. With me so far?

If you bond everything at dirt level together, bond everything above dirt together, and then connect the two together at 1 point with 1 wire you have a single point ground. For current to flow there must be an entry point and an exit point.

So it does not matter if your reference point rises to 50,000 volts because everything else around rises 50,000 volts and thus no voltage difference is felt. What we are discussing is not my analogy, rather a well known and studied subject. Telephone companies, tower operators, military, CATV and any other operations use what is known as Single Point Ground. In fact the NEC is based on that very principle.

I use to do a lot of Mission Work with the church. We focused on health, water, and lightning protection to villages in area with frequent storm activity. Places like Cuba, Jamaica, India and other like tropical areas. These poor folks live in shacks with dirt floors. One village in India has several people every year killed by lighting while in their home. Lighting strikes by a nearby tree resulted in Step Potential Rise. As a group with my knowledge and background can put a stop to it very easily. All that has to be done is install a Ground Ring around their shacks. We used any kind of wre or metal we could come up with. Did not have to be copper. We used a lot of telephone cable, mattress spring wire, steel cable, or any scrap we could come up with. The Ground Ring Shunts the current around the schack. After we leave. no one is ever injured again.

[max2k]

"completely isolated" and "DC'-' already connected to the frames" are in direct opposition.

"400V DC on the frame is not a problem on its own unless someone touches it with one hand and touches the
opposite wire with another." In the industry, voltage #s with the reference not specified IMPLY that the reference
is ground. If a frame is at 400V, it CERTAINLY IS a hazard to any person (on the ground).

For the record, none of my wiring is in direct contact with the ground (except rods). Bare #6 is mounted on
aluminum frames. I am concluding what is in place here is optimum for personal; equipment protection
could be improved. Bruce Roe

I meant AC and DC parts completely isolated from each other. In such setups DC- is typically grounded so you get 400V DC on the frames 'for free' and as I said it's not a problem by itself.

It is my understanding modern transformerless inverters don't provide AC - DC isolation and require both DC wires to be floated. Fault to the ground of either of those wires will be immediately detected by inverter and it will stop feeding AC disconnecting DC from AC part. At that point it will show error so one of us could go and fix it. Faulted DC on the frame at that moment is still not a problem or dangerous as it will be pretty much like transformer based setup.

You're probably correct that more precise term would be isolating / non-isolating inverters as it has profound consequences to the grounding. Non isolating inverters can be treated just like any other AC equipment with only EGC requirement while isolating ones require GEC to tie up its isolated AC output to house AC circuit.

It sounds you have non-isolating inverter and then what SK said in #36 fully applies. OTOH that 10A part to the ground and 'completing circuit' in your #37 contradicts it- non isolating inverters switch DC inputs between L1-N-L2 directly if I understand it correctly so the DC input is completely symmetrical and floating. Fault to the ground should make it stop switching but no breakers will trip although I didn't try this.