Ungrounded PV

Hi All,

In an ungrounded/floating PV system, what components are different from grounded systems (besides lack of bonding wire)?

Do string measurements (voltage/amperage) and/or wire calcs change? How about I-V curve testing?

Thanks!

jw
«13

Comments

  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Ungrounded PV

    Welcome to the forum.

    Functionally the two systems would be the same. The only time grounding actually does anything is if something goes wrong. Otherwise ground conductors carry no current.

    Someone is bound to point out that the neutral-ground bond in a AC system sinks one leg of the AC output to zero Volts in reference to ground making it "safe to touch" and that the same can be said for grounding the negative (or positive in some cases) on the DC side. So it might as well be me. :D

    But otherwise all Voltages and currents and the calculations thereof are the same.
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Ungrounded PV
    PVJW wrote: »
    Hi All,

    In an ungrounded/floating PV system, what components are different from grounded systems (besides lack of bonding wire)?

    Do string measurements (voltage/amperage) and/or wire calcs change? How about I-V curve testing?

    Thanks!

    jw

    None of the measurements really change, but the measurements can change the wiring. Under the NEC, the way you must wire a system may be different depending on whether the system voltage is greater than 50 volts. So if you are using a 450 volt panel string into a GTI inverter, you may not have the option of running the panel circuit ungrounded. While with 18 volt panels charging 12 volt batteries through a PWM controller you may.

    Since no current flows through the grounding conductor under normal circumstances, there will be no effect on your voltage drop versus current calculations. But whenever you measure voltages, you must connect one meter lead to the - wire rather than to ground.

    There will be absolutely no difference in I-V curve testing, except that if the I-V tester has one terminal grounded, you must isolate the panel from any conflicting ground connection.
    SMA SB 3000, old BP panels.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV

    The differences--A floating system should use two ganged circuit breakers--So if one trips, the other turns off too--Just like the 240 VAC two pole breakers used in North American 120/240 VAC panels.

    If you have a "grounded" neutral, the neutral/white wire cannot become hot, therefore does not need a breaker to prevent short circuit current overload.

    Other issues... A grounded system prevents static electric charging (i.e., an antenna/tall tower/solar array on a tower) can become charged by the natural earth's static electric field (around 100-300 volts per meter) or become electrified just before a lightning strike with thousands of volts per meter (obviously, a direct strike does even more interesting things).

    Improper grounding/connecting of DC power can energize water/gas pipes and cause corrosion.

    Regarding solar panels--The (German) Sunpower Solar Panels require positive grounding or they build up an internal charge that reduces output voltage (significantly). This is temporary and can be reversed in a few hours (days?) if reconnected as positive ground.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Ungrounded PV
    BB. wrote: »

    Regarding solar panels--The (German) Sunpower Solar Panels require positive grounding or they build up an internal charge that reduces output voltage (significantly). This is temporary and can be reversed in a few hours (days?) if reconnected as positive ground.

    -Bill

    What happens if you try to use Sunpower panels in a series string to an MPPT CC or GTI? Do you just ground the + of the whole string?
    Or do you have to use them with compatible micro-inverters?
    SMA SB 3000, old BP panels.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Ungrounded PV

    Bill, I've just got to take a little side step here and ask the rhetorical question: how well do those positive-ground panels work with the NEC's new DC GFCI regs? :p
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV

    With standard solar power charge controllers--They cannot run an isolated ground for the array vs battery bank. So you have the choice to run the panels as positive ground and the battery bank too (done all the time with Telecom based systems)--But be careful, there are lots of assumptions made regarding "signal" and power grounds--It can cause lots of issues with communications gear (i.e., RS 232 assume "negative ground"--If your charge controller is positive ground and your terminal/laptop assumes negative ground--you can easily smoke the cabling or a mother board trace).

    If you choose to run as negative ground, you make get something like 20-40% or so reduction in output power (as I recall--But it has been years and the original documentation is gone from the web).

    For a low voltage string and/or possibly the effect is not as bad as written about (Tony/Icarus has a couple of the 90 Watt panels, as I recall, and has been very happy with them in his "negative grounded (I think)" system).

    Regarding the NEC--They do not care about negative or positive ground--The ground fault detection system does not care--The sense fuse/breaker is not polarity sensitive (as far as I can see--but could be wired in reverse if needed).

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Ungrounded PV
    Hi All,

    In an ungrounded/floating PV system, what components are different from grounded systems (besides lack of bonding wire)?

    Do string measurements (voltage/amperage) and/or wire calcs change? How about I-V curve testing?

    Thanks!

    jw

    If you have enough strings to require string fusing (3 or more), you must fuse both the positive and negative conductors. If you have a DC disconnect, you must break both positive and negative conductors.

    EGC grounding (module frames, racking, etc.) is the same as for a grounded system; when you say "lack of bonding wire", if it's the wire connecting the lugs on the modules and racking to ground you are referring to, that is incorrect. When a PV system is said to be ungrounded, that refers to the absence of a link through a GFI circuit between one of the DC conductors (usually the negative) and ground; it DOES NOT mean that you do not need to ground all exposed metal that may be energized if there is a fault.

    String voltage and current measurements are unaffected. Wire sizing calcs are unaffected. I-V curve testing? I'm not sure what you mean, but I-V characteristics of modules do not change.

    Inverters are either for grounded or ungrounded DC and you cannot change them. They are all grounded on the AC side*.

    * assuming a grid tied system. I don't know all that much about off grid inverters.
  • PVJW
    PVJW Registered Users Posts: 3
    Re: Ungrounded PV

    Thanks, All, for the information, and happy to find this forum.

    For larger systems, are there fusible combiner boxes that will break both positive and negative lines? Any recommendations?

    Same question for DC disconnects... I've seen a spec sheet for Siemens 3-pole discos that break positive through 2 poles and negative through the third. Don't look different from normal 3-poles. At the risk of changing the topic--why break positives twice in DC disconnects? I see this a lot, but don't fully understand the reasoning behind it. Just providing redundant protection, or is this good practice to always follow (even if it requires additional disconnects?)?

    Affirmative on grounding the metallic parts, just not connecting positive or negative to the ground through the inverter. I-V curve testing was referring to string testing, rather than module, but I wasn't considering that both leads would be fused--I was wondering about the effects of I-V string testing with negatives bussed in the combiner (but still showing negative voltage to ground). With negative fusing, this wouldn't be an issue during testing, as long as both positive and negative fuses were opened.

    jw
  • jaggedben
    jaggedben Solar Expert Posts: 230 ✭✭
    Re: Ungrounded PV

    The NEC has some specific requirements for ungrounded systems (690.35). Most have been alluded to already, but one that hasn't (unless I missed it) is that ungrounded systems require PV wire instead of USE-2.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV
    PVJW wrote: »
    For larger systems, are there fusible combiner boxes that will break both positive and negative lines? Any recommendations?

    I have not seen any... And, while I understand the intent, I am not sure that even NEC requires double pole breakers in this application.

    More or less, solar power systems are "floating output" devices, unless the system has been "hard grounded" between "DC Return" and an Earth/Frame ground connection (and there is the DC GFI in NEC--A whole different discussion). So, a floating output to be "maximum safe" should have both wires switched/protected in the case of a short circuit so that the circuit is completely de-energized if a breaker pops. There are not many off the self fuses that can open both poles like a ganged breaker can.

    From a single point fault point of view--A single pole fuse/breaker is enough to stop a shorted panel from being feed "too much current" from two or more other strings in parallel.
    Same question for DC disconnects... I've seen a spec sheet for Siemens 3-pole discos that break positive through 2 poles and negative through the third. Don't look different from normal 3-poles. At the risk of changing the topic--why break positives twice in DC disconnects? I see this a lot, but don't fully understand the reasoning behind it. Just providing redundant protection, or is this good practice to always follow (even if it requires additional disconnects?)?

    Do yo have a link for that product?

    My guess---DC is very difficult to "break" vs AC... DC tends to sustain arcs very nicely in lower voltage DC circuits (DC Arc Welders are much nicer than AC arc welders). Having two breakers could give better maximum interrupt current ratings... Also, some breakers are polarity sensitive (contact shape, arc control orientation, use a permanent magnet to "blow out" a DC arc, etc.)--So putting two breakers "back to back" could support a DC current interrupt in either direction of current flow (common with a battery bank where they are charged and discharged through the same wiring for some equipment).
    Affirmative on grounding the metallic parts, just not connecting positive or negative to the ground through the inverter. I-V curve testing was referring to string testing, rather than module, but I wasn't considering that both leads would be fused--I was wondering about the effects of I-V string testing with negatives bussed in the combiner (but still showing negative voltage to ground). With negative fusing, this wouldn't be an issue during testing, as long as both positive and negative fuses were opened.

    Still should not matter... Fuses and Solar panels do not play together well like batteries and fuses do...

    Batteries, when shorted, can output 10-1,000 of times the nominal amperage--Very easy to trip a fuse or breaker.

    For solar panels which are, for the most part, current sources... A panel under full power loading may output Imp... With a short circuit it will output ~1.25*Imp (Isc)... Not enough to (reliably) trip a fuse or breaker. So, if you have a fault to ground (or return), they can draw pretty close to Imp, not trip a fuse, and eventually create a sustained arc--Which can start a fire/cause other damage/dangers.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • jaggedben
    jaggedben Solar Expert Posts: 230 ✭✭
    Re: Ungrounded PV

    About positive grounding of Sunpower panels, some clarifications:

    -With GT-only inverters, it's very simple. Bring the positive conductors to the grounded section of any combiner or disconnect. Then connect the grounded output to the positive input of the inverter. With SMA, for example, it's simply a matter of getting the polarity correct when you connect the integrated combiner/disco to the inverter.
    -Sunpower panels cannot be used in ungrounded systems.
    -Sunpower is based in Richmond California. They are not German.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV

    Sorry about calling Sunpower a German company--Must have been one of my Google Searches a few years ago sent me to a German version of their website.

    "Sun Power" appears to be a commonly used name and (from my reading) did not alway mean "Sunpower Corp. USA"--And I want to make sure we are talking about the correct company.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • jaggedben
    jaggedben Solar Expert Posts: 230 ✭✭
    Re: Ungrounded PV
    BB. wrote: »
    I have not seen any... And, while I understand the intent, I am not sure that even NEC requires double pole breakers in this application.

    That is correct. The NEC simply requires overcurrent protection on both sides. This can be through independent fuses.
  • jaggedben
    jaggedben Solar Expert Posts: 230 ✭✭
    Re: Ungrounded PV
    PVJW wrote: »
    Same question for DC disconnects... I've seen a spec sheet for Siemens 3-pole discos that break positive through 2 poles and negative through the third. Don't look different from normal 3-poles. At the risk of changing the topic--why break positives twice in DC disconnects? I see this a lot, but don't fully understand the reasoning behind it. Just providing redundant protection, or is this good practice to always follow (even if it requires additional disconnects?)?

    Take the following with a grain of salt, because it has been a while since it was explained to me and I may have significant aspects backwards or otherwise incorrect. With that said...

    When you put two poles of a disconnect in series you can add together the voltage rating of each pole. So if you have a 2-pole 300VDC, 30A rated disconnect, you can turn it into a 600VDC, 30A rated disconnect by wiring two of the poles in series.

    I might have it backwards on the voltage rating vs the amperage rating, but voltage makes sense to me because this is in series.

    Bill is correct that DC is 'harder to break' so a disconnect will often have a lower DC rating than AC rating, if it is rated for DC at all. For PV output circuits you need a disco that is DC rated.
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Ungrounded PV
    jaggedben wrote: »
    The NEC has some specific requirements for ungrounded systems (690.35). Most have been alluded to already, but one that hasn't (unless I missed it) is that ungrounded systems require PV wire instead of USE-2.
    Good catch; I forgot about that one.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV
    jaggedben wrote: »
    That is correct. The NEC simply requires overcurrent protection on both sides. This can be through independent fuses.

    Which would be true and correct (safety wise) for a floating PV system...

    Most fixed installations are going to DC ground the battery bank to a ground rod/cold water pipe/building safety ground system. And most vehicle mounts are going to DC ground to the vehicle frame ground.

    At that point, the neutral is "bonded" to safety ground and would no longer require a DC fuse/breaker on both sides of the panel.

    Which does bring us back to the NEC DC GFI system (~1 amp fuse between safety and earth ground to "detect" ground fault current flow)... The system uses a fuse or breaker to "trip/open" if there is >1 amp of current flow through the "sense" fuse/breaker. That effectively lifts the (typically) negative system ground of the DC battery bus and effectively turns the whole system into a positive ground referenced electrical plant...

    That violates the whole idea of safety/neutral bonding and only using single pole over current devices on the "hot side" of your power system (AC or DC--Hot vs Neutral with respect to earth ground is all I care about here).

    One of the reasons I believe that the whole DC GFI system as currently implemented by code requirement is absolutely a disaster (safety and system design wise).

    To quote myself from the Midnite forum thread about DC GFI:
    One other quick way to look at the DC GFP setup... Imagine you designed a Negative Ground system and put fuses/breakers on all positive wiring. No fuses/breakers on negative wiring, and one positive solid connection between the negative bus and earth ground (including grounding to water pipe, fixtures, wiring boxes, conduit).

    If there is a fault from anywhere in the Positive circuit path, positive wire will be protected by an appropriate fuse/breaker which will be tripped by the excessive current flow.

    If there is any negative/return wire to earth ground, virtually nothing will happen as both are very near the same voltage. (it is possible with ground loops to overheat return wiring, but for now, NEC does not address over current from ground loops other than recommending/requiring single point grounding).

    Now, you replace that one Earth Ground with a 1-5 amp fuse. And now earth ground, randomly, anywhere in the Positive Wiring Path (PV wiring or Battery wiring). Pops the 1-5 amp fuse, and stops power flowing from solar panels to/through charge controller... Now you have a positive earth grounded system with fuses in the earthed positive paths.

    No current limits on Negative Grounds. Items that were assumed to be "touch safe" (like lamp sockets) are now powered by battery or PV Array voltage/power. And the current limit available anywhere on the negative/return wiring is dependent on where, specifically, the Positive to Earth Fault was made.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Ungrounded PV
    PVJW wrote: »

    For larger systems, are there fusible combiner boxes that will break both positive and negative lines? Any recommendations?

    Yes, for some values of "larger", at least. Midnight Solar makes combiners that can fuse 12 circuits grounded or 6 circuits ungrounded, depending on which busbars you specify. SolarBOS makes dual output combiners as well.
  • vtmaps
    vtmaps Solar Expert Posts: 3,741 ✭✭✭✭
    Re: Ungrounded PV
    jaggedben wrote: »
    When you put two poles of a disconnect in series you can add together the voltage rating of each pole. So if you have a 2-pole 300VDC, 30A rated disconnect, you can turn it into a 600VDC, 30A rated disconnect by wiring two of the poles in series.

    That's correct. I believe that Midnite's 300 volt breakers are two ganged 150 volt breakers. --vtMaps
    4 X 235watt Samsung, Midnite ePanel, Outback VFX3524 FM60 & mate, 4 Interstate L16, trimetric, Honda eu2000i
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Ungrounded PV
    Now, you replace that one Earth Ground with a 1-5 amp fuse. And now earth ground, randomly, anywhere in the Positive Wiring Path (PV wiring or Battery wiring). Pops the 1-5 amp fuse, and stops power flowing from solar panels to/through charge controller... Now you have a positive earth grounded system with fuses in the earthed positive paths.

    No current limits on Negative Grounds. Items that were assumed to be "touch safe" (like lamp sockets) are now powered by battery or PV Array voltage/power. And the current limit available anywhere on the negative/return wiring is dependent on where, specifically, the Positive to Earth Fault was made.

    Amen! Now look at the system used by Midnight, with the common trip unit of 1.5 amp breaker and 60+ amp breaker. Now the presence of current to ground (from a ground high enough up on the string to deliver that much current) will trip both breakers, thereby isolating the no-longer floating negative from ground and also interrupting the positive lead connection to the CC or inverter. You now have a disconnected panel system except for the negative still being connected to the CC input. And nothing whatsoever has been done for the safety of the panel array except that you now know that something is wrong.

    What if you used a ganged triple breaker which sensed ground current and then opened both the + and - leads? (Basically, I just do not see how to properly use a fuse as a GDI. I do not have so much of a problem with requiring a Ground Detector or a GDI, but rather with accepting a simple fuse as a way to meet the intention of that requirement.!)

    Now the use of a Ground Detector for ungrounded systems is another animal entirely, but that has its owns set of design considerations to argue about.
    SMA SB 3000, old BP panels.
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Ungrounded PV
    vtmaps wrote: »
    That's correct. I believe that Midnite's 300 volt breakers are two ganged 150 volt breakers. --vtMaps

    I have a concern about that approach when used uncritically. The two 150 volt breakers in series will, when open, withstand 300 volts lead to lead based on the 150 volts contact to contact in each section, at least assuming that there is a low current leakage path which can equalize the middle contact set to halfway between the two lead voltages.

    But I am not nearly as willing to believe that the DC interrupting capability of the two add up to make the unit capable of interrupting 300 volts. Comments?
    SMA SB 3000, old BP panels.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Ungrounded PV
    inetdog wrote: »
    I have a concern about that approach when used uncritically. The two 150 volt breakers in series will, when open, withstand 300 volts lead to lead based on the 150 volts contact to contact in each section, at least assuming that there is a low current leakage path which can equalize the middle contact set to halfway between the two lead voltages.

    But I am not nearly as willing to believe that the DC interrupting capability of the two add up to make the unit capable of interrupting 300 volts. Comments?

    The theory, which is sound, is that the simultaneous interruption of the two breakers (one on negative, one on positive) creates a circuit interruption capable of handling 2X the Voltage. Voltage handling, in this case, being a matter of creating a gap large enough that the Voltage can not arc across it and doing so fast enough so that any arcing that occurs in the process does not last long enough to do any appreciable damage.

    Remember that this is for use as a breaker, not a switch. The difference being how often the device is expected to disconnect the power. A switch is 'continuous operation'; on and off whenever. A breaker is 'occasional operation'; mostly connected but capable of breaking the circuit if needed.

    In that respect it should work fine.
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Ungrounded PV
    inetdog wrote: »
    Amen! Now look at the system used by Midnight, with the common trip unit of 1.5 amp breaker and 60+ amp breaker. Now the presence of current to ground (from a ground high enough up on the string to deliver that much current) will trip both breakers, thereby isolating the no-longer floating negative from ground and also interrupting the positive lead connection to the CC or inverter. You now have a disconnected panel system except for the negative still being connected to the CC input. And nothing whatsoever has been done for the safety of the panel array except that you now know that something is wrong.

    What if you used a ganged triple breaker which sensed ground current and then opened both the + and - leads? (Basically, I just do not see how to properly use a fuse as a GDI. I do not have so much of a problem with requiring a Ground Detector or a GDI, but rather with accepting a simple fuse as a way to meet the intention of that requirement.!)

    Now the use of a Ground Detector for ungrounded systems is another animal entirely, but that has its owns set of design considerations to argue about.
    I'm not sure that this is relevant, but most (maybe all) inverters for ungrounded PV systems detect ground faults with differential current sensors on the positive and negative conductors. A mismatch between them above a set threshold sets an alarm flag and shuts down the inverter.
  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Ungrounded PV

    The system inetdog is describing there is for DC GFCI. Somewhere on this forum is quite an in-depth discussion of this requirement, including commentary from knowledgeable folk about its shortcomings.

    BTW, you left out mention of the resistor that connects (-) to GRND when the breaker is tripped to keep it "safe".

    And this is not to be confused with MidNite's arc-fault protection, which is probably of more value for a PV array than ground-fault.

    (There should be a bad pun here somewhere about having a "fault-sense of security" but I'm not saying it.) :p
  • inetdog
    inetdog Solar Expert Posts: 3,123 ✭✭✭✭
    Re: Ungrounded PV

    (There should be a bad pun here somewhere about having a "fault-sense of security" but I'm not saying it.) :p
    But I will hold you responsible for it anyway. Enough (un)said.
    SMA SB 3000, old BP panels.
  • PVJW
    PVJW Registered Users Posts: 3
    Re: Ungrounded PV

    Re: Combiner boxes with capacity to fuse both positive and negative home runs...

    Based on the suggestions, I found the SolarBOS (http://www.solarbos.com/data/files/15/SolarBOS%20Multi-Combiners.pdf) combiner that is designed for supplying two inverters--I assume this would also serve to isolate, fuse, and buss both positives and negatives for an ungrounded system. Am I wrong about this?

    Also found the Midnite Solar combiner (http://www.midnitesolar.com/pdfs/frontBack12.pdf) that looks as though it would work for exactly this type of application, though limited to 5 strings.

    Re: DC disconnect for fusing positive and negative feeders

    I had said Siemens, but it was actually an Eaton disconnect (http://downloads.eatoncanada.ca/downloads/Safety%20Switches/02%20-%20Product%20Guides/B35J01SK%20-%20600VDC%20Solar%20Disconnects%20Canadian%20Product%20Guide.pdf). Note this is intended for the Canadian market...

    jw
  • jaggedben
    jaggedben Solar Expert Posts: 230 ✭✭
    Re: Ungrounded PV
    BB. wrote: »
    Most fixed installations are going to DC ground the battery bank to a ground rod/cold water pipe/building safety ground system. And most vehicle mounts are going to DC ground to the vehicle frame ground.

    It's true that most systems are grounded, but the NEC doesn't require that. It does seem that there are few products designed for ungrounded stand-alone operation with batteries. For grid-tie-only applications, this is much less true. (The NEC does not apply to vehicles.)
    At that point, the neutral is "bonded" to safety ground and would no longer require a DC fuse/breaker on both sides of the panel.

    Correct, although if you'll forgive a couple nitpicks, it's not a 'neutral', and the DC fuse/breaker would actually be prohibited.
    Which does bring us back to the NEC DC GFI system (~1 amp fuse between safety and earth ground to "detect" ground fault current flow)... The system uses a fuse or breaker to "trip/open" if there is >1 amp of current flow through the "sense" fuse/breaker.

    What you are calling the 'NEC DC GFI system' is not actually required by the NEC. The NEC simply requires ground-fault detection and interruption. The fuse method is the most common method and requires a grounded conductor, but that is not the only method. Inverters designed for ungrounded applications use a different method.
    That effectively lifts the (typically) negative system ground of the DC battery bus and effectively turns the whole system into a positive ground referenced electrical plant...

    It usually floats the array. It should not lift the grounding of the batteries; that is not required by the NEC. (Are there products that do that? They would be mis-designed.) As for becoming 'positive ground referenced' that is of course only true if the ground fault is on the positive PV conductor. But if the array is floated, then so what?
    That violates the whole idea of safety/neutral bonding and only using single pole over current devices on the "hot side" of your power system (AC or DC--Hot vs Neutral with respect to earth ground is all I care about here).

    I don't follow. Since the NEC requires the array power to be disconnected, there should be no danger on the load side of equipment. A floated array may be slightly more dangerous to someone fixing the problem on the array, but presumably they will be aware of that.
    One of the reasons I believe that the whole DC GFI system as currently implemented by code requirement is absolutely a disaster (safety and system design wise).

    I don't think I follow. I would tend to agree that the 1amp fuse system is not the best we can do, especially for larger systems. But again, that method is not actually required by the NEC. I'm not sure what makes it a 'disaster.' Again, only the PV array gets floated, everything else can remain grounded.

    To go into your posting on the midnitesolar forum...
    If there is a fault from anywhere in the Positive circuit path, positive wire will be protected by an appropriate fuse/breaker which will be tripped by the excessive current flow.

    Well, that is not correct for every fault involving the PV source and output circuits. PV source circuits are inherently power limited, and as you know we fuse them at greater than their highest possible current. We fuse them to protect the output of multiple strings from overloading the conductors for a single string. That is the only kind of fault that will trip a PV circuit fuse or breaker. For the others we need a ground-fault detection system or we have no way of knowing about it, and no way of stopping ground paths from being overloaded.
    If there is any negative/return wire to earth ground, virtually nothing will happen as both are very near the same voltage. (it is possible with ground loops to overheat return wiring, but for now, NEC does not address over current from ground loops other than recommending/requiring single point grounding).

    A ground return path could easily be overloaded given certain conditions. That is why the NEC requires interruption of the PV power supply when a ground fault is detected.
    Now, you replace that one Earth Ground with a 1-5 amp fuse. And now earth ground, randomly, anywhere in the Positive Wiring Path (PV wiring or Battery wiring). Pops the 1-5 amp fuse, and stops power flowing from solar panels to/through charge controller... Now you have a positive earth grounded system with fuses in the earthed positive paths. No current limits on Negative Grounds. Items that were assumed to be "touch safe" (like lamp sockets) are now powered by battery or PV Array voltage/power.

    If the negative PV side is interrupted by the GFDI as required by the NEC, then how is current going to return to the PV negative, exactly? I don't where see the safety problem is, or how load devices that are touch-safe become not so.
  • jaggedben
    jaggedben Solar Expert Posts: 230 ✭✭
    Re: Ungrounded PV
    PVJW wrote: »
    Re: Combiner boxes with capacity to fuse both positive and negative home runs...

    I'm pretty sure SMA also makes one.
  • ggunn
    ggunn Solar Expert Posts: 1,973 ✭✭✭
    Re: Ungrounded PV
    PVJW wrote: »
    Re: Combiner boxes with capacity to fuse both positive and negative home runs...

    Based on the suggestions, I found the SolarBOS (http://www.solarbos.com/data/files/15/SolarBOS%20Multi-Combiners.pdf) combiner that is designed for supplying two inverters--I assume this would also serve to isolate, fuse, and buss both positives and negatives for an ungrounded system. Am I wrong about this?
    No, you are correct. The SolarBOS dual output combiners can be used in this manner.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV
    jaggedben wrote: »
    ...
    It usually floats the array. It should not lift the grounding of the batteries; that is not required by the NEC. (Are there products that do that? They would be mis-designed.) As for becoming 'positive ground referenced' that is of course only true if the ground fault is on the positive PV conductor. But if the array is floated, then so what?

    Because all of the "typical" DC GFI systems I have seen use a ~1-5 amp fuse or breaker between "return" and earth/safety ground--The very act of "detection" removes the ground bond (fuse/breaker opening).

    The NEC only requires the system to stop current flow (originally was designed for "arc fault" protection) and the standard devices are either only 1 pole breakers (lift "hot") or just simply stop the charge/GT controller (modern GT inverters and some MPPT Solar Charge controllers) from taking energy from the array (remember, designed to stop arc faults, not to provide any other safety improvements).

    With the small fuse/breaker between return and safety ground, the opening of the fuse does, by itself, stop the current flow in the fault path (with a very unsafe method--leaves everything "hot and ungrounded").
    I don't follow. Since the NEC requires the array power to be disconnected, there should be no danger on the load side of equipment. A floated array may be slightly more dangerous to someone fixing the problem on the array, but presumably they will be aware of that.

    With system grounding, we have only home ground bond per system to prevent ground loops. If we look at battery based systems, there is only one ground bond (DC GFI Fuse) for both the Solar array side, and the Battery Bus side. So, the fuse is not tripped by only a Hot to earth fault on the array side, any hot to earth fault in the system will pop the GFI fuse too (Battery + hot to earth, MSW AC output to earth, etc.).

    This causes the "hot" to be earth referenced (i.e., battery + bus to earth fault) with the former "ground referenced" return now not--With the current potential of where the Hot to Earth fault occurred (as an engineer, I reserve the right to pick the worse place fault location--I.e., battery bank... And not a 2 amp LED branch circuit).

    Since a grounded return system should not have fusing in the return (neutral) wiring, there are no controls on 1/2 of the wiring in the power system.
    I don't think I follow. I would tend to agree that the 1amp fuse system is not the best we can do, especially for larger systems. But again, that method is not actually required by the NEC. I'm not sure what makes it a 'disaster.' Again, only the PV array gets floated, everything else can remain grounded.

    Again, the rest of a DC power system does not have galvanic isolation with respect to the solar array (this would require an "isolation transformer" in the Charge Controller). The DC return/"ground" is carried through the entire system. So lifting the ground bond for the solar array, also lifts the ground bond for the rest of the system. The this now makes the entire "grounded neutral/return" hot with respect to earth--and no fusing/breakers/double pole ganged breakers to limit current flow in the return wiring.

    With regard to GT inverters--The maximum working voltage is limited to 600 VAC... If you run two arrays through a common conduit/junction box, a "ground reversal" can cause the + lead of one array to be at +600 volts and the - lead of the second array to be at -600 VDC -- That is a 1,200 VDC potential between two wires in the same conduit--Appears to be a violation of the 600 Volt maximum differential allowed for "standard" voltage wiring (I asked--I could not get a reply to that violation either).
    To go into your posting on the midnitesolar forum...

    Well, that is not correct for every fault involving the PV source and output circuits. PV source circuits are inherently power limited, and as you know we fuse them at greater than their highest possible current. We fuse them to protect the output of multiple strings from overloading the conductors for a single string. That is the only kind of fault that will trip a PV circuit fuse or breaker. For the others we need a ground-fault detection system or we have no way of knowing about it, and no way of stopping ground paths from being overloaded.

    The ground path will never be overloaded anyway (if you design your grounding system correctly--The ground needs to be able to "safely" carry the fault current until the protective device trips--A ground bond for a battery bank with a 400+ amp fuse needs to be much heavier than that of a 100 amp AC main service). So DC GFI's do not address that issue other than by placing a "1 amp fusible link between return and safety ground" (in only all my other power system designs could have been fixed that easily).

    Plus, since PV Hot to earth fault can be anywhere on a DC system (no isolation), then the entire non-isolated part of the system needs to be addressed "as a whole"... Not just the solar array, not just the battery bus, not the battery bus return, not just the output of the MSW inverter, not just the RS 232/422/ethernet/etc. com lines, not just the random control relay outputs, not just the battery shunt leads, not just the random voltage sense leads, etc.
    A ground return path could easily be overloaded given certain conditions. That is why the NEC requires interruption of the PV power supply when a ground fault is detected.

    By placing a 1 amp fuse between return and safety ground--We have now guaranteed that a shorted D cell flashlight can pop the safety ground.... This would not have passed any ground test that I ever had to pass with my commercial equipment designs that went through UL/NRTL Registration.
    If the negative PV side is interrupted by the GFDI as required by the NEC, then how is current going to return to the PV negative, exactly?

    The only reason that this DC GFI system even "works" is because the fuse/breaker (fuse/breaker which must be rated for >600 volt for the application) stops the current flow. The "ganged breaker", inhibiting the MPPT Charge Controller, GT Inverter turning off, etc. does nothing to limit the faulting current flow.

    For example, replace the Fuse/Breaker with a magic current detector (AC and DC fault current detector). If that link does not blow--The current will still flow through the fault path (i.e., turning off the charge controller/GT inverter did not stop the current flow).

    When working with a power system--you never "assume" that leads are not energized--Not even the white/neutral or the ground bonded DC return (I have been shocked by a neutral AC power connection because of different ground potentials in a large salt water aquarium installation).

    But, the whole idea that ground bonded neutrals can never be energized is the 1/2 the basis of safety in the NEC.

    Imagine you are working on a home AC power system and the "hot wire" became ground, and the neutrals became hot--How would you react (simplest example is the screw-in lamp--the threaded socket is supposed to be neutral to prevent electrocution--that neutral would now be hot).
    I don't where see the safety problem is, or how load devices that are touch-safe become not so.

    Since my concern is DC GFI (AC GFI is very nice managed with GFI breakers/outlets--very easy to measure 5 milliamps of leakage current with a current transformer--DC--not so easy), let me ask you...

    If the DC return of your power system became hot (say a battery + to earth fault)... What would happen with your RS 232 communications signal ground (I have had RS 232 cables and video terminal PC boards smoke when RS 232 signal ground "went hot"--Many times, Communications Grounds are referenced to Safety Ground, other times they are referenced to DC Power Ground--What if those ground potentials "switch" at random times).

    What if you had neutral (bus negative) short to your LED lamp... The only fuse in the system is the 400 amp fuse to the battery bank (if there is a fuse at all in many installations). Battery + (12-48 volts) is now at earth ground (cable short, wrench/bolt dropped between battery + and a grounded metal battery rack). That negative lead now has 400+ amps of battery bus current available and no fuse to limit current to rating of the wire.

    Or, if you have a MSW inverter that gets one of its leads shorted to safety/earth ground. Easily pops the 1 amp fuse and now your entire battery +/- bus has a modified square wave of dozens to hundreds of volts with respect to earth ground. So, the popped fuse/breaker stopped the charge controller from charging or opened the PV+ lead--nothing changes on the DC Battery side.

    I do not see how a 1 amp fuse between a system designed with a "bonded" neutral to earth connection can still be "safe" when that bond (originally designed to pass sufficient fault current to pop the upstream breaker--usually rated to at least several hundred amps of fusing current) is replaced with a 1 amp fuse.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • BB.
    BB. Super Moderators, Administrators Posts: 33,431 admin
    Re: Ungrounded PV

    To add one more... To me, it is kind of interesting that "they" require a 600 or 1,000 volt rated fuse/breaker between the return wire bond and safety ground...

    To me, if they require a >600 rated device between return/neutral/bonded to earth neutral and safety ground--then I can no long use a bonded neutral/return as an "unergized" conductor in my system layouts. Now--I have to double pole breaker on every DC +/- connection, and all electronics must have 500-1,500+ volt isolation between DC +/- and anything that may be accessible to humans (plugs, communications wiring, changing bulbs, etc.).

    Basically, low voltage (under 60 VDC or less, depending on the specification) needs to be managed just like 120 VAC with isolation/barriers/locked doors/double insulation or insulation/ground screen/etc...

    It is not a pretty picture. And difficult to do with DC... AC is "easy" to isolate (transformers), easy to make safe (double insulation or other methods), easy to current limits (again, the magic of transformers), does not sustain arcs nearly as well as DC, and easy to measure current/leakage current (current transformers).

    All of the above can be done with DC--Just not cheaply or easily.

    There is a consortium to make ~370 VDC as the standard for computer data centers (261 VAC * sqrt(2))... There are many levels of safety that we take for granted in the AC world that they have yet to address in the DC distribution world. In their world, they have trained people, locked facilities--In "our world" we have toddlers.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset