Floating an array

Status
Not open for further replies.

Zee

Senior Member
Location
CA
Floating an Array


I just got the biggest shock of my PV career last Summer. As my installation partner and I were laying down panels on a rack, he was shocked. ANY SHOCK CAN KILL YOU. EVEN THE "LOW" AMPS IN PV PANELS ARE 1000'S MORE AMPS THEN NEEDED TO STOP AND SEIZE YOUR HEART MUSCLES.

We analyzed the situation until we could figure out why he was shocked.
I checked online.
I checked with other experts. (Wiles included)

I found that ?floating an array? is key to installer safety.I was surprised to find I never heard the term "floating" before. Nor was I ever taught the practice. (I won't detail my experience, but it's lengthy)
I WANT TO SHARE WHAT I LEARNED AS PART OF MY NEW YEAR'S SAFETY RESOLUTION.
:p

Basically: BEFORE installing PV panels, remove the GFDI fuse or otherwise disconnect the grounded conductor (usually negative) from ground (usually in the inverter). This simple action could be a lifesaver.
The Handbook states: " (GFDI) Devices meeting 690.5 are intended to prevent fires in DC PV circuits due to ground faults." That means that the fuse will not protect life.

Background: NEC 690.5 requires that most PV systems be GFI protected. If you land all wires in DC disconnects and inverters before installing panels, then the GFDI fuse links PV negative (usually, sometimes positive) to ground. (And, landing wires BEFORE energizing them (adding PV PANELS) is a best practice.) That means all metal racking on the roof, all conduit and grounding wire every where is effectively charged (usually negative) - and exposed to the touch. Like one big, exposed and charged bus bar. All I need to do in order to get hurt is to accidentally contact PV positive. I am all but guaranteed to be simultaneously in contact with racking, grounding wire or a module frame by holding, leaning, or sitting on a grounded component, like a rail, i.e. negative. Maximum potential would then exist between the positive wire I accidentally touch and the exposed - negatively charged - metal of anything. Between the two.... my body would be the connection!
Is this unlikely? I used to think so.
Yet....this just happened to my partner on a rooftop. ( I also had a buddy shocked through the seat of his pants while sitting on a rail and wiring an array.) While tightening the set-screw in the bonding lay-in-lug on the underside of the module frame with his screw-driver, and laying the #10 bare copper grounding conductor into it, his screwdriver must have slipped into the MC Connector of the positive lead of the module! He was shocked and yelled.

However, if you remove the GFDI fuse then both polarities should be separately derived from ground - and thus the racking. Potential - in my understanding - should really exist only between the two polarities. If you manage to connect these two with your body, then ....well..... you are talented.

Anyways, just my two cents.

tip #2 Wear gloves. Gorilla brand neoprene fit like a secondskin for me. So I actually wear them!
tip #3 Use insulated screwdrivers. Even when grounding! (as we were)
tip #4 Remove pointy drill bits and pointy tools from toolbelts while installing panels. (MC Connectors are ?touch safe?, only to adegree - but not when handling pointy drill bits and screwdrivers. When handling 100's of modules it is just a matter of time before something slips inside the MC Connector or you crush or nick a wire)
 

kwired

Electron manager
Location
NE Nebraska
I know very little about PV equipment, but my replies apply to all electric equipment in general.

Background: That means all metal racking on the roof, all conduit and grounding wire every where is effectively charged (usually negative) - and exposed to the touch. Like one big, exposed and charged bus bar. All I need to do in order to get hurt is to accidentally contact PV positive. I am all but guaranteed to be simultaneously in contact with racking, grounding wire or a module frame by holding, leaning, or sitting on a grounded component, like a rail, i.e. negative. Maximum potential would then exist between the positive wire I accidentally touch and the exposed - negatively charged - metal of anything. Between the two.... my body would be the connection!
Is this unlikely? I used to think so.

Why is this any different from any other voltage source no matter how derived? Disconnect from the source and you have no more potential for shock. Like I said I know little about PV but if it is easier to disconnect the negative in this case then do so, it still isolates you from the potential. Ground means nothing until it is connected to a system conductor - even in AC systems.

Why would you work on this while it is "live"? Make the final connection that energizes it last then it is "dead" for the rest of the work. Better yet would be a switching method to make this final connection - again I know very little about PV, but this is general information that applies to all types of electrical work.
 

Zee

Senior Member
Location
CA
iwire - I am not surprised PV can kill you. I am very conscious of that, that is my point. I am also surprised that despite our safety planning a member of my team got zapped.

There is a lesson to be learned here for some, I guess not you. ;) I am happy to admit my faults and to learn. I AM surprised - however - that there are simple solutions to NEVER working live that even some of the largest and most experienced installation outfits do not follow. I think that is - shocking.

I have also never been shocked. In 11 years and 300 commercial and residential systems. I have always followed best practices and taken precautions.

resistance - Is it slow for you? I am busy busy!

kwired - exactly, why would you ever work on it live? You nailed it.
I just am not sure many make the mental connection, when on the roof, that the rails are effectively "live".

1. PV is still different from what you do: at some point you add panels, which are ALWAYS live. This presents its own unique challenges. Many outfits run and land all wire first. That is a good idea in order to never have to strip and terminate live or potentially live wires. But it carries the issue I detailed.
2. In ac repair work, do electricians disconnect all neutrals from ground as a precaution? I am recommending the equivalent of that, in an easy step. In PV it is easy to unground all conductors. Just some don't bother.
 

kwired

Electron manager
Location
NE Nebraska
iwire - I am not surprised PV can kill you. I am very conscious of that, that is my point. I am also surprised that despite our safety planning a member of my team got zapped.

There is a lesson to be learned here for some, I guess not you. ;) I am happy to admit my faults and to learn. I AM surprised - however - that there are simple solutions to NEVER working live that even some of the largest and most experienced installation outfits do not follow. I think that is - shocking.

I have also never been shocked. In 11 years and 300 commercial and residential systems. I have always followed best practices and taken precautions.

resistance - Is it slow for you? I am busy busy!

kwired - exactly, why would you ever work on it live? You nailed it.
I just am not sure many make the mental connection, when on the roof, that the rails are effectively "live".

1. PV is still different from what you do: at some point you add panels, which are ALWAYS live. This presents its own unique challenges. Many outfits run and land all wire first. That is a good idea in order to never have to strip and terminate live or potentially live wires. But it carries the issue I detailed.
2. In ac repair work, do electricians disconnect all neutrals from ground as a precaution? I am recommending the equivalent of that, in an easy step. In PV it is easy to unground all conductors. Just some don't bother.

The difference from most utility supplied systems is the fact that the rest of the grid is still grounded.

With your small system if you disconnect the grounded negative lead you have opened the circuit and there is no shock potential.

With an AC system - especially utility fed, if you disconnect the neutrals there is still potential to ground because the system is still grounded someplace (at utility transformer, and usually every utility structure). If you disconnect all places where the system is grounded then you have the same type of situation you are getting at with the PV system. But there is still potential to other system conductors if it is more than a two wire system.

You could disconnect the positive lead(s) and still have a safe working condition, because it only a two wire system. If you were to have a center tap and make it a three wire system (not necessary but not impossible) you have a situation similar to common AC systems.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Here is the way you deal with this problem, which is better than removing the GFDI fuse, and also kills a couple of other birds with the same stone.

Install whips outside of the junction box. Leave these unconnected while you install panels and plug them into home-runs out on the array. Then test voltages for ground fault indications, then plug all the home runs into the whips at the same time. (For extra safety, plug in the ungrounded conductors first.)

-Your array is 'floated' without having to open up the inverter or leave it open until after the panels are installed.
-The procedure incorporates the proper testing of each string, which you should be doing anyway. You test string voltage to make sure you don't have a broken circuit. Then you test both positive and negative to ground to make sure you don't have a ground fault. You cannot test at a combiner where all strings are already connected if you want to know which string is the source of a problem.
-It's nice to have the whips for servicing if you have to come back later and replace or modify a home run on the array.

Finally, note that floating the array doesn't solve the saftey problem if you have a ground fault. Try not to stick your screwdriver in the connectors. ;-)
 
Last edited:

Zee

Senior Member
Location
CA
jaggedben -

sounds good . I see you have given it some thought!
The whips are certainly better than leaving unlanded conductors, as those still have to be landed....live. (I have heard that some leave the negatives to be landed in the inverter at the very end)

With a whip - or a fuse in a holder - a touch-safe method of final connection is ensured. That is basically what i like about both methods.:D
I will say though, that the GFDI fuse is there for (relatively) easy removal. Whips are extra work, and - well - I am lazy. :weeping: That said.... I may just try your idea.

Yeah, floating is not a solution if you have a ground fault. To get shocked, you would still have to create a ground fault PLUS touch the opposite polarity! That likelihood seems infinitely small. With the array grounded only one accidental touch of a live conductor is needed. The rack is "charged" already and receptive to current flow. waddya think?
 

BillK-AZ

Senior Member
Location
Mesa Arizona
jaggedben -

(I have heard that some leave the negatives to be landed in the inverter at the very end)

I do not recommend the above because it generally means handling the cable end(s) in the confines of the inverter and a high degree of a chance of uninsulated contact with the conductor.

Two years ago we were installing a series of systems in hot and dry Phoenix that somehow we were producing a minor shock on the negative (ungrounded polarity) at the inverter in spite of open positive connections to the module strings and having passed a Hi-Pot test on all wiring with over 2000 MegOhms resistance. A digital multimeter would show a negative 50-90 volts that would decay with just the load of the multimeter. Could not find any specific source. Standard Operating Procedure became leaving all end of string connectors open until all wiring was complete, especially the inverter and DC combiner. Extra work as the connections were necessary for the HiPot testing, then had to be opened.
 

Zee

Senior Member
Location
CA
Billk-AZ

Absolutely, I pull all wire and land all wire terminations first to avoid ever having to handle live - or even potentially live conductors. Having a loose conductor in a confined enclosure (inverter) for the duration of an installation and then having to land it in the terminal is not something I want as a standard operating procedure.
 

macmikeman

Senior Member
Why not just dark tarp the panels during install and therefore no voltage generated? Probably not practical on a large commercial job, but not so undoable on a resi pv system.
 

kwired

Electron manager
Location
NE Nebraska
A similar situation where you can't turn off the source would be a bank of batteries.

You can leave the jumper to ground as the last connection - which would be same thing as "floating" it.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Why not just dark tarp the panels during install and therefore no voltage generated? Probably not practical on a large commercial job, but not so undoable on a resi pv system.

Sounds like a pain to me at least half the time. (What if it's windy? What if it's a steep roof?)

But to repeat, handling of this issue is best integrated with proper testing and commissioning which everyone should be doing anyway. Positive and negative conductors for each string should be brought to touchsafe plug connections at the same location (at least close enough that the leads of your voltage meter can span the distance.) Everything else should be finished first, and then as the last task of the installation, every string should be voltage tested (+ to -, + to ground, - to ground) before being plugged in. This has benefits both safety and quality control.

The whips method is not the only way conceivable. For example, in larger installations with repeating string patterns, U-pattern string arrangements can be used so that positive and negative ends of strings are co-located. On tilt-ups and ground mounts, put string ends on the back of arrays for easy accessibility..and so on. Depending on the details, it needn't even cost any extra connectors.
 

Zee

Senior Member
Location
CA
@ MacmikemanWhy not "dark tarp" ?Because removing a 1-amp fuse in an insulated fuseholder is about as easy as it gets.But it is not 100%.Also wind is a major issue.How to attach it? How would I move tarp, panel by panel as i install?Are the panels live if a little light shines through? How about if a corner is exposed?It could work, i just don't know how i would do it. In an easy and practical way.I think it is a great idea when servicing an existing system that has no disconnecting means.@ kwired Yeah, it would seem to be a great idea for batteries too! Haven't dealt with those in a while....@jaggedben Good design considerations all! I re-read your postl and will consider accessibility as a design consideration and factor in string layout and conductor routing in future. btw - I am using Comb. Boxes on all jobs - even the smallest resi installation. I can isolate the ungrounded conductors easily by removing fuses. That, and floating the array, ie. breaking the grounded to ground bond is sufficient safety for me.Yes U routings may actually decrease conductor length and may cost no more connectors.You said, "every string should be voltage tested ".so lemme see....if i had shorted pos. to ground accidentally while mounting panels......and neg. is the grounded conductor in this example..........then i go to plug in the gfdi fuse, what would happen??? It would burn the fuse. But....Would there be an arc? How big? A shock hazard?
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
I am using Comb. Boxes on all jobs - even the smallest resi installation. I can isolate the ungrounded conductors easily by removing fuses. That, and floating the array, ie. breaking the grounded to ground bond is sufficient safety for me.

That's one way to do it, if you're combiners have touchsafe fuseholders. Note that if you experience a ground fault you will still have to disconnect the ungrounded conductors to isolate which string the problem is on. PV voltages have a way of finding their way through the ungrounded conductors and looking the same on all strings if one side is paralleled.

Yes U routings may actually decrease conductor length and may cost no more connectors.You said, "every string should be voltage tested ".

Mostly checking strings is for quality of work. If you do not test each string separately you may well walk away from a job leaving, say, two out of three strings functioning on a system. Or 10 out of 40. It is a comparative waste of money to come back later when the customer is asking why production is so low.

The safety aspects are an additional benefit. Leaving whips to test strings covers both bases.

so lemme see....if i had shorted pos. to ground accidentally while mounting panels......and neg. is the grounded conductor in this example..........then i go to plug in the gfdi fuse, what would happen??? It would burn the fuse. But....Would there be an arc? How big? A shock hazard?

There will be an arc and then the fuse will blow almost simultaneously. The size of the arc will depend on the size of the system and the resistance in the ground fault(s). I think it is safe to say that with small systems of less than 30amps you can do this thing of plugging in the gfdi fuse with a touchsafe fuse holder. Not so good an idea, maybe, on a 50kW central inverter. (I think arcflash hazard is more worrisome than shock hazard in this situation.) But here is the question with that. How could you know for sure that you're not completing the circuit for multiple ground faults when you put in that fuse? Answer: by testing each string beforehand.

By comparison, leaving each string unplugged until you test it involves no comparable possibility of a high current arc, even on huge systems, because by testing you can avoid plugging in any strings that are ground faulting. If you are doing this, it really avoids the need to mess with the GFDI fuse. And to repeat, it helps ensure a fully functioning system.

(In principle it is always safer to connect the ungrounded conductor first, since if the grounded conductor is connected first, the ungrounded conductor can short to ground, where as in the opposite case they can only short to each other. In practice, one should have methods like I've described that make either unlikely.)

I have heard of guys lighting cigarettes with the bare ends of live PV string conductors. I don't recommend doing this. Smoking kills. :lol:
 

PaulWDent

Member
PV panels don't produce serious voltages until connected in series

PV panels don't produce serious voltages until connected in series

Typically, a PV panel is not going to produce more than 45 volts DC open circuit. You will barely feel that, it at all. So to get a shock, you must have connected a string of panels in series fifrst. So there is the answer: Complete the series connections between panels (usually made by PV connectors) last. There are some 150 volt panels around (usually the thin film variety) but they are rarer and deserve their own degree of caution in handling.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
So there is the answer: Complete the series connections between panels (usually made by PV connectors) last.

I can tell you that in the real world, where a crew of installers is trying to work efficiently to get a job done in a timeframe, this strategy simply isn't that realistic.
 

Zee

Senior Member
Location
CA
So there is the answer: Complete the series connections between panels (usually made by PV connectors) last. There are some 150 volt panels around (usually the thin film variety) but they are rarer and deserve their own degree of caution in handling.

Do you install or design? :happyno: You cannot access the connectors once the panels are mounted in an array with multiple rows!
BUt it is still irrelevant: ultimately you are still connecting one to the other and adding voltage. So it is just like what we all do, plug one into the other.

However Removing a fuse, is...... what, 1-2 seconds?
 

Zee

Senior Member
Location
CA
That's one way to do it, if you're combiners have touchsafe fuseholders. Note that if you experience a ground fault you will still have to disconnect the ungrounded conductors to isolate which string the problem is on.

Mostly checking strings is for quality of work.

How could you know for sure that you're not completing the circuit for multiple ground faults when you put in that fuse? Answer: by testing each string beforehand.

Yup always touch safe fuse holders in CB's.

Yes all strings are checked. Good idea always agreed.

Finding a ground fault is extremely rare. Very rare. I am happier to go back at end of install to find a ground fault than having it fault to ground...thru me......while installing. I understand you are able to avoid this danger also with your method. Awesome!

" But here is the question with that. How could you know for sure that you're not completing the circuit for multiple ground faults when you put in that fuse? Answer: by testing each string beforehand."

Certainly your solution is valid and practical.
ALternatively, with large systems and multiple strings and hence BIG AMPS, I do as follows:
I remove all ungrounded (POSITIVE) fuses
and dc switches off.
Plug in GFDI fuse after all panels are in.
PLug in fuses.
TUrn DC on.
If fuse blows you weren't there to burn your face.
Your whips work also.

BUT i gotta say.....:roll: if you or your installers are regularly crushing those poor, defenseless 600 VDC PV wires so that you have multiple shorts in a single system, then hire my younger brother! Even he can do better work.:p:lol:


""I have heard of guys lighting cigarettes with the bare ends of live PV string conductors. I don't recommend doing this. Smoking kills. :lol: ""

Funny. HOw much do lighters COST in your area?? These guys need more work. haha
 

Zee

Senior Member
Location
CA
Complete the series connections between panels (usually made by PV connectors) last.
I just realized by "last" you may have ,meant at end of complete SYSTEM install - of ALL components and wiring.
That is how all systems are built - or should be - agreed. I mount panels at very end in a couple hours.
But mounting panels has to be concurrent with making series connections. No way around it.

BUt agreed: no reason to have panels all wired up to 600 VDC with power having nowhere to go!
 
Status
Not open for further replies.
Top