Fusing parallel strings

larcal · May 2013

Hi Gang,

Not sure if correct etiquette is to post this in 2010 thread or not, but long time ago and didn't answer this.

Embarrassed to say I've been limping along on a 1990's system which was hacked together with a few 80's HP articles and no internet, plus now a little hazy (okay, maybe a lot) on what I did know. 1000 watts of different manufacture panels and pwm controller all paralleled. Kept to 12 volt cause Mppt was still out there. Time to bone up and improve maybe. Not sure yet if willing to abandon 12 volt wiring cause panels all different and kind of spread out. Efficiency losses with differnt vpm's and Ipm's, for sure. Got that straight. Found 4/0 cable then for 1 buck/foot Geez. Maybe get mppt controller but not sure I'd really gain more then 10% when panels still 12 volt. So a gamble but not the subject here.

Anyhoo, to get to it. Recently became aware that one is supposed to fuse the positive wire between each panel in a parallel string after the second panel. Did I know that then and chose to ignore it to avoid voltage drop? Can't remember. Hoping you all can explain what the risk is, or the odds of any particular scenario, and why.

For simplicity, Let's isolate to one 10 panel group of 55 watt (vpm=17.4, 3.1 rated amps) panels in parallel so 550 watts. Rack on dirt, no fire risk to property.

So if a short occurs in one panel what is going to happen is the query. So the others pour amperage into the short but that panel is toast anyway. How likely is it that I would also lose other panels on the rack?. If so, why? from fire? How could a fire spread to other panels when there is so little to burn in a panel? Maybe there's another reason I'd lose other panels? If I move to 24 series strings that are then paralled would that change the risk much? Doesn't seem like it to me. The volt drop from 8 fuses alone just for that group is a bummer, but losing all those panels, argh!

Second thing is I just fused on battery side of controller. I mean, not sure, but aren't there diodes in controller or in panels that would keep juice from flowing from battery to the panels and frying them in the event of this panel short?

Lastly, basic electricity I quess, but how does this work? I mean, if each panel is producing 3 amps in a 10 panel parallel group then this means that there is 30 amps flowing through the whole group in a circle, i.e, thru each interconnect wire between panels, correct? so 8 bigger then 30 amp fuses fuses needed. Somehow I know this is totally wrong. No doubt really stupid. trying to remember (or learn first time) but so much to do.

Any elucidation much appreciated

westbranch · May 2013

Re: Fusing parallel strings

I'll take a crack at it. think of the whole array, one short so it now has 27 amps it can send into the shorted one, than a second one goes , now there is 24 amps to send through the 2 shorted ones... it has a cascading potential, so that is why when over 2 panels in parallel, you need circuit protection for reverse flow through any panels.

hth

as to the rest of your problems it may be better to build a new one with a post describing just what you have.

ps MidNite Classics are a favourite of some on this board...

Cariboocoot · May 2013

Re: Fusing parallel strings

Any given panel is designed to take its Isc plus a bit, hence it will have a maximum series fuse rating - usually about 10 or 15 Amps.

What happens when one shorts is that the remaining parallel panels can push their Isc to it. if this is just one more panel of the same type the current rating of the shorted panel will not be exceeded so there's no potential over-current problem. What happens with three or more panels is parallel is that the shorted one can receive 2X or more its Isc rating and the current maximum can be exceeded.

Now, what happens when you put more current to a conductor than it is designed to happen?

Sparks, flames, fire on the roof maybe.

With many panels of different Imp ratings you may find that the fuse requirements are not the same for each and every panel. A panel with an Imp of 7 and another with an Imp of 8 will probably both be covered by 10 Amp fuse rating. But if you have a 70 Watt panel it may have an Imp of 4, so any single larger panel could exceed the current maximum.

Hopefully they are all "12 Volt" panels so the Vmp ratings won't be too far off (not more than 10% difference for preference).

Second thing is I just fused on battery side of controller. I mean, not sure, but aren't there diodes in controller or in panels that would keep juice from flowing from battery to the panels and frying them in the event of this panel short?

The main problem with reverse current flow (occurs at night) is not frying the panels but draining the batteries. For the most part any system with a decent charge controller will not suffer from this because the controller itself uses a switching semi-conductor which will block reverse current flow.

In the even the controller shorts you need that fuse between the battery and controller otherwise the battery can dump huge amounts of current to the short and set the wire on fire.

Fuses do not cause any appreciable Voltage drop. Blocking diodes do. The two are not interchangeable.

vtmaps · May 2013

Re: Fusing parallel strings

larcal wrote: »

Rack on dirt, no fire risk to property.

So if a short occurs in one panel what is going to happen is the query. So the others pour amperage into the short but that panel is toast anyway. How likely is it that I would also lose other panels on the rack?. If so, why? from fire? How could a fire spread to other panels when there is so little to burn in a panel?

I doubt (but I'm not certain) that the fire would spread from one panel to another. If panels are on roofs, a fire could be devastating. My panels are not near any flammable structure, but they are fused anyway... I wouldn't want to risk even a grass fire.

larcal wrote: »

Second thing is I just fused on battery side of controller. I mean, not sure, but aren't there diodes in controller or in panels that would keep juice from flowing from battery to the panels and frying them in the event of this panel short?

To expand on Cariboocoot's response: One failure mode (not sure how common) for a charge controller is to short out the battery. The fuse will keep it from burning down the house.

--vtMaps

larcal · June 2013

Re: Fusing parallel strings

My panels are not near any flammable structure, but they are fused anyway... I wouldn't want to risk even a grass fire.

Thanks, Vtmaps. This means you fused between each panel in a parallel string, correct?

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Thanks, Vtmaps. This means you fused between each panel in a parallel string, correct?

A 'string' is one or more panels connected in series.
The fusing goes on each and every string, before the parallel connection to the other strings.
It is only done on the positive side; the negative side does not need fusing as well.

larcal · June 2013

Re: Fusing parallel strings

Thanks all, but I'm not sure we are on the same page on a couple of things so please allow me to rephrase myself a bit, again using the hypothetical example of 10 55 watt panels connected only in parallel with the rack on dirt.

* First, do you agree that one is supposed to (by code) have a fuse between each positive terminal in said string? Not sure you do. Regardless is it necessary?

* Second, Do you agree the code requires a fuse between the panel and controller? Regardless, is it necessary? Yes I'm aware of the need for a fuse between controller and battery and have allways had one there.

* Third, can you clarify this. Panel *1 produces 4 amps and so 4 amps flows between *1 and *2. Then *2 creates 4 amps and so 8 amps flows between *2 and *3. 12 amps between 3 and 4 and so on until 36 amps are flowing between 9 and 10 so you would need increasingly larger fuses as you progress. Now I think this is not true because the same 5 amp fuse seems to be required between each panel but I don't get why. How do we get 40 amps flowing out of the rack if it's not increasing between each panel?

Now, what happens when you put more current to a conductor than it is designed to happen?

Sparks, flames, fire on the roof maybe

. ---Caraboocoots

I get that all the current flows to the shorted panel, and that some kind of flash flame will occur, but do you agree with VTmaps that it is pretty unlikely to lose the other panels, (assuming no other surrounding combustibles), because silicon, aluminum, and glass don't burn? I ask because I'm not sure and others may have more experience. I'm worried about losing the whole rack, not a house fire, because on dirt. And do you agree that the fire on the one panel would be very brief? not likely to ignite something like plywood if one does have that nearby. Paper etc. sure.

The main problem with reverse current flow (occurs at night) is not frying the panels but draining the batteries. For the most part any system with a decent charge controller will not suffer from this because the controller itself uses a switching semi-conductor which will block reverse current flow.

---caraboocoots

Interesting about the switching semi conductor but my understanding was that there was a concern that if a panel shorted an immense amount of battery current would flow back to the panels (thus frying them all) and that this was the reason for a fuse between panels and controller. But I'm unsure. Is this truely the concern? If this is not a realistic concern, why the fuse in question 2, i.e, between panels and controller? My original question was why would the diodes in the panels or controller (or semi conductor) not stop this shorted back flow from happening?

Fuses do not cause any appreciable Voltage drop. Blocking diodes do. The two are not interchangeable.

This statement is why I asked question #1. Really?, am I wrong on this? Never actually measured this but kind of learned from all those old Perez articles to be absolutely paranoid of each connection, soldering being the (usually unachieved) ideal. We're talking 2 extra crimped and heat shrinked connections per fuse or 16 xtra for just that rack. And out in elements.

Hoping Icarus (?) sees this and responds as I think I somewhere saw him make an inferrence on this subject. Sorry for so much length but internet here kind of happenstance so need to be as complete as possible with every chance. And second typing cause lost my log in.

Best to all. Thanks again.

mike95490 · June 2013

Re: Fusing parallel strings

Attachment not found.Attachment not found.Attachment not found.

So here are some diagrams of a SERIES STRING, Parallel (not called a string) and a mix of Series & Parallel

If you imagine these are PV panels instead of batteries , can you see how series, POS connects to NEG, and that parallel Pos connects to Pos.

larcal · June 2013

Re: Fusing parallel strings

Cariboocoot wrote: »

A 'string' is one or more panels connected in series.
The fusing goes on each and every string, before the parallel connection to the other strings.
It is only done on the positive side; the negative side does not need fusing as well.

Sorry, caraboocoots, posted my second (#7) at 18:52 PDT before seeing yours. Please bear with me on the rest of post anyway. But yeh, this seems to answer part of it. In your example a series string is like one panel in my paralled group rack. Each panel needs a fuse between positive terminals, starting after the second panel, or 8 fuses in a group of 10 55 watt panels or 550 watts at 17-18 volts

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Thanks all, but I'm not sure we are on the same page on a couple of things so please allow me to rephrase myself a bit, again using the hypothetical example of 10 55 watt panels connected only in parallel with the rack on dirt.

* First, do you agree that one is supposed to (by code) have a fuse between each positive terminal in said string? Not sure you do. Regardless is it necessary?

More than two panels or strings of panels in parallel requires individual string fusing/breakers.

* Second, Do you agree the code requires a fuse between the panel and controller? Regardless, is it necessary? Yes I'm aware of the need for a fuse between controller and battery and have allways had one there.

Depending on jurisdiction a disconnect of some sort may be required. Actual circuit protection would function in the even of a short on the controller input.

* Third, can you clarify this. Panel *1 produces 4 amps and so 4 amps flows between *1 and *2. Then *2 creates 4 amps and so 8 amps flows between *2 and *3. 12 amps between 3 and 4 and so on until 36 amps are flowing between 9 and 10 so you would need increasingly larger fuses as you progress. Now I think this is not true because the same 5 amp fuse seems to be required between each panel but I don't get why. How do we get 40 amps flowing out of the rack if it's not increasing between each panel?

Current flows as a function of power production and current path. Each matched panel exposed to the same amount of light will produce the same amount of current. Since the panels themselves are semi-conductors the current does not flow from one panel to another but collectively to the charge controller input.

If one panel should fail (short) it is capable of handling its maximum current (Isc and a bit) as per its series fuse rating. If there is only one other panel to produce current this is not a problem as its output will not exceed that rating. With multiple panels it can. The shorted panel being the new "path of least resistance" and no longer functioning as a semi-conductor means the collective current from the other panels will try to flow through it, exceeding its maximum current ability and possibly starting a fire.

This action will not affect the good panels, only the shorted one. What it will ignite depends on what is near it when it goes up. Depending on the amount of current, size and type of wire, and Voltage this could either be a brief flash or a sustained ignition point.

---caraboocoots

Interesting about the switching semi conductor but my understanding was that there was a concern that if a panel shorted an immense amount of battery current would flow back to the panels (thus frying them all) and that this was the reason for a fuse between panels and controller. But I'm unsure. Is this truely the concern? If this is not a realistic concern, why the fuse in question 2, i.e, between panels and controller? My original question was why would the diodes in the panels or controller (or semi conductor) not stop this shorted back flow from happening?

A shorted panel will not cause current to back-flow from the battery. Only a failure of the controller will do that. That is why the wiring between the battery and controller is fused against maximum current: if something goes wrong that will blow rather than some uncontrolled spot in the wiring.

This statement is why I asked question #1. Really?, am I wrong on this? Never actually measured this but kind of learned from all those old Perez articles to be absolutely paranoid of each connection, soldering being the (usually unachieved) ideal. We're talking 2 extra crimped and heat shrinked connections per fuse or 16 xtra for just that rack. And out in elements.

No, this is why combiner boxes were invented. You do not want to put in-line fuses on each panel connection.

Really one day I shall have to draw a schematic of panel wiring options. One day.

larcal · June 2013

Re: Fusing parallel strings

Gosh Mike, I certainly understand how Parallel and series connections are made and combined. Can't figure what I did wrong, how you could have thought this.

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Sorry, caraboocoots, posted my second (#7) at 18:52 PDT before seeing yours. Please bear with me on the rest of post anyway. But yeh, this seems to answer part of it. In your example a series string is like one panel in my paralled group rack. Each panel needs a fuse between positive terminals, starting after the second panel, or 8 fuses in a group of 10 55 watt panels or 550 watts at 17-18 volts

If you have ten panels in parallel you need ten fuses: one on each.

larcal · June 2013

Re: Fusing parallel strings

Cariboocoot wrote: »

If you have ten panels in parallel you need ten fuses: one on each.

Huh? No fuse between first two as you said and nothing after #10 on way to controller so this time drew a little diagram and still get 8.

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Huh? No fuse between first two as you said and nothing after #10 on way to controller so this time drew a little diagram and still get 8.

No, if you have only two panels in parallel you do not need a fuse.

Three or more panels in parallel and every one must have a fuse.

The reason being you do not know which panel may short: any one that does will be subjected to the combined current of the remaining panels which will exceed any one's maximum rating.

With the fuses in place, the one on whatever panel shorts will blow first: controlled current interruption capable of breaking the circuit in a sealed container that can not start a fire.

larcal · June 2013

Re: Fusing parallel strings

Depending on jurisdiction a disconnect of some sort may be required. Actual circuit protection would function in the even of a short on the controller input.

Well I was wrong. How the hell did I get that idea? Will have to review. Thanks

Current flows as a function of power production and current path. Each matched panel amount of light will produce the same amount of current. Since the panels themselves are semi-conductors the current does not flow from one panel to another but collectively to the charge controller input

.

Well you're the expert so I'll have to take your word for it, but I don't get it and wish I did. I don't suppose there is some kind of overly simplistic metaphor or 1/2 wrong example to explain how it doesn't build up cumalatively between panels? Hah! Dreaming I know. So, all fuses same size. Good

A shorted panel will not cause current to back-flow from the battery. Only a failure of the controller will do that. That is why the wiring between the battery and controller is fused against maximum current: if something goes wrong that will blow rather than some uncontrolled spot in the wiring.

So hypothetically, with a shorted panel if there was no controller in line and the fuse at the battery failed to blow would not juice flow backwards and fry all the panels? And so with the controller in line is it that semi conductor you mentioned that stops the back flow?

No, this is why combiner boxes were invented. You do not want to put in-line fuses on each panel connection.

Hah! this is great you telling me this, but I don't get how any kind of "combiner box" I can visualize could do that. Never seen one. Some kind of fancy junction box but how can you avoid all those fuses. How many would you need for said group of 10 panels? Yes wish I could see a diagram of innards. Maybe have to order one just to see. then send back and get right size.

is there a time limit before one loses log in and therefore post?

larcal · June 2013

Re: Fusing parallel strings

Cariboocoot wrote: »

No, if you have only two panels in parallel you do not need a fuse.

Three or more panels in parallel and every one must have a fuse.

The reason being you do not know which panel may short: any one that does will be subjected to the combined current of the remaining panels which will exceed any one's maximum rating.

With the fuses in place, the one on whatever panel shorts will blow first: controlled current interruption capable of breaking the circuit in a sealed container that can not start a fire.

Geez, I'm sorry Carabboocoots, I'll wear you out so letting it go, but this drove me nuts. Regarding your first sentence, yes I more then get it, and the reason. No fuse required between first 2. So that leaves 8 left, starting with space between 2 and 3. Maybe I need a trip to see the wizard

Great name, by the way.

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Well you're the expert so I'll have to take your word for it, but I don't get it and wish I did. I don't suppose there is some kind of overly simplistic metaphor or 1/2 wrong example to explain how it doesn't build up cumalatively between panels? Hah! Dreaming I know. So, all fuses same size. Good

It does "build up between panels" but it doesn't flow backward through them because they are semi-conductors: current only "flows out of them". Diode: --->|--- Current flows in the direction of the arrow, but can't flow back across the bar (unless something goes wrong with it).

So hypothetically, with a shorted panel if there was no controller in line and the fuse at the battery failed to blow would not juice flow backwards and fry all the panels? And so with the controller in line is it that semi conductor you mentioned that stops the back flow?

This part gets tricky. If the panels were connected directly to the panels and the panels were not illuminated the battery can send current backward across the panels and drain the battery. Whether or not this happens depends on the Voltage involved; generally a 12 Volt system will not suffer from this because the battery Voltage is too low to affect the panels' reverse Voltage breakdown point. But if one of the panels is shorted, illumination or not, it will act like a piece of wire and allow current flow from any source in any direction across it. As such the battery can dump its full current potential into it which is almost certain to be many times higher than the panel was designed to handle: instant fire.

Hah! this is great you telling me this, but I don't get how any kind of "combiner box" I can visualize could do that. Never seen one. Some kind of fancy junction box but how can you avoid all those fuses. How many would you need for said group of 10 panels? Yes wish I could see a diagram of innards. Maybe have to order one just to see. then send back and get right size.

Take a look at this MidNite combiner box: http://www.solar-electric.com/misomnsoarco5.html
Basically the panel wires all go inside where their negatives connect to a common bus bar and their positives go to individual breakers, the other side of which connect to another bus bar. Two wires then go from bus bars to the charge controller.

is there a time limit before one loses log in and therefore post?

Yes, and it seems to be random at times. What's more there's a time limit on post writing (recently lengthened because too many posts were getting lost). If you have to sign back in and start over in the same thread there's a box at the bottom left that will say something about "restore from autosave"; a function that saves a 'draft' copy of your post to the server before completion. You may note the little yellow message popping up on the right as you type saying "autosaved".

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Geez, I'm sorry Carabboocoots, I'll wear you out so letting it go, but this drove me nuts. Regarding your first sentence, yes I more then get it, and the reason. No fuse required between first 2. So that leaves 8 left, starting with space between 2 and 3. Maybe I need a trip to see the wizard

No; no fuse required between only two.
Three (or more) panels in parallel = three (or more) fuses, one on each panel.

Great name, by the way.

A "Cariboocoot" is a mythical bird.

larcal · June 2013

Re: Fusing parallel strings

Hi, Well, tough week and just getting back to this. Not too practiced at quoting but the quote below comes as a response to my objection that every fuse meant more voltage drop if only because any break in a wire involves new crimps etc. So you said:

No, this is why combiner boxes were invented. You do not want to put in-line fuses on each panel connection

.

I thought this meant that somehow a combiner box allowed one to use a lot fewer connections somehow but I see from your picture of one (thanks for that, very helpful) and description that it does not. Clearly the box seals against weather but why does it not create a lot more volt drop? I.e,

* In my example of 5 panels on a rack one just has 1 foot of wire or less between each positive terminal, wheras with a box there is no direct connect between panel terminals and the amount of wire is vastly increased. In addition to vd, there is a real aesthetic issue here imo.

* Each wire is still broken in the middle for the fuse or cb so no saving there as far as vd. You said elsewhere that fuses don't create vd but my understanding is that even if they don't the crimping does. Is this wrong?

Regardless, it looks like the midnight is using cb's where you just push a stranded wire under a screw plate and tighten and I have found those to be the worst. Hate them. Unsealed connections oxidize and you can't sand off the stranded wire later.

Anyway, I'm just surprised that it's done this way. Is this really the norm do you think? I mean, assuming one can mount the panels in a fire resistant way like on a metal roof or dirt wouldn't it be electrically more efficient to do it my way and skip the fuses? Attachment not found.

Thanks for the tip on losing log in. If I chose the "go advanced" function I notice there is no tab in the lower left for restoring auto save, though the yellow label does flash like you said so what to do if i lose it?

Cariboocoot · June 2013

Re: Fusing parallel strings

There is no reason for a fuse, breaker, or connection of any kind to create Voltage drop. V-drop is a function of an increase in resistance @ current. Proper sizing and installation of components will not cause any noticeable increase in resistance. Bad connections, too small wiring/components - that will cause problems. If you are concerned about corrosion (some locations are worse than others for this) a little dialectric grease will seal out the possibility. The MidNite boxes are a much better way of making connections than spliced wires, believe it or not.

You should be able to restore from autosave first, then return to "advanced" before posting if you need to use the extra functions.

larcal · June 2013

Re: Fusing parallel strings

But if one of the panels is shorted, illumination or not, it will act like a piece of wire and allow current flow from any source in any direction across it. As such the battery can dump its full current potential into it which is almost certain to be many times higher than the panel was designed to handle: instant fire.

You said this in reference to my hypothetical of there being no controller in the line post #15.

* So even if no fire, all this current would destroy the whole group of panels, yes?

* you have said that the inline cb or fuse betwixt controller-battery as well as the "switching semi conductor" (which I just understand as some kind of diode) in any decent controller will protect against losing panels (or fire). I have a 60 amp Square D cb between controller and battery that I don't necessarily trust to work or work fast enough. My current xantrex c60 is dioded but believe I've read the Morningstar Tristar is not, which was thinking of trying. I thought the Panels had a diode also but I quess from your answer that was wrong. Is adding a diode between the controller and panels a dumb or unnecessary idea do you think? Lot more volt drop?

I do kind of gather a panel short is very rare so maybe it's silly just on that basis

Tell us more about this mythical duck Attachment not found.

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

You said this in reference to my hypothetical of there being no controller in the line post #15.

* So even if no fire, all this current would destroy the whole group of panels, yes?

No. Panels are capable of handling their full current rating and only allow it in one direction. In the case of multiple good panels in parallel, nine can not feed their total current "backwards" through one because it will not allow the current to flow backwards. This problem only occurs if one of them shorts, becoming a bi-directional conductor. It then acts like a piece of wire connected to the others: the combined current of the remaining panels will flow through it.

Note that this applies to parallel sets of the same panels; things get more complicated when you start to mix them (differing reverse Voltage values and such).

* you have said that the inline cb or fuse betwixt controller-battery as well as the "switching semi conductor" (which I just understand as some kind of diode) in any decent controller will protect against losing panels (or fire). I have a 60 amp Square D cb between controller and battery that I don't necessarily trust to work or work fast enough. My current xantrex c60 is dioded but believe I've read the Morningstar Tristar is not, which was thinking of trying. I thought the Panels had a diode also but I quess from your answer that was wrong. Is adding a diode between the controller and panels a dumb or unnecessary idea do you think? Lot more volt drop?

All charge controllers are made from switching semi-conductors. It's how they work. That fact will prevent current from the battery from back-feeding dark panels and draining it. The circuit protection on the wiring from controller to battery is there to protect the circuit in case the charge controller should short. If the wiring and CP are sized correctly for the controller (based on its maximum output rating with/without NEC derating as needed) then the CP will go before the batteries can dumb hundreds of Amps into the circuit and burn things up. If the controller is shorted it is unlikely any current would flow back to the panels, and again they are designed to take 100% Isc indefinitely and will block current in reverse if illuminated (and V under reverse breakdown levels). Putting an additional diode in line will do nothing but drop Voltage under normal operation.

I do kind of gather a panel short is very rare so maybe it's silly just on that basis

Almost totally unheard of, in fact. The known cases tend to be homemade or badly made panels (falsely labeled as certified). You're more likely to have a problem with the average store-bought extension cord.

Tell us more about this mythical duck Attachment not found.

Didn't you see the picture?

larcal · June 2013

Re: Fusing parallel strings

No. Panels are capable of handling their full current rating and only allow it in one direction. In the case of multiple good panels in parallel, nine can not feed their total current "backwards" through one because it will not allow the current to flow backwards. This problem only occurs if one of them shorts, becoming a bi-directional conductor. It then acts like a piece of wire connected to the others: the combined current of the remaining panels will flow through it

Yes, and if the combined current of the 9 can flow thru this bi directional conductor, then why won't the entire battery current flow thru the whole 10, and because it is so massive, destroy them all? that is, if my controller does allow reverse and my cb fails as they often do. Thus my query of putting extra protection in.

Damn Guy, you're a hero. And so fast. Must mean there is a way to quickly erase everything in a post except what you are quoting? without gluing your finger to the backspace key while you make dinner. Attachment not found.

Anyway, beating a dead horse but did learn a lot.

Attachment not found.

vtmaps · June 2013

Re: Fusing parallel strings

larcal wrote: »

Yes, and if the combined current of the 9 can flow thru this bi directional conductor, then why won't the entire battery current flow thru the whole 10, and because it is so massive, destroy them all? that is, if my controller does allow reverse and my cb fails as they often do. Thus my query of putting extra protection in.

In order for battery current to melt down a panel, there would have to be two faults... the controller would have to fail and the panels would have to be shorted out. Fortunately, this meltdown won't happen because you have circuit breakers. If they also failed you would indeed be SOL. The NEC is designed to prevent disasters caused by faults, but it can only go so far... Three simultaneous faults are so unlikely that the code does not anticipate that it can happen.

--vtMaps

ggunn · June 2013

Re: Fusing parallel strings

FWIW, string fuses are not just to protect the modules but the wiring as well. If you have 10 8A strings/modules in parallel with no string fuses wired with #10 wire (which is typical), and you have a wiring fault somewhere, then all 10 modules/strings will feed the fault with 80A. #10 wire is rated at 40A; 80A will fry it and quite possibly start a fire. If every string is fused at 15A (also typical), then the available fault current is only 23A (8A from the faulted string plus 15A from the other strings through the fuse before it blows), so your wiring is saved. Wiring faults are a lot more common than module faults.

BB. · June 2013

Re: Fusing parallel strings

Specifically for Code and UL type testing, they only assume single fault scenarios... I.e., a controller faulted OR a solar string shorted. Not both at the same time.

However there is something that it implied in the code but now well understood by many (in my humble opinion) about subsequent faults.

A fault/protective action cannot make a system "more dangerous" or more likely to fail/cause damage/hurt someone.

For example, with a North American Power type system where the Neutrals are bonded to earth ground. We never put fuses/breakers or even on/off switches in the neutral/return side of a wiring circuit. If the neutral switch/protective device is opened, the current is stopped, but the device/wiring is still energized.

If the same switch/breaker is in the "hot side" of the power line (black or red wire), then the device is "cold".

The only time we can put a breaker in the return side is if there is a breaker in the hot side too, and they are ganged together so that if one breaker is tripped, the other is tripped too (i.e., 240 VAC split phase double pole breakers as used for water heaters, electric stoves, etc.).

Which brings us back to the DC GFI ground fault detection fuse installed between "DC Return" and earth ground. The power system is grounded through the "sense" fuse/breaker until the ~1 amp device trip current is exceeded. Then it opens and the system is not longer a neutral to earth bonded system. This fuse will stop the current (not really part of the design, but how it works in real life) and the initial fault current is stopped. However, the system now has a "hot DC Return wire" (no longer grounded) and various wiring paths no longer have over current protection (i.e., the rest of the return wires are "hot" and have no fuses/breakers in them).

This is a major example of where a tripped protective device as required by Code and UL makes the system less safe after a "trip" and subsequent failures may no longer be protected.

Note: I have a big problem with DC GFI Breakers and think they are unsafe to use in a "typical" off grid solar power system--regardless of what the NEC says. Normally, you would never ever put a fuse or breaker in the main system DC safety ground.

If you want to read the details "why" I believe this... Here is a discussion on the Midnite Forum:

Ground Fault Protection

-Bill

Cariboocoot · June 2013

Re: Fusing parallel strings

larcal wrote: »

Yes, and if the combined current of the 9 can flow thru this bi directional conductor, then why won't the entire battery current flow thru the whole 10, and because it is so massive, destroy them all? that is, if my controller does allow reverse and my cb fails as they often do. Thus my query of putting extra protection in.

What vtMaps said.

The controller would have to fault to allow battery current to flow back to the panels, then the panels would have to fault to allow current to flow backwards through them. And there would have to be no circuit protection available to halt current flow if it exceeds the maximum the circuit is designed to handle. This comes under the heading of almost no chance whatsoever of happening.

Power is going to take the path of least resistance. If the controller shorts, the battery power will flow through it (not the panels) causing the current to rise above the maximum and trip the circuit protection.

Blocking diodes are confusing you. They are used in about two or three instances. The first is in small, unregulated panels of the "battery maintainer" type that you plug in to your vehicle to keep its battery up while it sits unused. There the blocking diode prevents battery discharge through the panel at night (when the panel would conduct backwards). The second is in certain higher Voltage applications where one part of an array may be partially shaded at times, causing an imbalance in Voltage to the point where the higher output Voltage could exceed the reverse conducting Voltage of the other panels (to some extent all diode devices act like Zeners in that they do have a breakdown Voltage allowing reverse conducting). The third would be where there is a mixture of different Voltage panels in parallel strings: whereas several lower Voltage panels add up to the same Vmp of a higher Voltage panel their reverse Voltage is lower so again it is possible under some conditions to cause them to reverse conduct.

12 Volt systems with charge controllers and matched panels would never need blocking diodes. Some people also get confused by bypass diodes which are built-in to panels and are not the same thing: they allow current flow around cell segments that are not illuminated.

Fusing parallel strings

Comments

Categories