NEC Overcurrent Protection Devices and Over-Paneling

[calinb]

According to what I've read here, https://diysolarforum.com/resources...nes-for-solar-panels.143/version/324/download
the NEC does not accommodate highly "over-panelled" solar installations.

I live in a very solar-adverse location. If I did not fear the high likelihood of losing commercial electrical power for a lengthy or indefinite period of time, I would not even consider solar power as anything more than a hobby and curiosity. Since before the last winter solstice, I have been taking output measurements of portable Ecoflow 220W bi-facial panels as well as a pallet of SunPower E-Series: E20-330-COM-MLSD panels. It would behoove me to install the SunPower PVs in two highly over-paneled 2s9p arrays on the roof of my shop to feed my two Ecoflow Delta Pros.

If I were to install small 2s2p arrays, no OCP devices would be required by the NEC, because the arrays would contain only two strings each. On the other hand, a small array isn't really worth installing where I live but 2s9p arrays could support food refrigeration and a well pump but the NEC says they require very large fusing and wire sizes.

The NEC requirements don't make sense to me in the over-paneled 2s9p system, because the worst case currents used in the NEC calculations will never occur with the 2s9p array feeding an Ecoflow Delta Pro SCC (15A / 150V max. solar input) unless there is a fault somewhere and then I would happily welcome under-NEC-spec fuses to blow as a result. Wiring and fuses could easily and economically be sized to protect the system from faults in connectivity, backfed shorted panels, charge controller, etc.

Has anyone run into the above considerations in designing an "over-paneled" system?

Then there's the abstract here:
https://ece.northeastern.edu/groups/power/lehman/Publications/Pub2011/2011_9_Zhao.pdf

"This paper focuses on the challenges to overcurrent protection devices (OCPDs) in a PV array under two types of unique fault scenarios. One is a line-line fault that
occurs under low irradiance conditions. In this circumstance, the fault current may not be large enough to trip the OCPDs in the PV array, even when high irradiance occurs later in the day."

My 2s9p arrays would be operating in low irradiance conditions most of the time where I live and smaller fuses mitigate the problem.

Any thoughts and tips would be much appreciated!

-Cal

 

[zanydroid]

Your post is a bit confused.

Can you point specifically to the NEC rules that you think are busted by your design? Plenty of solar installers and forum members are able to make progress under the NEC rules, without IMO undue requirements.

I don't find the NEC rules constrained at all for the normal 9s2p systems (9 series, 2 parallel strings). In fact these are the string configs targeted by 600VOC limit MPPTs, which is standard in modern AIOs.

If you do 2s9p (2 series, 9 parallel), then yeah that is far outside the norm for a house-scale system. If you want to configure a system like that NEC will require you to use something like AWG 1 conductors. The main reason I can think of to do something as wide as 9p is if you want to aggressively defend against shading; the standard NEC compliant approaches to this include wiring microinverters in parallel, or using optimizers to allow comparable shading restrictions while still using a series string.

If you insist on 2s9p, you might be able to get away with putting the 9p combiner on the roof, with 9x 15A or 20A fuses on the input of the combiner and 20A fuses on the output.

BTW using a Delta Pro would not be NEC compliant unless that is 1741 or 9540 listed.

If you used a regular inverter for a house-sized use case like 6000XP instead of the Delta Pro (which is for camping/small systems) you will not have these issues with those 330W panels. And installers are using at least this tier of equipment.

As a former academic (I published in computer science journals/conferences in likely comparable or higher tier as that paper), I would go out on a limb and say that particular paper (as well as 99.9% of other conference and journal papers), NEC doesn't give a crap about unless someone makes a lot of noise about it within the rules making committees.

 

[calinb]

Your post is a bit confused.

Can you point specifically to the NEC rules that you think are busted by your design? Plenty of solar installers and forum members are able to make progress under the NEC rules, without IMO undue requirements.

I don't find the NEC rules constrained at all for the normal 9s2p systems (9 series, 2 parallel strings). In fact these are the string configs targeted by 600VOC limit MPPTs, which is standard in modern AIOs.

Thanks very much for you prompt and helpful reply, zanydroid! I think you have educated me that my 2s9p array would be highly unusual for a whole house scale system, but one 2s9p array per Delta Pro is what I need to achive my solar goals. (A pair of Delta Pros is required for 240V.)

The NEC minimum output circuit cable rating for 9p is the killer, which I'll cover later, but you already know the problem, given my Delta Pro based system.

The Ecoflow Delta Pro maximum solar input is 150V, 15A and 1600W. In order to get close to the 1600W max. (1600W is almost impossible), it is necessary to provide a Vin as close to 150V as the cold weather Voc temperature coefficent calculation (or alternate table or 1.25x "guardband") permits. I found very few panels featuring a Voc that, after cold weather minimum temperature are ideal for the Delta Pro. The SunPower E-Series: E20-330 Voc (64.9V) is pretty close to Ideal for 2s (129.8V before the cold weather adjustment for Idaho winters). So I don't have the luxury of a 600VOC limit with the Delta Pros.

If you do 2s9p (2 series, 9 parallel), then yeah that is far outside the norm for a house-scale system. If you want to configure a system like that NEC will require you to use something like AWG 1 conductors. The main reason I can think of to do something as wide as 9p is if you want to aggressively defend against shading; the standard NEC compliant approaches to this include wiring microinverters in parallel, or using optimizers to allow comparable shading restrictions while still using a series string.

I only briefly thought about microinverters but what I'm designing is really not a whole house system and I'll only use it and my house electrical panel interlock and 240V input to run essential items during power outages. Shade mostly falls simultaneously on all panels due to clouds or high terrain to the south when the sun is low. I'd rather use a petroleum fuel powered generator but, with solar, I'm planning for fuel shortages.

The SunPower panel Isc is 6.52A so the NEC minimum output circuit cable rating for 9p is 9 x 6.52 x 1.56 = 91.5A. Yikes! This requires a minimum 2, 3 or 4AWG wire, depending on temperature rating and 1AWG would probably be appropriate for a long run, if the wire ever actually had to carry 91.5A (I've not calculated the drop). Very large wires are expensive and difficult to mate to commonly available connectors and also difficult to route and install. The Delta Pro MPPT should not accept more than 15A in operation so why plan for 91.5A, except to deal with it as an electrical failure or fault?

My main motivation for placing panels on my shop roof instead of using ground mounts is the shop roof is the sunniest location with about 2 hours of direct sun per day on the winter solstice before the sun disappers behind the canyon rim above. On the other hand, clearing snow from panels is much easier with ground mounts. Not counting our very sunny three summer months, most days are cloudy here. Actually, a little more than half of the days are cloudy here, even when including summer months!

If you insist on 2s9p, you might be able to get away with putting the 9p combiner on the roof, with 9x 15A or 20A fuses on the input of the combiner and 20A fuses on the output.

From my measurements, I know that 2s9p is the only way to get the power I need during the non-summer months--at least with the Delta Pros. The SunPower E330 max. series fuse spec is 15A but if that specification is accurately published, max is max, and 15A was what I was thinking too. However, using the NEC forumla, 1.56 x Isc = 1.56 x 6.52A = 10.17A for the SunPower E330 so I could look for a 12A fuse for the combiner inputs. It would be good to have a larger fuse on the combiner output too--just in cast the MPPT fails and it draws more current than the specified Delta Pro 15A max. input. A 20A combiner output fuse seems appropriate (15A is 75% of 20A). Given the 15A input current limit of the Delta Pro, it will never draw more than 15A from the entire array during normal operation, so the above fuse values and an appropriate wire sizes for the 15A input and 20A output protection sounds safe to me.

BTW using a Delta Pro would not be NEC compliant unless that is 1741 or 9540 listed.

I have no idea if the Delta Pro is listed. I was looking at NEC, because panels on even a shop's metal rooftop is considered a solar "installation", I htink. As you mentioned, the Delta Pro is a portable integrated MPPT charge controller, LiFePo battery, and inverter "power station."

I originally bought the Delta Pros and 220W folding bifacial panels thinking of the portable benefits, but I realized that, with palet of 36 SunPower panels, I would have enough emergency power to run refrigeration, water, efficient lights and occasonally a few other items so I had a palet from Signature delivered.

If you used a regular inverter for a house-sized use case like 6000XP instead of the Delta Pro (which is for camping/small systems) you will not have these issues with those 330W panels. And installers are using at least this tier of equipment.

As a former academic (I published in computer science journals/conferences in likely comparable or higher tier as that paper), I would go out on a limb and say that particular paper (as well as 99.9% of other conference and journal papers), NEC doesn't give a crap about unless someone makes a lot of noise about it within the rules making committees.

I understand what you're saying about academia vs. government rulemaking institutions! I'm sure not going to make a lot of noise with NEC! I'm a retired electrical engineer. In grad school I was an analog circuits and systems guy but I got sucked into the digital "dark side" when I took a job with Intel Corp. and I spent a total of about 25 years there.

 

[zanydroid]

Make sure you use matching MC4 connectors when you build this.

The Ecoflow Delta Pro maximum solar input is 150V, 15A and 1600W. In order to get close to the 1600W max. (1600W is almost impossible), it is necessary to provide a Vin as close to 150V as the cold weather Voc temperature coefficent calculation (or alternate table or 1.25x "guardband") permits. I found very few panels featuring a Voc that, after cold weather minimum temperature are ideal for the Delta Pro. The SunPower E-Series: E20-330 Voc (64.9V) is pretty close to Ideal for 2s (129.8V before the cold weather adjustment for Idaho winters). So I don't have the luxury of a 600VOC limit with the Delta Pros.

Those panels are very unusual for residential use, and I hadn't factored that into my analysis. Usually residential 330W panels are 60s of solar cells, for somewhere in 30ish volts VOC. I guess the unusual config is why you got a good deal on the panels.

The SunPower panel Isc is 6.52A so the NEC minimum output circuit cable rating for 9p is 9 x 6.52 x 1.56 = 91.5A. Yikes! This requires a minimum 2, 3 or 4AWG wire, depending on temperature rating and 1AWG would probably be appropriate for a long run, if the wire ever actually had to carry 91.5A (I've not calculated the drop). Very large wires are expensive and difficult to mate to commonly available connectors and also difficult to route and install. The Delta Pro MPPT should not accept more than 15A in operation so why plan for 91.5A, except to deal with it as an electrical failure or fault?

You should be able to fuse the combiners input to 6.52 *1.56 = ~11A, and the output of the combiner to 15A. Then you can use the #14 cables that the solar panels come with (+ extension jumpers), and #14 or #12 to your inverter. I would recommend #10 or #8 because that is more future proof (depends on how easy it is to rerun the cables).

I believe this is code compliant and should not false-trip. I don't think any wires will melt. This does assume that the MPPT will not get into a weird state where it draws above 15A on its MPPT input (there are many more possible operating points with the same wattage when you overpanel).

That said, there might be some short impulse effects that a slower long timescale analysis will not catch. I attempted to discuss this here as a layperson:

Can I oversize 200%?? What would the implications be?

Hey Everyone, We have an AC coupled off-grid system in SEQLD Australia, though I am posting here as this question only really relates to the Fronius Primos, not the rest of the setup. Objective: Make more energy in the middle of winter by SIGNIFICANTLY oversizing the PV array, but staying...

diysolarforum.com

From my measurements, I know that 2s9p is the only way to get the power I need during the non-summer months--at least with the Delta Pros. The SunPower E330 max. series fuse spec is 15A but if that specification is accurately published, max is max, and 15A was what I was thinking too.

Max fuse refers to what the solar panel can tolerate when they are put in a parallel situation requiring fuses, IE it's the fault current they can tolerate, not what they can put out. NEC modeling says you have to factor for Isc * 1.25 (and 1.25 factor above that for continuous rating). If you look at the charts for the module they cannot possibly output 15A.

It would be good to have a larger fuse on the combiner output too--just in cast the MPPT fails and it draws more current than the specified Delta Pro 15A max. input. A 20A combiner output fuse seems appropriate (15A is 75% of 20A).

OK this is how I would build it, and also what NEC says you can do.

If you have trouble getting panels that can max out the 150V rating, you might consider using "hacked" SolarEdge optimizers to push panels to a specific output voltage. Optimizers have a regulated max output voltage and a regulated max current. So you don't have to do temperature compensation for voltage nor do 1.56 * output current, rather you can do just 1.25 (the extra 1.25 factor is for over STC weather conditions). Another potential advantage with optimizers is if there is a lot of shading / partial sun on the panels -- this type of optimizer will do current matching along the string along with MPPT on the input side.

So you can order SE optimizers hard coded to 49V output, and put 3 of them in series. With those high voltage solar panels, you could put 2P into one 400W SE optimizer and stay within their Isc limit.

I'm not sure if IndOp optimizers are certified to do something sane when operating in parallel

However, optimizers have an odd voltage curve, and they might confuse the MPPT of the Delta Pro.

Discussion I had recently on this, including a possible way to hack the power/voltage curve to make trick a confused MPPT to do the right thing again.

Solaredge optimizer IndOp experience requested

Thanks in advance...looking for some insight from anyone who runs optimizers (like the P320) on a non-SE inverter. They can be placed in a mode where they don't shutdown to 1V unless communicating with SE equipment, and in some cases you can buy used ones preconfigured to something like 37V. My...

diysolarforum.com

I understand what you're saying about academia vs. government rulemaking institutions! I'm sure not going to make a lot of noise with NEC! I'm a retired electrical engineer. In grad school I was an analog circuits and systems guy but I got sucked into the digital "dark side" when I took a job with Intel Corp. and I spent a total of about 25 years there.

Haha. I didn't read the paper much, but it seemed to be leading to discussing some plausible (for a grad student scenario) edge cases where the operating voltage is sustainably held at an usually low level. In my area, a lot of grad student scenarios are semi-fictional to make the paper work better (yet usually still logically consistent).

 

[zanydroid]

How are the bypass diodes in this panel arranged? You might be able to take advantage of that in the design, by aligning them properly against the dominant shading direction from the mountains etc. I think these are full cut so they work terribly with one shade orientation.

With half-cut they have more viable scenarios for horizontal and vertical shading (relative to the panel).

There are some pictures in here, and elsewhere (This compares shingled vs half cut [and I think they call half cut conventional which is kind of weird], but it might help you get the intuition. And the half cut pictures show how half cut has good properties in both orientations, they just don't show how much better they are than single cut)

https://sunpower.maxeon.com/au/sites/default/files/2024-04/COLLAT-71_Performance_Power_White_Paper_Feb22.pdf