diy solar

diy solar

Am looking for some electrical code help

LBen

New Member
Joined
Feb 6, 2022
Messages
59
I've been trying to get some help locally and even paid for a consult, but got nothing useful. So coming here to see if I can get some help.

The site is in Colorado and I'm trying to figure out the best way to do the grid interconnect and be code compliant. I've talked with the local state inspector and he says that he just inspects - very little guidance. This will be a grid tied whole house solar. Was thinking about going completely off-grid, but the rural co-op killed that idea

House was built in 2017 and was approved with a breaker service disconnect at the utility pedestal. The main service panel is inside the garage and the electrical service comes up out of the concrete floor and is behind a 2x4 formed wall cavity. The structural wall is 13" ICF wall (8" concrete and 2.5" foam on both faces). I have a limited amount of wall space on that wall due to windows, the main service panel is between two windows. The 2x4 wall cavity is 36" wide, but has a structural steel support post in the left most space between wall studs, just leaving a single 13" stud bay for any electrical pass throughs.

My current thought is that the SolArk15 will be mounted to the left of the main service panel on the inside wall. I'd have an 8x8x36 wire gutter below to do the wire runs inside. On the outside will be another 8x8x36 gutter for the wire runs outside. Looks like I need 3 pass throughs, a pair of 2.5" EMT and one 1" EMT to pass all of the wires from inside to outside and the other direction. Boring those holes through the ICF wall will be no fun. Outside I think I need a DPDT Xfer switch and I have to have a meter can for the rural co-ops production meter.

The PV is three strings using 10 awg wire plus I need something for an EGC - guessing I have to do 10 awg for that also? Was planning on using PVC conduit to bring PV wire into outside gutter. I currently have an 11kW Generac so those wires have to go through the wall also. The system will also have an LFP battery system positioned in the garage.

I have spent a bunch of time reading through NEC 2020 and have gotten bogged down as to what actually applies and under what conditions. It appears that the local state inspector does not adhere strictly to NEC 2020, which is good (acceptable batteries) and bad (not sure what he will allow). I've also spent hours and hours reading through threads on this forum gaining little bits and pieces of useful info for the system design in general - and I thank everyone for making this resource available.

So I'm hoping someone on this forum can help me out. Is what I've described viable : need more info, better way to do things, can't do it that way - here is how to do it, can't do it at all, or whatever.

Thanks,
Lowell
 
I've got some wiring/layout diagrams crudely drawn up so that I can visualize things.
 

Attachments

  • Solar - Grid Interconnect.pdf
    232.9 KB · Views: 62
Ben,
Your diagram doesn't make much sense to me. You've got a generator "box" through which you're running "PV".
I don't see the grid tie (is it "PM"?) - I do no see a transfer switch. The Sol-Ark isn't even in this diagram.
Block diagram is fine, but we need all the pieces... And then we're going to ask you for the wires.

Might START with Sol-Ark's diagram and adapt that.

Last I read code now requires metal conduit for all PV coming into and out of a structure. You can't use PVC.

Your local jurisdiction should be able to tell you which version of NEC they follow. Many are behind 2020.

The areas where I've done PV installs where they allow owner-installs, they're really just inspecting the interconnection and looking for proper location and labeling of AC and DC shut downs. Mine doesn't even allow rapid shut down (allowed by 2020 NEC) - they still require physical disconnects.

Have you done residential wiring before?
 
Several pictures are just equipment layout diagrams. No wiring details implied - of course PV doesn't go through the Generac box. The Inside Layout (page 2) diagram clearly shows the placement of the SolArk to the left of the load panel. Page 5 of my attachment shows wire sets that would go into the bottom of the Sol Ark (Batt, PV Com, Gen, Load, Grid). I also have the SolArk manual and I have used Diagram 3 on page 9 as the reference.

As I described in my text, the grid feed comes up out of the floor of the garage, will go into the outer gutter, then to the production meter and to the DPDT switch as shown in my Page 4. On page 4 and 5 of the attachment I've described which wires go where and I spec wire sizes as 4/0 Al (grid), show where I switch to 2/0 Cu and then the gen and PV are all 10 awg.

I mentioned in my original post that I was working with NEC 2020. That is what Colorado uses with variations.

I did not want to add a lot of extraneous information that was not salient to the help I needed. It is gonna be a ground mount 9.7 kW array with a 230' run of PV wire running 3 strings.

Where I am at in Colorado I have to pull a permit from the county (primarily the ground mount and batteries), the State (the whole installation) and apply to the rural electric co-op (application approval and inspection).

On your comment about the PV going into the house must be metal, not PVC - I need more detail. Everything I've seen on YouTube and related sites is going from the array to the structure with PVC. Are you saying I must run metal all the way from the array to the house. Or to transition from PVC to metal once the PV conduit comes out of the ground next to the structure? I was going to run an EGC wire from the array to the the metal parts on the house to ensure that everything was grounded together. From the outer gutter, the through the wall conduit, the inner gutter and the conduit to the SolArk will all be metal.

Yep, I'm familiar with residential wiring. Did a pool installation, rewired a garage and did the final electrical in a 5000 sqft house that we live in and lots of other small things over the years. Also way back in the distant past, I worked as an instrument/avionics tech on heavy jets (B-52H, KC-135A) aircraft in the Air Force. Then I was a "rocket scientist" in the manned space program for 32 years - which has absolutely no bearing here.

Thanks for your reply. All constructive input is welcome.

Lowell
 
I mentioned in my original post that I was working with NEC 2020. That is what Colorado uses with variations.
"It appears that the local state inspector does not adhere strictly to NEC 2020,"
Is the statement that made it unclear to me. You're asking for help with code. I'm trying to help with code. In my state, the state does not set what standard the local authority follows, so it can vary. Thank you for clearing up that it's 2020.

Where I am at in Colorado I have to pull a permit from the county (primarily the ground mount and batteries), the State (the whole installation) and apply to the rural electric co-op (application approval and inspection).
Is it the co-op doing the inspection? Because again, just because the state says 2020 NEC, the local co-op can ignore that. They may have their own standards. I could not get rapid shut down allowed, even though it's 2020 NEC. They simply said "no".

Here the inspections by the co-op are very trivial. Disconnects, labels, nothing obviously wrong.

On your comment about the PV going into the house must be metal, not PVC - I need more detail. Everything I've seen on YouTube and related sites is going from the array to the structure with PVC. Are you saying I must run metal all the way from the array to the house. Or to transition from PVC to metal once the PV conduit comes out of the ground next to the structure? I was going to run an EGC wire from the array to the the metal parts on the house to ensure that everything was grounded together. From the outer gutter, the through the wall conduit, the inner gutter and the conduit to the SolArk will all be metal.
NEC Section 690.31(D).


Apparently it's for "improved arc protection" with high voltage DC. I've always used PVC also, but it's changed.


Yep, I'm familiar with residential wiring. Did a pool installation, rewired a garage and did the final electrical in a 5000 sqft house that we live in and lots of other small things over the years. Also way back in the distant past, I worked as an instrument/avionics tech on heavy jets (B-52H, KC-135A) aircraft in the Air Force. Then I was a "rocket scientist" in the manned space program for 32 years - which has absolutely no bearing here.
Just making sure, no issues with your experience.. I explain it different if you've never done residential electrical.
I'm an EE. Trust me, I get it when you say "it has no bearing here" - totally agree.

For me to help more, I need wiring diagrams. I haven't seen a lot of "code help" here, but I know of a great place where they will help you with code (lots of master electricians) - but you'll need to do diagrams.
 
I could not get rapid shut down allowed, even though it's 2020 NEC. They simply said "no".

I would hope RSD is allowed, just not sufficient for disconnect. In fact, a visible blade disconnect could trigger RSD. It does for GT PV, taking away the keep-alive. I'm planning a 3-pole disconnect so 3rd pole can carry power to keep-alive from my battery inverter. That way one switch is disconnect and RSD.

Just be sure PV panel to RSD is actually compatible, listed pair of MC4 connectors mated. Otherwise, it will earn its keep by stopping the fire it starts.

Apparently it's for "improved arc protection" with high voltage DC. I've always used PVC also, but it's changed.

I'm inclined to use rigid/IMC not EMT for PV in building, more burn-through resistance.
All mine is outdoors for now, but I'm going to do a PV run to my lab for bench testing.

Then I was a "rocket scientist" in the manned space program for 32 years - which has absolutely no bearing here.

One time "Aerospace Engineer" aka "Rocket Scientist" here too.
Normally I'm called EE of some form or another.
 
Hey guys, thanks for the discussion. Can't do a good reply right now. Working on tractor hydraulics out in the rain. Gotta get that done before dark.
 
I would hope RSD is allowed, just not sufficient for disconnect. In fact, a visible blade disconnect could trigger RSD. It does for GT PV, taking away the keep-alive. I'm planning a 3-pole disconnect so 3rd pole can carry power to keep-alive from my battery inverter. That way one switch is disconnect and RSD.
Just be sure PV panel to RSD is actually compatible, listed pair of MC4 connectors mated. Otherwise, it will earn its keep by stopping the fire it starts.
You have it right, the PoCo said "not sufficient" for disconnect despite NEC. Whatever, bungholes.
The reality here is that most electricians do not want to do PV - they're too busy with residential and what they need is "help" (labor) not to learn something new. The PoCo guys that inspect, they're been doing this for 10+ years, they're mainly focused on people installing platforms for their transformers, etc. They do concrete inspections... They're not electricians. I could have had a big box full of canarys and as long is it was labelled "solar disconnect" and had an arm, they would pass it.

I don't thing the PV panels need to be rated RSD compatible (I may be wrong) the RSD is basically at the "combinder" and it's got signal wires going to the inverter. If the inverter stops talking to the RSD, it "fails open".

One thing I learned with generators is that you can run generator signal wire in the same conduit as PV, but it needs to be V rated.
 
The utility should care, for lineman safety. Visible blade switch can be trusted, transistors can not.
PG&E now says disconnect switch is optional. But hey reserve the right to unplug meter and leave all power to house off.

Earlier version RSD was just that voltage had to be low > 12" from array. A switch with extension handle could do that. Or inverter at array.

New RSD is no wires can be > 80V from any other wires, even within array.
So for new version, box per PV panel (or several panels) is required. They have MC4 connectors, but mixed brand connectors caused fires so now only same model or UL listed pairs of models are allowed (not just for RSD, for all connections.)
 
The utility should care, for lineman safety. Visible blade switch can be trusted, transistors can not.
I agree. That's why these are designed to "fail open".

So for new version, box per PV panel (or several panels) is required.
Seriously? My RSD came with the inverter, which makes sense but it's rated for about 400V DC. You're saying that I need a RSD PER panel now? I get it for micro-inverters, but this would surprise me. Again, I learn something new every day.
 
Something like NEC 2010 (?) was RSD no voltage outside array.
NEC 2017 (?) was RSD no wires more than 80V from any other. Probably paid for with donation from Enphase.

See which applies at this time in your state. You would want RSD boxes compatible with the keep-alive built in to your inverter. Or, some like SolArk close a relay to power external keep-alive transmitter. The inverter is usually expected to do the job of discharging its input capacitor within 30 seconds.

The deal is, a fireman might spray water on a fire or put a pickaxe through something and get hit with nearly 600VDC.



Arc-fault may not always work. I would like RSD to also interrupt string in series when AFCI triggers, but not sure how the implementations have been done.
 
NEC 2017 (?) was RSD no wires more than 80V from any other. Probably paid for with donation from Enphase.
Clarification, you mean the RSD can't have more than 80V "voltage drop" away?


The deal is, a fireman might spray water on a fire or put a pickaxe through something and get hit with nearly 600VDC.
But this is the same as without RSD. My DC disconnect has to be "readily accessible" by the fire-people. That's on the side of the shop. Upstream of that, you've still got 400V DC.
 
Seriously? My RSD came with the inverter, which makes sense but it's rated for about 400V DC. You're saying that I need a RSD PER panel now? I get it for micro-inverters, but this would surprise me. Again, I learn something new every day.
NEC 2017 requires module level in most residential rooftop cases. Previously it was array level.

Inverter can’t do RSD, it only transmits the keep alive. One of the tough things about faults from solar panels is that they’re always generating power and voltage in the sun. Similarly inverter can’t do as robust ground fault protection or arc fault protection as a breaker in a panel, since it’s in the wrong place to turn off the power by itself. Maybe series fault, but parallel fault?? Best it can do is work with some MLPE or detect the issue and raise a loud alarm/shutdown to safe mode.

Actually microinverters are exempt by how the rules are written. Other exemptions are there for solar roof tiles etc. basically cases where you don’t have arbitrarily long strings of TC-ER or PV at arbitrary voltages.

How can you guarantee that a FET fails open? Even a design with lower chance to fail short is probably not as good as a mechanical switch with enough clearances. It is also harder (impossible?) to visually confirm disconnect with RSD since it involves invisible radio / packet signals talking to hidden FETs up in the roof, vs a visual indicator on a physical switch that has been manipulated.
 
Clarification, you mean the RSD can't have more than 80V "voltage drop" away?

Panel Voc no greater than 80V

With keep-alive, RSD connects that 80V from PV panel to RSD output.
When keep-alive goes away, RSD disconnects PV. Or rather, puts out maybe 1V on its output, probably has a buck converter to power its electronics.

If you grabbed any two wires, one panel's PV + and it's PV- or any other two wires, they would be < 80V apart.
(Hmm, that doesn't quite comply with 11 panels worth of RSD each outputting 1.0V and last PV panel putting out 80V to its PV+. Isolated step down converter and both PV+ & PV- disconnected rather than just PV+?)

SolarEdge has its own proprietary optimizers, and they put out 1V each so voltage of string tells you how many panels in series. Being optimizers with buck converter, they allow more panels in series while limiting Voc actually seen by inverter.

But this is the same as without RSD. My DC disconnect has to be "readily accessible" by the fire-people. That's on the side of the shop. Upstream of that, you've still got 400V DC.

With per-panel RSD, no 400V anywhere. Unlike simply a DC disconnect.

RSD is required on building rooftop, and for PV which has DC wires in or on a building. Exception if building is a power shed.
I would think simple disconnect would be sufficient for ground-mount PV array and inverter on building (so long as inverter discharges input voltage)
Putting inverter on a post 1' away from building and running AC to building should eliminate that last requirement.
 
Ah yes the inverter is responsible for draining the capacitance to manage the string voltage up to the first RSD disconnecting device within the time period required by code. After that there are no big capacitors.
 
From the outer gutter, the through the wall conduit, the inner gutter and the conduit to the SolArk will all be metal.
I don’t think the whole run needs to be metal. I thought it was just the part inside the building. This is the section I’m aware of


Note my understanding is that different AHJ/HOA in California have different requirements for exterior PVC vs EMT. I vaguely recall people saying that some prefer one while others prefer the other. The installations I’ve driven past/personally know of in other towns use EMT on the roof and exterior
 
Panel Voc no greater than 80V
Thanks.

With per-panel RSD, no 400V anywhere. Unlike simply a DC disconnect.
But it's got to increase panel cost if you need a syncing RSD per panel.
RSD is required on building rooftop, and for PV which has DC wires in or on a building. Exception if building is a power shed.
By whom? NEC? Back when I started there was no such thing as RSD.
For me, RSD showed up around 2019/2020 on a Frontius inverter. And again, the POC said "doesn't count". So I've got RSD on the roof and 2 disconnects (AC/DC).

I would think simple disconnect would be sufficient for ground-mount PV array and inverter on building (so long as inverter discharges input voltage)
Putting inverter on a post 1' away from building and running AC to building should eliminate that last requirement.
This makes sense. Most of my systems have been ground mount (simply as I don't like being on roofs and do not want the liability of a roof penetration).
 
But it's got to increase panel cost if you need a syncing RSD per panel.

Yes, around $35 to $50 per RSD. I bought Tigo TS-R-F which is RSD shutdown only and has two pairs of PV inputs. The panels I'm going to use can have two in series for the voltage limit, so I'll have 4 panels total per RSD, cost of $11 per panel. Not so bad!


I bought this or similar keep-alive transmitter, planning to use with 30kW SMA TriPower (which only implements discharge, nothing else.)
But PG&E informed me that inverter doesn't have the grid support features required since June 2020.
So I changed my plan to Sunny Boy 7.7kW -41. That does have RSD transmitter, but apparently for different brand RSD, so I'll use what I've got.


By whom? NEC? Back when I started there was no such thing as RSD.

Two different schemes required in two revisions of NEC.
Have to see what revision is currently used by your state. Make use of exceptions, maybe.
My existing system predates RSD and AFCI, just has GFCI. New house and system will meet latest codes.

For me, RSD showed up around 2019/2020 on a Frontius inverter. And again, the POC said "doesn't count". So I've got RSD on the roof and 2 disconnects (AC/DC).

This makes sense. Most of my systems have been ground mount (simply as I don't like being on roofs and do not want the liability of a roof penetration).

I'm going to be very careful and deliberate with flashed standoffs lag screwed into rafters. Keeping water out is more important than any power savings.
 
Two different schemes required in two revisions of NEC.
I'm aware of NEC 2020 690.12 as I read it there are 3 compliant options:
  • Install a PV hazard control system listed for the purpose. (I believe this is what I have, an all-in-one RSD)
1683864446010.png
  • Limit the controlled conductors inside the array boundary to not more than 80 volts within 30 seconds of rapid shutdown initiation. (this is the 80V limit per limit that you're referencing, it's not one needed PER panel, it's one needed per panels in series producing less than 80V Voc - as you mentioned above)
  • Install the PV arrays so that they have no exposed wiring methods or conductive parts and install them more than 8 feet from exposed grounded conductive parts or ground. (this seems really difficult unless your array is "ground mount" - insulated, and at least 8' up)
Have to see what revision is currently used by your state. Make use of exceptions, maybe.
I'm good. It's the POC that sets our standards, they're on some old ass version of NEC and frankly they don't even look at the array.

Thank you for your reply. I like staying current.
 
Ok, you guys have kinda taken over my thread and I'm now more out in the weeds than when I started. I also need to figure out the quote function that every one is using.

First, PV and PVC - After reading and re-reading 690.31(D) it appears to me that it is legal to run PVC to the outside gutter and then the PV wire is in metal from the outer gutter, through the EMT, through the inner gutter and then into the SolArk. 690.31(D)'s paragraph title says "on or in a building" but the text only talks about "in" a building. It would appear to me that the "on" part only refers to roof mount systems. Mine will be a ground mount.

RSD - I thought that I had this one packed away a long time ago. Thought that I had read somewhere that ground mount systems did NOT require RSD. I've read 690.12 and that text is as clear as mud to me. Am I correct in believing for a ground mount array that an RSD is not required? In any case I plan on including a manual disconnect at the array. And then the SolArk has a PV disconnect switch on it and the PV wire enters the wall less than 3' from the SolArk. So I think I'm covered on the disconnect requirement?

As dcg9381 had queried about, Colorado adheres to NEC 2020, but the local state inspector (when he was willing to talk to me) said that certain aspects are not enforced. For example on the batteries, NEC 2020 says any single ESS cannot be more than 20kWh and I think that there is a total limit of 80 kWh and the individual ESS must be at least 30" apart. And then there is the 9450 or is it 9350 that says that the inverter/charger must be certified to work with a matched ESS. He said that he only looks for an NRTLS cert on the batts. I had also asked him if it was ok to put the SolArk to the side of the service panel. He said yes as long as there was access to the service panel. And then at that point he said that he only inspects and can't recommend anything. So my thought process is that if I understand how things must be done to be compliant with NEC 2020, then I will work to that goal. If I come up against an issue where there is a conflict, then I'll see if there is a work around.

The rural co-op rep told me that they have their own inspection criteria. He said that if the state passed the installation, that MOST LIKELY they would also pass it. I live in a rural, technically backwards county. I'd classify them as at least 10 years behind other places. The rural co-op does not like residential solar, but need the offsets for the state. They are into commercial solar that they then resell to the consumer. This has been an uphill battle. For the county inspector, I had to show proof that LFP batteries were relatively safe, did not require a vented cabinet and did not require a drain.

I had thought that I had reduced my list of issues/questions to just figuring out how to interface with the grid: gutter sizes, types of splices or power distribution blocks, wall pass throughs, most efficient way to wire the xfer switch and production meter, conduit fill, EGC, etc. But I think I'm doing a circle back on things I thought I had understood.

I'm gonna be out of pocket for the next couple of days, so may not be able to reply to anything during that time. I thank you all for your observations.
 
Back
Top