6kW system design

[foo1bar]

One idea is to open stucco ...

Can you access it from the other side of the wall? (ex. cut drywall from the other side of the wall, then repair that, rather than break out and repair stucco)

standard EMT 90 degree elbow

You mean a factory-created 90 sweep?

Probably a drawing of what you're planning would make it easier for people to suggest better approach.
And/Or start a new thread with an appropriate title that might get more eyeballs of people who have exprience with fitting conduit in walls and dealing with stucco.

[sensij]

OP's approach sounds sort of OK if there is enough wall space for the sweep. If it is only a 2x4, that would be too small. I guess I'd ask why you don't just go straight through to the exterior at the top, where you have access? Running that length of conduit outside alongside the inverter won't look too bad if you paint it nice.

[max2k]

Can you access it from the other side of the wall? (ex. cut drywall from the other side of the wall, then repair that, rather than break out and repair stucco)


You mean a factory-created 90 sweep?

Probably a drawing of what you're planning would make it easier for people to suggest better approach.
And/Or start a new thread with an appropriate title that might get more eyeballs of people who have exprience with fitting conduit in walls and dealing with stucco.

Interesting idea, 'theoretically' I have access to the interior side of the wall except there's big shelf rack and moving that around would probably be more work than patching stucco. I have some questions for you regarding permit process so if you could find few minutes to answer them over email I'd really appreciate it: michael at hyper-plaza dot com.

[max2k]

OP's approach sounds sort of OK if there is enough wall space for the sweep. If it is only a 2x4, that would be too small. I guess I'd ask why you don't just go straight through to the exterior at the top, where you have access? Running that length of conduit outside alongside the inverter won't look too bad if you paint it nice.

That's what I wanted to know- standard 90 degree sweep wouldn't fit in there.

Each inverter has 3 conduits going to it (DC, AC & Ethernet) so overall I'd be looking at 6 of them coming from the top. I can have only 2 coming from the top (DC + Ethernet) and AC one going down with 3 junction boxes between inverters but that would look ugly IMO. Besides AC conduit going close to the ground level would need to pass across garage door opening on its way to MSP. This is shortest path for AC BTW but I haven't came up with a way to cross garage door as routing EMT around its top would look funny and burying it into concrete floor at the bottom of the door would require some hammering + rigid conduit if I'm reading NEC correctly.

How about using FMT in the walls while using EMT in the attic space (to simplify pulling vs FMT all the way)? I'd need to open stucco only below inverters in this case to fit in 90 degree 'pulling elbows' with FMT connectors while letting FMT bend to form that 90 degree turn inside the wall. I need to confirm it is allowed to run DC through FMT (roof penetration)- EMT(1st attic) - FMT(wall between attics) - EMT(2nd attic) - FMT(exterior wall cavity) combination. House has 2 separate attics and array is installed on top of the upper one so DC wires have some way to travel.

[foo1bar]

Each inverter has 3 conduits going to it (DC, AC & Ethernet)

I'd just do CAT5 cable zip-tied to the conduit for the ethernet. I don't think it's necessary to run separate conduit for ethernet.
But if you think it'll be better aesthetically, it's your house.

My ethernet cable comes out the crawl space, goes down a bit for a drip loop, then comes up ~4' to the inverter cabinet.
No conduit - but I did paint it same color as the wall.

How about using FMT in the walls

FMT or FMC?
Probably FMC - but FMT does exist.
I don't know the rules with either one.

I have another idea that you can think about.
What if you have a junction box (or multiple ones) embedded in the stucco?
You could bring conduit down to it inside the wall, then do conduits from that box/gutter to the inverters.

I think a gutter that extends under the two inverters and has a conduit in the center going up the outside to attic level would probably be OK looking.

burying it into concrete floor at the bottom of the door would require some hammering + rigid conduit if I'm reading NEC correctly.

I think you could do IMC under the concrete floor of a garage or under a driveway. I think under a driveway you could even do direct burial if you were deep enough for the cover rules (table 300.5)
I think for inverter's AC output you could do PVC conduit too. (again with appropriate cover for the conduit)

BUT my inclination to get from one side of garage to the other would be to run it in the ceiling or attic space above the garage.

[max2k]

I'd just do CAT5 cable zip-tied to the conduit for the ethernet. I don't think it's necessary to run separate conduit for ethernet.
But if you think it'll be better aesthetically, it's your house.

My ethernet cable comes out the crawl space, goes down a bit for a drip loop, then comes up ~4' to the inverter cabinet.
No conduit - but I did paint it same color as the wall.


FMT or FMC?
Probably FMC - but FMT does exist.
I don't know the rules with either one.

I have another idea that you can think about.
What if you have a junction box (or multiple ones) embedded in the stucco?
You could bring conduit down to it inside the wall, then do conduits from that box/gutter to the inverters.

I think a gutter that extends under the two inverters and has a conduit in the center going up the outside to attic level would probably be OK looking.


I think you could do IMC under the concrete floor of a garage or under a driveway. I think under a driveway you could even do direct burial if you were deep enough for the cover rules (table 300.5)
I think for inverter's AC output you could do PVC conduit too. (again with appropriate cover for the conduit)

BUT my inclination to get from one side of garage to the other would be to run it in the ceiling or attic space above the garage.

Thank you for the input, for now I'm going to go with attic ran conduits but I'll keep your suggestions as backup plan in case I hit a wall doing it.

[max2k]

I got the system built and test ran, it produced 6.5 kW around 1 PM on a sunny day. All is well from that point of view.

My current problem is I can't pass the inspection. I already went through 2 of them with 2 different inspectors and unfortunately they came up with different list of issues. So I'm looking for advise how to handle this and I'm well aware of the main 2 rules when it comes to dealing with AHJ so it needs to be realistic.

1st inspection:
1. All PV wire to be tied to rack or panel it cannot be supported by roof - fixed
2. All raceway must be exposed for inspection. - fixed: I had to partially remove drywall panel hiding some conduits. Second inspector didn't look there.
3. All warning stickers must be on - fixed
4. Open all boxes for inspection - fixed: opened AC combiner/disco, subpanel and inverters and it seems was satisfactory.
5. Open attic with ladder for inspection - fixed and the second inspector was fine looking through there.

Second inspection:
1. work not complete per approved plans (missing placard) - fixing now.
2. metal raceways with bonding bushings and jumpers required.
3. complete electrical wiring per approved plan.
4. smoke and carbon monoxide detector alarms don't comply.
5. need ladder for safe roof access.
6. equipment grounding conductors are not inside raceways. Here it means DC raceway only as AC one has EGC inside.
7. Subpanel below main is not on plans.

I have question about 2.2, 2.3, 2.6 as they all relate to the same issue- according to inspector EGC originates on the roof, where it bonds all racks and panels then gets spliced into stranded wire which goes along DC conductors inside the same conduit to inverter(s) and then connects to common ground inside MSP. He wants to see bonding bushings along the way as lock nuts + connectors are not acceptable for bonding.

Currently I have 6 AWG bare copper GEC coming from the roof >outside< conduits. DC conduit is bonded to that on the roof using bushing due to use of raintight connector but there's no EGC inside DC conduit coming down. DC conduit run ends in electrical box in the garage where wires from 2 strings go to each individual inverter, there's no intermediate boxes anywhere else. Can I just bond inverters EGC to that #6 AWG to satisfy inspector's requirement? It doesn't make sense to me as I built this under assumption that GEC is better to get to the grounding electrode as fast as possible and bonding is completely separate problem especially for ungrounded arrays.

Another question is about 2.4- is it normal to include items irrelevant to the current project into inspection scope to make them code complaint?

[J.P.M.]

I got the system built and test ran, it produced 6.5 kW around 1 PM on a sunny day. All is well from that point of view.

My current problem is I can't pass the inspection. I already went through 2 of them with 2 different inspectors and unfortunately they came up with different list of issues. So I'm looking for advise how to handle this and I'm well aware of the main 2 rules when it comes to dealing with AHJ so it needs to be realistic.

1st inspection:
1. All PV wire to be tied to rack or panel it cannot be supported by roof - fixed
2. All raceway must be exposed for inspection. - fixed: I had to partially remove drywall panel hiding some conduits. Second inspector didn't look there.
3. All warning stickers must be on - fixed
4. Open all boxes for inspection - fixed: opened AC combiner/disco, subpanel and inverters and it seems was satisfactory.
5. Open attic with ladder for inspection - fixed and the second inspector was fine looking through there.

Second inspection:
1. work not complete per approved plans (missing placard) - fixing now.
2. metal raceways with bonding bushings and jumpers required.
3. complete electrical wiring per approved plan.
4. smoke and carbon monoxide detector alarms don't comply.
5. need ladder for safe roof access.
6. equipment grounding conductors are not inside raceways. Here it means DC raceway only as AC one has EGC inside.
7. Subpanel below main is not on plans.

I have question about 2.2, 2.3, 2.6 as they all relate to the same issue- according to inspector EGC originates on the roof, where it bonds all racks and panels then gets spliced into stranded wire which goes along DC conductors inside the same conduit to inverter(s) and then connects to common ground inside MSP. He wants to see bonding bushings along the way as lock nuts + connectors are not acceptable for bonding.

Currently I have 6 AWG bare copper GEC coming from the roof >outside< conduits. DC conduit is bonded to that on the roof using bushing due to use of raintight connector but there's no EGC inside DC conduit coming down. DC conduit run ends in electrical box in the garage where wires from 2 strings go to each individual inverter, there's no intermediate boxes anywhere else. Can I just bond inverters EGC to that #6 AWG to satisfy inspector's requirement? It doesn't make sense to me as I built this under assumption that GEC is better to get to the grounding electrode as fast as possible and bonding is completely separate problem especially for ungrounded arrays.

Another question is about 2.4- is it normal to include items irrelevant to the current project into inspection scope to make them code complaint?

Following the two rules you cite to the letter: What anyone here thinks matters not one whit - as you already know. You're after clear direction, not unwinnable arguments.

I'd get the 2 inspectors and their supervisor in the same space at the same time, and politely but just as clearly and professionally say "W.T.F. guys ? Give me some help and direction in complying by providing clear, complete and unambiguous direction to what you all agree will result in a safe, to the code system." Add "Please" to the end of that sentence. If you still get J.O.'d you can start walking up the chain of command, but you'll do so with a record of cooperation on your side.

[max2k]

Following the two rules you cite to the letter: What anyone here thinks matters not one whit - as you already know. You're after clear direction, not unwinnable arguments.

I'd get the 2 inspectors and their supervisor in the same space at the same time, and politely but just as clearly and professionally say "W.T.F. guys ? Give me some help and direction in complying by providing clear, complete and unambiguous direction to what you all agree will result in a safe, to the code system." Add "Please" to the end of that sentence. If you still get J.O.'d you can start walking up the chain of command, but you'll do so with a record of cooperation on your side.

I'm not after fight with them at all as outcome is well known. It looks like my 2 inverter setup throws them off as they used to see single inverter systems. They were more or less consistent in asking to see EGC green wire inside DC conduit. Is it required by code to be in DC conduit for ungrounded array? Do I have leg to stand here in my use of #6AWG for GEC and EMT + lock nuts for bonding DC conduit or just pull green EGC as they expect and stop surprising inspectors?

[max2k]

I'm not after fight with them at all as outcome is well known. It looks like my 2 inverter setup throws them off as they used to see single inverter systems. They were more or less consistent in asking to see EGC green wire inside DC conduit. Is it required by code to be in DC conduit for ungrounded array? Do I have leg to stand here in my use of #6AWG for GEC and EMT + lock nuts for bonding DC conduit or just pull green EGC as they expect and stop surprising inspectors?

Answering my own questions: NEC 690.43 is what seems to be a problem. It clearly requires EGC to be run together with the rest of conductors so my installation is not compliant. Looks like all I have to do is to run green EGC inside DC conduit and put bushings/bonding jumpers through DC and AC paths as part of EGC from inverter to AC combiner grounding bus is also considered DC GEC per 690.47(C)(3) so bushings are required to mitigate choke effect. Since my AHJ is on 2014 cycle I don't have leg to stand (even without invoking 2 base rules) as they expect quite different arrangement regardless of what makes sense from electrical point of view.

[sensij]

Answering my own questions: NEC 690.43 is what seems to be a problem. It clearly requires EGC to be run together with the rest of conductors so my installation is not compliant. Looks like all I have to do is to run green EGC inside DC conduit and put bushings/bonding jumpers through DC and AC paths as part of EGC from inverter to AC combiner grounding bus is also considered DC GEC per 690.47(C)(3) so bushings are required to mitigate choke effect. Since my AHJ is on 2014 cycle I don't have leg to stand (even without invoking 2 base rules) as they expect quite different arrangement regardless of what makes sense from electrical point of view.

If your DC conductors are in continuous EMT all the way from the roof to the inverter, and your conduit work was done well, there should be an opportunity to suggest that the conduit itself is the EGC, if the 6 AWG bonded to the array can be bonded to the EMT on the roof. It sounds like you might have trouble selling that since it wasn't what was in your plans, though.

[max2k]

If your DC conductors are in continuous EMT all the way from the roof to the inverter, and your conduit work was done well, there should be an opportunity to suggest that the conduit itself is the EGC, if the 6 AWG bonded to the array can be bonded to the EMT on the roof. It sounds like you might have trouble selling that since it wasn't what was in your plans, though.

Yep, it didn't fly and partially as you pointed out - because it wasn't in 'my' plan or actually- in county's Standard Plan which comes as read only pdf.

Anyway, I now realize entire section 690.47(C) does not apply in my case as my inverters are transformerless and do not provide isolation between DC & AC parts. Basically the whole thing looks like ungrounded current source inverter switches between L1 and L2 at the frequency of 15-25 kHz. In case of ground faults on DC side it can easily detect that and stop switching. Since array is ungrounded the fact that one of its wires developed fault to raceway is not dangerous as there's no return path for someone get shocked.

Overall it looks like I just need to provide dedicated EGC for DC part of the system and not rely on the conduits themselves for that purpose.

[sensij]

The DC GEC is required on ungrounded, transformerless inverters in 2014 NEC. It runs from the inverter to the grounding electrode system (GES). Usually, that means running it to the MSP ground bar.

From the inverter to the roof, only an EGC is required.

[max2k]

The DC GEC is required on ungrounded, transformerless inverters in 2014 NEC. It runs from the inverter to the grounding electrode system (GES). Usually, that means running it to the MSP ground bar.

From the inverter to the roof, only an EGC is required.

with all due respect and I mean it I'm failing to locate part of NEC which requires it: IMO 690.47(C) doesn't apply as it starts from words
"Photovoltaic systems having dc circuits and ac circuits with no direct connection between the dc grounded conductor and ac grounded conductor shall have a dc grounding system."

For TL inverters they do have direct connection between DC and AC circuits including grounded conductor making it irrelevant from common sense point of view. OTOH coming from the guy who failed 2 inspections already it is more important how AHJ interprets it regardless what laws of physics say . I can try to present it this way for my next inspection to see what AHJ thinks about that. As I understand the consequence is quite profound- instead of running unspliced irreversibly split GEC from inverter to MSP I could simply run EGC with all needed splices along the way. All I have to provide is low impedance on higher frequencies by using grounding bushings on both ends of conduit runs for both DC and AC runs.

[sensij]

There is no direct connection on your inverters between the AC grounded conductor (neutral) and the DC grounded conductor because there is *no* DC grounded conductor. The only AC/DC connections are between the *ungrounded* conductors, L1/L2 on the AC side. Therefore, the DC grounding system is necessary.