1ft array rule question

[maverick69]

Doing some brainstorming and wanted a sanity check.

I have a very tall house making maintenance a concern. I am looking a string inverter system with panel optimizers.

Would it be feasible to route panel wiring into the attic to optimizers mounted on the interior of the roof to allow for easy access in the event of a component failure?

I think these could be mounted to within the 1 foot array boundary rule. What else am I missing? Or should I be concerned with?

Is my concern about optimizer failure irrational?

 

[Solar Guppy]

That would not be allowed ( entrance through roof ) without it being totally contained within a fire-proof methods, like metal conduit or armored cable, and even if allowed, your looking at a hole for each panel, terrible idea for a roof.

As for lifespan, I would not want optimizers or micro-inverters for the reason your concerned about, the solution is the string solution, which is the lowest cost option. With 600V string inverters you can get ~10 panels in a single string, for code you would need RSD per panel if NEC2020 or greater or NEC2017 an RSD per string.

 

[maverick69]

That would not be allowed ( entrance through roof ) without it being totally contained within a fire-proof methods, like metal conduit or armored cable, and even if allowed, your looking at a hole for each panel, terrible idea for a roof.

As for lifespan, I would not want optimizers or micro-inverters for the reason your concerned about, the solution is the string solution, which is the lowest cost option. With 600V string inverters you can get ~10 panels in a single string, for code you would need RSD per panel if NEC2020 or greater or NEC2017 an RSD per string.

Thanks for the input. The optimizers would act as the panel level RSD as well. Was trying to devise a solution that could allow easy access to the rsd devices. It may not be possible. Can RSD devices be clustered. Ie place them under more accessible panels? Obviously this would extend some cable runs.

 

[Solar Guppy]

Thanks for the input. The optimizers would act as the panel level RSD as well. Was trying to devise a solution that could allow easy access to the rsd devices. It may not be possible. Can RSD devices be clustered. Ie place them under more accessible panels? Obviously this would extend some cable runs.

If NEC2020 or newer for your permit the answer is no. If NEC2017, can be within the array boundary ( hence a single RSD per string or ~10 panels ) at the edge of the array

 

[zanydroid]

If NEC2020 or newer for your permit the answer is no. If NEC2017, can be within the array boundary ( hence a single RSD per string or ~10 panels ) at the edge of the array

Disagree.

IMO you can cluster MLPE/RSD no problem. The voltage within the boundary remains <60V (or was it 80V) at all points when shutdown is triggered (RSD) and all the time (Microinverters). Regardless of what crazy wiring you use to get from the panel to the MLPE/RSD.

I put my multi-port microinverters at the array edge per this understanding of code. And anyway multiport microinverters are mathematically impossible to mount under all solar panels that they service. It can only be under one and MAYBE two. Four, no way.

Now, you MAY run into trouble with the installation instructions if all the pictures in the manual show them pre-installed directly under the panel, and technically code requires you to follow all installation instructions.

 

[smoked_g60]

I might look into the new UL 3741 hazard control listing that Ironridge has released with Tesla Inverters along with the Mid Circuit interrupter (MCI-1). Tesla also has a similar listing but it uses more of the mid circuit interrupters (MCI-1 or MCI-2).

An Ironridge system with one large continuous array separated by no more than 2' would need just 2 MCI per string, and those could be located near the edge of the array.

 

[zanydroid]

I might look into the new UL 3741 hazard control listing that Ironridge has released with Tesla Inverters along with the Mid Circuit interrupter (MCI-1). Tesla also has a similar listing but it uses more of the mid circuit interrupters (MCI-1 or MCI-2).

An Ironridge system with one large continuous array separated by no more than 2' would need just 2 MCI per string, and those could be located near the edge of the array.

Interesting. Is the idea here to make the array qualify for the "no exposed wiring" exemption in NEC?

What is the value prop vs just using RSDs? Is the extra IronRidge mechanical hardware actually going to be cheaper than RSDs? Call me skeptical. Objectively there would be fewer points of failure without RSDs.

(Also, flush mount? not that many DIYs using it. And flush mount might produce less on a hot day with the worse ventilation. There are already people that do not light trim kits for this reason)

 

[smoked_g60]

UL 3741 is called out under 690.12.B.2 as one of the options for controlled conductors. You follow the installation manual and in this case the 3741 Hazard Control Addendum.

You do not need anything different mechanically, but need (2) Tesla MCI-1 per string within a discrete subarray separated by more than 2'.

There is also a path to UL 3741 compliance with the Tesla Inverter and any other racking, but that needs one MCI per 3 modules or fraction of 3 modules within a string.

When module level RSD is otherwise required, the MCI solution is quite a bit cheaper.

For the average roof and the average persons needs, flush mount is perfect for your pitched and ok for your flat roof depending on snow and lattitude. With module prices so low, the specific production of each module needs to be weighed against mounting on a tilt, and the additional roof space that will require. YMMV

 

[zanydroid]

UL 3741 is called out under 690.12.B.2 as one of the options for controlled conductors. You follow the installation manual and in this case the 3741 Hazard Control Addendum.

You do not need anything different mechanically, but need (2) Tesla MCI-1 per string within a discrete subarray separated by more than 2'.

There is also a path to UL 3741 compliance with the Tesla Inverter and any other racking, but that needs one MCI per 3 modules or fraction of 3 modules within a string.

When module level RSD is otherwise required, the MCI solution is quite a bit cheaper.

For the average roof and the average persons needs, flush mount is perfect for your pitched and ok for your flat roof depending on snow and lattitude. With module prices so low, the specific production of each module needs to be weighed against mounting on a tilt, and the additional roof space that will require. YMMV

OK, I get it now.

What does UL3741 product promise on top of regular RSD, that allows for softer rules? IE 600V limit within array with IronRidge and 120V ish with non IronRidge?

Is there sunspec work to provide a non-Tesla option?

The way the MCI is in series with string instead of parallel to one module may make inverter compatibility harder.