I can understand the NEC-2014 requirement to get the PV wire homerun under 30 volts for firefighers (within 30 seconds of shut down). Less-experienced poorly-supervised Mike could have would have done his attic wire run a foot down from the ridge... a nasty DC arc flash for a firefighter ventilating vertically with a chainsaw.
But the NEC-2017 requirement to bring wires within the PV-array boundary under 80 volts within 30 seconds (module-level rapid shutdown)... in the real world, this doesn't meaningfully support firefighter safety (feedback welcome: mike@slowpower.wtf). Firefighters treat PV arrays as energized, and stay away from them (let alone cut thru them). If a firefighter fell onto an energized PV array... the module frames and racking are bonded to equipment ground and the module glass isn't conductive. Wires are underneath the modules or in EMT between arrays, and they're thickly insulated ("PV wire"). Firefighter gear is very insulated. Let's get real: If a firefighter falls onto a PV array... her/his problem is falling — not getting electrocuted.
When I got into the industry in 2007, the two leading inverter companies were SMA and Fronius. They offered war-horse string inverters. Enphase and SolarEdge started offering module-level power electronics (MLPEs) in 2008 and 2010 (respectively)... boasting module-by-module monitoring and better shade management. And Tigo in 2007, for use with standard string inverters from other companies.
With the NEC-2017 requirement, Enphase and SolarEdge took over the residential market. There was (what I assumed was) an industry conspiracy that Enphase, SolarEdge, and Tigo were behind the NEC-2017 revision. But — as shown in derek the solarboi's excellent video — they were straight-up three out of the four PV-industry companies that submitted the formal proposal.
Enphase has worked hard to make their micro-inverters reliable. With that said... As argued by solarboi, a bunch of electronics in a high-temperature-swing environment... this is a common source of PV rooftop failure. Twice as many electrical connections... common source of failure. I've replaced failed MLPEs but I've never needed to replace a failed module.
Adding to the trouble: With MLPEs installed, the string's native open-circuit voltage can't be confirmed during commissioning. I'm personally aware of a related fire.
Module-level electronics have caused building fires. They've increased the cost, complexity, and liability of PV installations.
With the adoption of NEC 2020, there's a return-to-reasonable approach. If the racking + module + inverter + string-level rapid-shutdown device + wire management altogether have the UL 3741 listing (as an assembly)... MLPEs aren't required. Just bring the wire homerun voltage below 30 volts within 30 seconds (like it was with NEC 2014).
IronRidge, for example, has an option (that isn't Tesla)... A Solis inverter + MidNite Solar's 💪 MNSSR-600S-SS string-level rapid-shutdown receiver.
The wire-management requirement tl;dr: Use approved wire clips / ties and do it in a workmanlike manner (no dangling wires, bro).
Solis is a Chinese company with associated data-privacy risks. I'd just establish an internet connection temporarily for commissioning (if the system owner agrees to keep it offline). But I assume other residential-inverter companies will follow suit (or already offer residential UL 3741 options with other racking companies).
MidNite's rapid-shutdown product came out ~ten years ago, when modules weren't so huge. ...Its maximum-rated string voltage is 600. Its maximum short-circuit current is 12 and maximum MPPT current is 10 (according to IronRidge documentation). This rules out most year-2025 400+ watt modules. But compatible high-power options exist. For example, this 450-470 watt series from REC. Alas, the unusually low current comes with an unusually high voltage... each string could only have eight or nine modules.
Whether from MidNite or other companies... I assume we'll see more higher-rated string-level options now that a majority of states have adopted NEC 2020 (solarboi shows an option in this short video).
For an 18-module system, my cost for two MidNite rapid-shutdown devices would be around $260. For 18 module-level rapid-shutdown devices... around $700. Plus extra labor time with the latter. Plus extra liability cost around increased likelihood of future failure (more rooftop electronic devices = more liability).
Generally speaking, fuck third-party ownership (TPO). To compete with the continued TPO 30% tax credit, I need to find ways to make cash-purchased (or local-credit-union loan purchased) PV systems less expensive. Moving to string-level rapid shutdown is one of the ways I plan to do this.
COPY-PASTE FOOTER STUFF:
Not on my once-a-month email-newsletter list, and want to be? Click here.
On it, and don't want to be? Click here.
This blog and AI: I use ChatGPT all the time, to understand things better. But I'll NEVER peddle AI writing as my own. These blog posts are 100% written by Mike (for better or worse).
This blog is managed thru cabin