Rapid Shutdown with SolarEdge Optimizers in 2:1 series module design

[solarken]

Does anyone have any experience with SolarEdge commercial optimizers used in a 2:1 series arrangment (2 modules in series connected to one optimizer) like for the P1101 optimizer? I am specifically trying to find out how that meets NEC 2017 Rapid shutdown requirements of 80V within the array boundary. I am looking at options for a commercial system using Q-Cells modules that have just under 60V open circuit voltage at low temperature, so two in series would be just under 120V. I cannot find anywhere in solaredge docs how the optimizer keeps the voltage below 80V when it is not producing (in shutdown) with this arrangement of 2:1. Has anyone measured the voltage across the optimizer inputs (i.e. across two series modules) while it is off to see if it limits the voltage somehow to under 80? I have a vm into Solaredge design team but they often don't respond very promptly. Thanks.

 

[ggunn]

That's a good question for which I have no answer. I have used P1101 optimizers in a lot of commercial PV systems; that question never occurred to me and it has never come up in plan review or inspection. Have you asked someone in SE tech support about it?

 

[solarken]

I found this in SolarEdge training academy. It states their series connected commercial optimizers do not meet NEC 2017 rapid shutdown unless two modules in series have less than 80V Voc when combined, which no commercial module would meet. This seems like a really big deal. Still waiting for call back from sales/design support.

 

[electro7]

The optimizers go to 1V when the inverter is turned off, from my experience. It doesn't matter if it's a 2:1 optimizer or 1:1. So if there is a string of 20 optimizers (40 panels) there would be 20VDC on that string when the inverter is turned off. I don't believe the Voc of the module matters.

Sent from my SM-G998U using Tapatalk

 

[ggunn]

Our SE rep says that the P1101 is RSD compliant with NEC 2020 690.12(B)(2)(1) under UL3741 as referenced on the data sheet.

 

[wwhitney]

I don't believe the Voc of the module matters.

I don't believe that's correct. The conductors between the optimizer and the panels are controlled conductors, and so if two of them will have >80V between them after shutdown, that would not comply with 690.12(B)(2)(2).

But per ggunn, it sounds like if you are under the 2020 NEC, it would alternatively comply with 690.12(B)(2)(1).

Cheers, Wayne

 

[jaggedben]

Does anyone have any experience with SolarEdge commercial optimizers used in a 2:1 series arrangment (2 modules in series connected to one optimizer) like for the P1101 optimizer? I am specifically trying to find out how that meets NEC 2017 Rapid shutdown requirements of 80V within the array boundary. I am looking at options for a commercial system using Q-Cells modules that have just under 60V open circuit voltage at low temperature, so two in series would be just under 120V. I cannot find anywhere in solaredge docs how the optimizer keeps the voltage below 80V when it is not producing (in shutdown) with this arrangement of 2:1. Has anyone measured the voltage across the optimizer inputs (i.e. across two series modules) while it is off to see if it limits the voltage somehow to under 80? I have a vm into Solaredge design team but they often don't respond very promptly. Thanks.

P1101 seems to have a 125V absolute max DC voltage rating. Is that 60V the STC Voc or the lowest temp Voc? Could be a problem if it's the former.

Our SE rep says that the P1101 is RSD compliant with NEC 2020 690.12(B)(2)(1) under UL3741 as referenced on the data sheet.

I find that bizarre but okay, I guess that's on UL.

 

[solarken]

P1101 seems to have a 125V absolute max DC voltage rating. Is that 60V the STC Voc or the lowest temp Voc? Could be a problem if it's the former.



I find that bizarre but okay, I guess that's on UL.

The module I am looking at has a Voc of about 53V at STC, and just under 60V at Extreme min temp, so the max DC rating is not a problem. The problem is that using the P1101 or S1201 that are designed for 2:1 appears to not meet NEC 2017 unless you go 1:1. These optimizers are almost $100 each.

 

[ggunn]

The module I am looking at has a Voc of about 53V at STC, and just under 60V at Extreme min temp, so the max DC rating is not a problem. The problem is that using the P1101 or S1201 that are designed for 2:1 appears to not meet NEC 2017 unless you go 1:1. These optimizers are almost $100 each.

You *might* be able to persuade an AHJ to accept that the 2020 NEC clarified an issue in the 2017 code and therefore a design with dual module P1101 optimizers should pass muster. Nothing in the hardware has changed; if it is safe now it was safe in 2017.

Thanks for the heart attack, by the way.

 

[Carultch]

At one point, it was not rapid shutdown compatible to use dual module optimizers that series the two modules together, except in the rare case that both modules are less than 40V at your location's ASHRAE low temperature. Around 2019 with the onset of module level rapid shutdown becoming a requirement, SolarEdge phased out their optimizers that connected two modules in series, in favor of paralleling optimizers to be 2017+ compatible.

Now SolarEdge is bringing back the optimizers for series connected modules, and phasing out the paralleling dual optimizers that they once had. Ask SolarEdge for documentation on this issue, to show how they are getting an exception the 80V limit. They clearly know it exists, and had paralleled the modules at one point to comply with it.

 

[jaggedben]

It seems like somehow UL was willing to sign off but if I were an AHJ I'd be quite skeptical.

 

[ggunn]

As long as it is code compliant with the UL listing it's good enough for me. My top priority is safety, and I do not believe that an isolated series pair of live modules is significantly less safe than a single isolated live module. The point of RSD is to not have strings of high DC voltage everywhere under the array and IMO the P1101 still accomplishes that.

 

[jaggedben]

Yeah, I get that, I just wonder how much higher UL's voltage cuttoff point is than the CMPs? At what point do you consider the voltage 'high'? Or is there more to it than that?

 

[wwhitney]

Or is there more to it than that?

You can read UL3741 online with a free account at ul.com. I looked at it a bit without trying to digest the whole thing. Apparently it involves modeling the resistance of a firefighters body and PPE and the likelihood of various ways of interacting with a PV array, so there's a lot to it.

Cheers, Wayne

 

[pv_n00b]

I've seen a couple of UL 3741 listed arrays that have the full 1,000V string voltage in the array. Under UL 3741 it's not the voltage that matters, it's the interaction between a firefighter and the PV array and what will be their possible exposure to a body current over a maximum value.

 

[pv_n00b]

You can read UL3741 online with a free account at ul.com. I looked at it a bit without trying to digest the whole thing. Apparently it involves modeling the resistance of a firefighters body and PPE and the likelihood of various ways of interacting with a PV array, so there's a lot to it.

UL 3741 is kind of like a choose your own adventure book. It's impossible to tell how a particular manufacturer navigated all of the choices to end up with a listed system unless you know how they went through the standard. But in the end, all that matters is that the system is listed and therefore NEC compliant.

 

[jaggedben]

I've seen a couple of UL 3741 listed arrays that have the full 1,000V string voltage in the array. ...

Do you mean to say that still applies after rapid shutdown is initiated?

 

[wwhitney]

UL 3741 is kind of like a choose your own adventure book. It's impossible to tell how a particular manufacturer navigated all of the choices to end up with a listed system unless you know how they went through the standard.

I didn't read the standard closely enough, but I'm surprised that the manufacturer isn't required to publish their analysis. It seems like it's just "trust us, we did a bunch of computations and it's fine."

Cheers, Wayne

 

[pv_n00b]

Do you mean to say that still applies after rapid shutdown is initiated?

Yes, for instance, the Sollega rack with the SMA Core1 inverter are UL 3741 listed and there is no reduction in the string voltage after rapid shutdown is initiated. There are an array of similar products that we will see getting UL 3741 listing in the future. The AC voltage from the inverter outside the array is still reduced to 30V or less in 30 sec.

 

[pv_n00b]

I didn't read the standard closely enough, but I'm surprised that the manufacturer isn't required to publish their analysis. It seems like it's just "trust us, we did a bunch of computations and it's fine."

Cheers, Wayne

The analysis and how they got through the standard to get listed is considered propriety information. I have not run into any manufacturer that is willing to share the test results, it's either UL listed or not and we don't get to know how the sausage was made.