Solar edge string fuses

hhsting

Senior Member
Solaredge is so confusing. I have solaredge inverter SE9K-US, 3 strings, each string 13 modules, each module model TSM 300DD has optimizer model P320. The circuits from each string go straight to inverter their is no combiner. Are string fuses required?
 
Last edited:

jaggedben

Senior Member
SolarEdge says they are no, but this may be somewhat contradicted by NEC requirements. I believe SolarEdge can provide a fuse kit if your AHJ requires it.
 

PWDickerson

Senior Member
Location
Clinton, WA
Have you considered wiring the array with just two strings? Which optimizer are you using? Have you used the SolarEdge online Designer tool? I designed a very similar system earlier this year using a couple of the 9 kW 3ph inverters, and I was not able to make the stringing work with the dual optimizers. Tech support gave me some tips that weren't in the manual on combining different optimizers on the same string. If you can make two strings fly, it will make the project cheaper and easier to do, and it will solve your fusing problem.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Solaredge is so confusing. I have solaredge inverter SE9K-US, 3 strings, each string 13 modules, each module model TSM 300DD has optimizer model P320. The circuits from each string go straight to inverter their is no combiner. Are string fuses required?
Some AHJ's will take SE's word for it and some will not. You can put 5250WDC (5700WDC if you are using a SE HD inverter) on a string; do you really need 3 strings?
 

Carultch

Senior Member
Location
Massachusetts
Some AHJ's will take SE's word for it and some will not. You can put 5250WDC (5700WDC if you are using a SE HD inverter) on a string; do you really need 3 strings?
It's very frustrating, because at one point, SE had string fusing come standard on all the inverters that anticipate 3 strings. Then they reverted to no fusing and bare terminal block connections. How is an installer supposed to anticipate whether an AHJ will accept a white paper from SE on this issue?
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
It's very frustrating, because at one point, SE had string fusing come standard on all the inverters that anticipate 3 strings. Then they reverted to no fusing and bare terminal block connections. How is an installer supposed to anticipate whether an AHJ will accept a white paper from SE on this issue?
Ask the AHJ before you build a system.
 

pv_n00b

Senior Member
Location
CA, USA
As the technology gets more complex it gets harder to apply the simple NEC rules of thumb to PV systems. There are really only two issues with the lack of OCPD in the SE inverter.

1) Will the module series current rating be exceeded if there is an internal fault in the module?
Modules have a maximum series fuse rating to limit fault current going back into the module if there is an internal fault. If there is a DC optimizer on the module and it does not allow reverse current flow into a module fault then it does not matter how many strings are in parallel, the module will never see a fault current greater than the series rating. SE though has never definitively stated that the DC optimizer will not allow reverse current flow, at least not anywhere I have seen. If the DC optimizer allows reverse current flow then the strings have the same OCPD requirements as a string with no DC optimizer. SE implies that reverse current flow can't happen but if anyone has read anything from them that specifically says it won't please post a link.

2) Are the string conductors rated for the maximum fault current?
This is pretty basic, assuming a fault on a string conductor between the last DC optimizer and the inverter, is the conductor rated to carry the maximum current from N-1 strings into the fault. Most of the time with a #10 conductor you can have quite a few strings in parallel before you hit the rating.
 

SunFish

Member
Location
ID
Ugh...SolarEdge has frustrated me with this as well. I wish they would just bring back the fused inputs with their three phase inverters. But, you can in fact combine three strings together with solaredge without fusing but it requires some additional work.

Here is a link to a white paper from solaredge on the subject of combining three strings: https://www.solaredge.com/sites/default/files/string_fusing_requirements.pdf

Unfortunately, in my opinion, what it comes down to is if you want to combine three strings together at the inverter you probably need to run number 8 from the wire transition box to the inverter. The optimizers won't pass through fault current from the other two strings so #10 wire is ok on the PV module side of the circuit. And #10 homeruns from the wire transition box to the DC optimizers is in free air so it's rated 55 A. But if you have a fault on the output side of the optimizers I dont think #10 is adequate from the wire transition box to the inverter because you would have 6 current carrying conductors in a raceway, which is a 0.8 derate and you will probably also be applying an ambient temperature derate which would push the current rating of a #10 THWN-2 conductor down below 30 A which is the continuous output current of two strings of solaredge.

So just to be safe when I'm doing three strings without fuses I'm doing #8 AWG from the wire transition box to the inverter. I wish solaredge would just bring back the fused inputs. They claim this is saving us $ but in reality the cost of upsizing to #8 wire probably ends up costing more than the fuses and fuseholder inputs. They also claim you can order fused inputs aftermarket and install them yourself but I've tried and nobody is even supplying these.

SolarEdge also has a letter from Bill Brooks on this subject floating out there somewhere. If you contact their tech support they should be able to send you a copy.
 

jaggedben

Senior Member
... They claim this is saving us $ but in reality the cost of upsizing to #8 wire probably ends up costing more than the fuses and fuseholder inputs...
The derating issue applies whether there are fuses or not though. Where it really costs money is in the delays and failed inspections if an AHJ doesn't just let unfused strings sail through.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Ugh...SolarEdge has frustrated me with this as well. I wish they would just bring back the fused inputs with their three phase inverters. But, you can in fact combine three strings together with solaredge without fusing but it requires some additional work.

Here is a link to a white paper from solaredge on the subject of combining three strings: https://www.solaredge.com/sites/default/files/string_fusing_requirements.pdf
I sent that document to our local AHJ and his response was that he is not going to take SolarEdge's word for it. He wants to see something from an independent testing lab.
 

BillK-AZ

Senior Member
Location
Mesa Arizona
With SolarEdge you have series strings of Optimizers, not series strings of PV modules.

The SE9KUS has two DC inputs on a terminal strip (the two input strings are parallel connected at the input terminal), each terminal position is usually used for a series string of Optimizers. The datasheet shows 'Maximum Input Voltage DC+ to DC-' is 500 VDC and 'Maximum Input Current' is 26.5A. Maximum DC Power (Module STC) is shown as 12,150W.

The datasheet for the P320 Optimizer shows that for THREE PHASE 208V 'Minimum String Length (Power Optimizers)' is 10, 'Maximum String Length' is 25, and 'Maximum Power per String' is 6000W. Maximum Output Current is 15A.

The original post does not identify the PV module rated power, but if it is 300W or less there would be no problem connecting one string of 20 and one string of 19 Optimizers to the inverter. No fuses required. If the PV modules are over 300W each, there is a problem.

The discussion at a SolarEdge training session was that the inverter will adjust the outputs of the Optimizers such that the inverter input current, 26.5A in this case, is not exceeded.

The installation instructions do state "You can connect systems with multiple DC strings in parallel to the DC input terminals of the switch. Inverters have one, two or three pairs of DC input terminals, depending on the inverter power rating. If more strings are required, they can be connected in parallel using an external combiner box before connecting to the switch." These are strings of Optimizers and the rules based on PV module ISC and related fuse requirements do not apply, but the output current of the Optimizers must be considered for wire sizing. If fuses are used to reduce wire gauge for each string, they are based on the continuous 15A of the Optimizer, 20A in this case and the 20A is used for calculation of the required wire gauge as generally done.
 

pv_n00b

Senior Member
Location
CA, USA
I sent that document to our local AHJ and his response was that he is not going to take SolarEdge's word for it. He wants to see something from an independent testing lab.
Unfortunately, this is not a test in UL 1741 and I doubt SE is going to have an NRTL run an independent test to demonstrate how it works ($$$$).

SolarEdge also has a letter from Bill Brooks on this subject floating out there somewhere. If you contact their tech support they should be able to send you a copy.
Bill must think AHJs roll over after seeing anything with his name on it. I've seen the letter, it makes a good argument for the safety of the SE system and how SE complies with the spirit of the NEC without complying with the text. As good a description as I have seen in this thread. It's basically a request for an AM&M exemption from the AHJ. Maybe an AHJ will accept it, maybe not. Find out before you install anything because the rework will cost you.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Unfortunately, this is not a test in UL 1741 and I doubt SE is going to have an NRTL run an independent test to demonstrate how it works ($$$$).
I know, and that's why we put in string fusing when we have to use three strings in most (all?) jurisdictions. Choose your battles.
 
Top