Several utility-scale solar farms began losing DC string fuses in their combiner boxes within the first year of operation, eventually by the hundreds. No faults were found on the strings. The fuses were listed for photovoltaic (PV) service, and the site's own current logs showed the strings carrying roughly half the fuse's continuous current rating. The owner wanted to know why fuses were opening on normal load current, and whether it would continue.
Simplified Explanation
A fuse is designed to melt first at intentional "weak" spots in its internal ribbon. PV string current rises and falls every day with the sun, and that daily heating and cooling flexes the ribbon. Over time the ribbon can stretch and thin at the weak spots until ordinary load current is enough to melt it. The incident fuses likely operated this way. However, enclosure temperature, how the fuses were mounted next to each other, and the condition of the ribbon when it was manufactured could not be ruled out as contributors without more testing, so the root cause was undetermined.
How a fuse opens
- Current through the fuse ribbon heats it (I²R heating).
- When any local section of the ribbon reaches its melting temperature, the ribbon separates and an arc forms across the gap.
- The silica sand surrounding the ribbon fills the gap and forces the arc to wind between the grains, which lengthens it.
- The arc's heat fuses the sand into a glass-like insulator called fulgurite, which lengthens the arc path further and raises the arc voltage.
- The higher arc voltage drives the current toward zero. When the gap exceeds what the system voltage can sustain, the arc extinguishes and the fuse has opened the circuit.
Fuse ribbons are notched to create the weak spots, and a tin overlay is usually added near the center so that melting starts at the notches in the middle of the ribbon. The location of the break and the amount of melted material show how the fuse operated.
Reading the evidence
- Location of the break. In overcurrent testing of new exemplar fuses, every fuse opened at the center notches, as designed. In the field fuses, about three-quarters had severed near one end, close to a contact, away from the center notches where the design initiates melting.
- Amount of fulgurite. A "high energy" (high current) operation consumes most of the ribbon and leaves substantial fulgurite. The field fuses showed very little melted ribbon and fulgurite, which is consistent with a "low energy" operation at or near normal load current.
- Ribbon shape. X-rays showed "S" and "V" bends in the ribbons of most field fuses. The manufacturer attributed these to cyclic loading. Some new and unused exemplar fuses also showed sharp bends, so the possibility that some bends were present at manufacture could not be excluded.
- Temperature-rise testing. Exemplar fuses were heat-run in fuse holders in the spirit of the listing standard's temperature test. One of three opened at rated current before the test was complete, which suggests little to no margin on this fuse model's continuous current rating.
The sizing check
NFPA 70 (NEC) Article 690.8 sizes PV source-circuit overcurrent devices from the module short-circuit current: 125% for the maximum circuit current, and then 125% of that value for the device, or 156% of short-circuit current in total. As an illustration, a string with a 20 A combined short-circuit current requires an overcurrent device rated at least 31.25 A. In this case the installed fuse appeared slightly undersized, by less than one ampere. A shortfall that small does not explain fuses opening at half their rating.
Why the root cause was undetermined
The site data and laboratory testing suggested the fuses should not have been operating under the measured conditions. Several factors could not be ruled out without more work:
- Combiner box internal temperature on a hot, sunny day with the door closed had not been measured. Site ambient readings do not represent the air around the fuse.
- The fuses were mounted in a vertical stack. Heat from one loaded fuse may raise the temperature of the fuse above it, and no thermal analysis of that arrangement had been done.
- Without the manufacturer's design and production records, the as-manufactured ribbon condition could not be assessed.
An undetermined root cause still records what the evidence supports and names the testing that would resolve it.
What it means for owners
- It is recommended that owners log temperature inside representative combiner boxes, with doors closed, through summer days before accepting a "no derating needed" conclusion.
- Check overcurrent device sizing against NEC 690.8 using the module data sheet, and remember that the fuse holder has its own continuous rating, which can be lower than the fuse's.
- Retain operated fuses, labeled with location and date, so the break location and fulgurite can be examined.
- Ask the fuse manufacturer for cyclic-loading qualification data for the specific fuse and holder combination as installed.
- Consider whether stacked mounting needs derating or spacing.
Standards referenced
- NFPA 70 (NEC) Article 690.8, circuit sizing and current for PV systems
- UL 248-1 and UL 248-19, low-voltage fuses (general requirements and PV fuses), including temperature-rise and current-cycling tests
Details have been generalized to protect client confidentiality.