If your solar fuse keeps blowing, the first thing to establish is whether the fuse is doing its job or whether it is operating when it should not. Those are two different problems with two different investigation routes, and treating them as one is how an array ends up with a box full of spent fuses and a fault that is still there.
A string fuse that opens because a real fault exists has performed as designed, and the answer is to find and repair the fault. A fuse that opens because it was the wrong rating, the wrong class, or running hotter than anyone allowed for is nuisance operation, and the fix lies in the device’s selection and installation. This guide separates the two and shows how to confirm each cause in the field.
Sizing is deliberately out of scope: string current, the applicable margin and temperature correction are covered in solar PV fuse selection for gPV 1000 V and 1500 V. This article assumes the fuse is already in place and asks why it keeps opening.
Read the evidence rather than the symptom. Both cases leave you with an open fuse, but the site gives different clues.
| What you observe | Fuse cleared a real fault | Nuisance operation |
|---|---|---|
| How often | Once, or after a specific event — a storm, impact damage, work on the array | Repeatedly, often after hot weather or re-termination, or after a string was added |
| Which fuse | One string, and the same one each time | Different strings on different occasions, or several at once |
| Fuse and holder | Soot, arc marks, a blackened body or holder, sometimes a cracked body | Element parted with a heat-discoloured barrel; a warm or lightly discoloured holder |
| Array beforehand | The affected string measured worse than its neighbours, or a defect was obvious nearby | All strings normal right up to the moment of failure |
| Marked rating | Consistent with the string current and the datasheet | At or barely above the string’s short-circuit current, or a class that is not PV-specific |
Neither column is a diagnosis by itself, but the pattern tells you where to spend your time. A real fault can also be intermittent: a connector that only fails when wet, or a cable that only faults when the array flexes in wind, will open fuses at intervals that look random.
Everything below assumes a competent person working to the site’s own procedures. Modules produce voltage as soon as light falls on them, and opening a fuse or isolator removes the path to the rest of the system rather than de-energising the cable between the modules and the open device.
Work through the sequence in order and record every result; the value is in the comparison between strings, not in any single number.
Strings in a combiner box are in parallel. If one develops an internal fault — a shorted bypass diode, a shorted module section, a failed connector — it stops behaving as a source and starts behaving as a load, and the remaining strings push their combined current into it. The fuse in the faulted string carries the sum of the others’ currents and is expected to open. This is exactly what a string fuse exists for.
So the question is not why the fuse failed but what is wrong in the string it protected. It only works if every string has its own fuse of the correct class and rating — a box that fuses groups of strings, or leaves some unfused, cannot limit back-feed at all.
Damaged insulation, chafing on a roof edge, rodent damage, a cable crushed under a mounting foot or water tracking inside a junction box creates a path from the string to earth. Depending on where the fault is and how the system is earthed, the result is a sustained leakage current the fuse sees as an overload, or an intermittent fault triggered by moisture. Insulation resistance testing confirms it; a healthy reading on a dry day and a clear failure after rain is a common signature.
Connectors are among the most frequent failure points in fielded arrays: a partly mated pair, mismatched types, a crimp never properly made, or corrosion in a housing that has taken water. A high-resistance joint heats under load, accelerating its own deterioration, and can end as an open circuit, an earth fault or a fire. Bypass-diode failure inside a module junction box produces the back-feed condition above. Confirm with thermal imaging under load, a physical check of every connector, and voltage comparison between modules.
Water in a module junction box, an inline connector or a combiner-box gland rarely announces itself. It corrodes contacts, provides a leakage path, and creates faults that appear after rain or condensation and disappear later. Look for staining, verdigris on copper, tide marks and swollen or cracked gaskets. Replace the affected connector or gland rather than drying it out and closing it up — the corrosion is already in the contact.
Here the array is healthy and the fuse still opens. Every cause below sits in the selection or installation of the fuse and its holder — and one of them, the wrong fuse class, is a safety problem rather than an inconvenience.
A PV string fuse is normally specified with margin above the string’s short-circuit current, because a module can deliver more than its nameplate value at high irradiance and low cell temperature, and the margin also absorbs manufacturing tolerance and the fuse’s behaviour over life. Without it, the fuse lives at or above its rating, ages quickly and eventually opens on a healthy string.
The method belongs to the dedicated guide — see how to select a gPV solar fuse for 1000 V and 1500 V systems. If the installed fuse was chosen by eye, copied from a similar project or inherited from a supplier’s default, verify it against that method before looking any further.
A fuse’s current rating applies at the reference ambient stated in its standard and on its datasheet. Above that temperature the element runs hotter for the same current, ages faster and carries less. The temperature that matters is not the weather forecast but the air inside the enclosure, which on a sunny day can sit well above the outside air and higher still once the devices inside are loaded. The correction factor is standard- and manufacturer-specific and must come from the datasheet — it is not a universal number.
If a design used the rating at the reference ambient and never corrected for a box that runs hot, the result is exactly the symptom in the title: a fuse opening a healthy string on the hottest days of the year while every measurement looks normal. Find out whether the enclosure has ever been temperature-surveyed at peak load.
A gG fuse is a general-purpose industrial fuse. It is not a PV fuse class. gPV fuses designed and tested to IEC 60269-6 — and to UL 248-19 for the North American market — are built for what a string imposes: sustained DC operation, a time–current characteristic matched to the slow thermal behaviour of a PV string, and the ability to interrupt the reverse current a parallel array can drive through a faulted string. A gG fuse may have no DC rating at all, or a DC voltage rating below the string’s open-circuit voltage; fitting one in a string holder is both unreliable and dangerous. The general principles are in AC vs DC fuses.
A fuse holder is part of the current path. A terminal never torqued to the manufacturer’s value, a ferrule not fully seated, or contacts that have lost tension adds resistance, and every amp through it produces heat that cycles with the day. The fuse body sits next to that heat source, so its element ages faster than the datasheet implies. The signature is a warm holder and a discoloured fuse in an otherwise healthy string. Fix it by re-terminating to the specified torque and replacing a heat-affected holder matched to the fuse body — see matching a DIN rail fuse holder to the fuse. Never by fitting a larger fuse.
Lightning and switching events can damage a fuse element without destroying it, leaving it to open later under normal load. A fuse that fails alongside a nearby strike is one symptom of a surge event, and the surge protection should be inspected at the same time — a surge that reached the fuse also reached the SPDs, whose indicators and condition are part of the evidence.
| Cause | How to confirm | Correct action |
|---|---|---|
| Reverse or back-feed current from parallel strings | One string’s fuse open; that string shows low or zero open-circuit voltage or a shorted section; neighbours normal; arc damage in the fuse | Keep the string isolated, find the faulted module, diode or connector, repair or replace, then fit the correct gPV fuse |
| Insulation or earth fault in a string | Insulation resistance to earth below specification, or changing between dry and wet conditions; visible cable damage | Repair or replace the damaged cable section, re-test, record the result against the baseline |
| Module or connector failure | Hot connector or junction box on thermal image under load; discoloured or partly mated connector; abnormal module voltage | Remake or replace the connector with a matched pair and correct crimp tool; replace the module if the fault is internal |
| Water ingress in a junction box, connector or gland | Water, tide marks, verdigris or corrosion inside a housing; faults appearing after rain or condensation | Replace the affected connector or gland; correct gland sizing and entry direction; re-verify enclosure integrity |
| Undersized rating for the string current | Marked rating barely above the string short-circuit current on the module datasheet | Re-select using the proper method, fit the correct gPV fuse, record the calculation |
| Missing ambient or enclosure derating | Measured internal temperature well above the fuse’s reference ambient; failures concentrated in hot weather or at peak load | Apply the datasheet derating for the real internal temperature, re-select the rating, reduce enclosure temperature where possible |
| Wrong fuse class — gG or AC-rated fuse in a string | Class marking on the fuse; datasheet with no DC rating, or a DC voltage rating below string open-circuit voltage | Replace with a gPV fuse of the correct voltage and current rating, in a holder rated for the same |
| Loose terminal or high holder contact resistance | Holder or terminal warmer than its neighbours on thermal image; terminal not at the manufacturer’s torque value | Re-terminate to the specified torque with the correct ferrule or lug; replace a heat-affected holder |
| High ambient inside the enclosure | Temperature survey of the enclosure interior at peak load | Shade the box, ventilate consistently with the IP requirement, or re-select devices for the real temperature |
| Transient or surge event | Fuse failed around a storm or utility switching event; SPD indicators at end of life or SPD damaged | Fit the correct replacement fuse, inspect and replace SPDs as required, verify earthing |
| Fuse not seated, or holder mismatched to the fuse body | Fuse loose in the holder or not gripped at the caps; holder voltage rating below the system voltage | Use the holder matched to the fuse body size and voltage; confirm the assembly before re-energising |
| Something about the string changed | New string, new module type, re-termination or a different fuse batch immediately before the failures began | Re-verify the affected string’s design, including fuse and holder, as a change rather than a repair |
One related temptation is worth naming: raising the fuse rating to stop it opening. That is not a repair. It removes overcurrent protection from the string and leaves the array outside its design basis, while the actual cause — a fault, a hot enclosure, a bad holder — carries on unchanged.
No. A blown string fuse is a symptom. Replace it only after the measurement sequence has been worked through and the cause is known — a repaired fault or a corrected selection. If it opens again immediately, stop and diagnose rather than repeating the cycle.
Usually because ambient and enclosure derating was never applied, the rating has no margin above the string’s short-circuit current, or a loose terminal is heating the holder. Hot-weather-only failures point at temperature effects rather than a fault in the array; the enclosure’s internal temperature at peak load is the number to measure first.
Not on a PV string. General-purpose classes such as gG are not built or tested for sustained DC operation, and may have no DC rating at all or a DC voltage rating below the string’s open-circuit voltage. Use a gPV fuse, with a holder rated for the same voltage and current.
Isolate the string, verify dead, remove the fuse and test continuity out of circuit. A visual check alone is not reliable, because an element can part inside an intact-looking barrel. Then compare the fuse with its neighbours to confirm it was the right class and rating in the first place.
It may stop the fuse opening, but it also removes the protection the string was designed around and leaves any underlying fault in place. Oversizing is not a solution to nuisance operation; correcting the selection, the enclosure temperature or the holder is.
Work through the sequence, keep the failed fuse and the measurements, and decide from the evidence whether you are repairing an array or correcting a specification. In most nuisance cases the answer is the fuse class, the derating or the holder rather than anything in the modules.
If you are specifying string fuses and holders, or replacing a fuse that keeps opening, send us the string datasheet values, the system voltage and the enclosure details — contact us and we will help match the device to the job.
