Ask most buyers what they need from a semiconductor fuse and the answer comes back as two numbers: rated current and rated voltage. Both matter, but neither is the number that decides whether an IGBT survives a fault. That number is I2t — the energy let through by the device before it clears, expressed in ampere-squared seconds. If you have ever watched a converter come back from a short circuit with a fuse intact and a semiconductor in pieces, the cause is almost always a mismatch on I2t rather than on current rating. This article explains what I2t measures, how to read it on a datasheet, and how to match it to the device you are actually trying to protect.
What I2t Actually Measures
I2t is the integral of current squared over time. Written out, it is the area under the i-squared curve from the moment the fault begins to the moment the current reaches zero. Why square the current? Because the heating that destroys a semiconductor junction, a bond wire, or a busbar is proportional to i-squared-R. The thermal damage done by 2000 A flowing for 2 ms is not the same as 1000 A flowing for 4 ms, even though the ampere-time product looks similar. Squaring the current captures what the silicon actually feels. So I2t is not an abstract figure of merit. It is a direct proxy for how much thermal energy the protected device has to absorb during the fault. That is exactly the quantity you need if the thing you are protecting is a piece of silicon with a thermal mass measured in grams.
Why Current Rating Alone Will Not Protect Silicon
A general-purpose fuse is designed to protect cable. Cable is forgiving: it has real thermal mass, and it takes a lot of energy to raise copper to a damaging temperature. That is why a conventional device can afford to let a fault run for a few cycles before it clears. A power semiconductor is the opposite case. The junction has almost no thermal mass, and the surge capability of a diode or IGBT is typically defined for a single half-cycle event. Between the fault starting and the junction exceeding its limit, there may be only a few milliseconds — and in a stiff DC link, less. This is the entire reason the semiconductor fuse exists as a separate category. It is built to operate fast enough, and to limit current hard enough, that the energy delivered to the junction stays below the device's withstand. Its fast action is not about convenience or selectivity. It is the only way to keep the i-squared-t within the capability of silicon. Select on rated current alone and you can easily end up with a device that carries the load current perfectly and still lets through four times the energy the IGBT can survive.
Pre-Arcing, Arcing, and Total Clearing I2t
Datasheets usually split I2t into two figures, and the distinction matters. Pre-arcing I2t, sometimes written as melting I2t, is the energy from fault inception to the moment the element begins to melt. It is the quantity you compare against the surge rating of the semiconductor, because from the device's point of view the current is still rising during this period. Arcing I2t is the energy delivered while the arc is being established and extinguished inside the body. Total clearing I2t is the sum of the two, and it is the figure used when you are checking the thermal burden on the enclosure, the busbar, or the fuse holder itself. There is a practical trap here. Pre-arcing I2t depends on the prospective current and on the time constant of the circuit, so a single number in a datasheet table is only valid under the stated test conditions. Good manufacturers publish curves. If your datasheet gives one number with no conditions attached, treat it with suspicion.
How to Read I2t on a Datasheet
When you open a datasheet, work in this order:
- Confirm the test voltage and whether the figures are AC or DC, because DC interruption is the harder case
- Find the time constant used for the published values and compare it with your circuit
- Locate the pre-arcing I2t curve rather than a single table value
- Read the let-through or cut-off current curve, which shows peak current against prospective current
- Check whether the values are given at the rated voltage or at some reduced value
You will also see class designations. Under IEC 60269, aR denotes partial-range breaking capability intended for semiconductor protection, while gR and gS denote full-range devices that additionally protect against overload. Knowing which class you are looking at tells you whether the device is expected to clear small overloads at all.
Matching I2t to the Device Withstand Limit
Every power semiconductor has a surge rating, usually published as an I2t figure for a specified pulse width and case temperature. Your job is simple in principle: the pre-arcing I2t of the protective device must be below the withstand I2t of the semiconductor, with margin. In practice, apply margin for three things. First, the published semiconductor figure is typically for a single event at a defined starting junction temperature; a device already running hot has less headroom. Second, repeated near-fault events accumulate damage even when each one is nominally within limits. Third, the fuse's own I2t varies with the actual prospective current, so check the worst case rather than the nominal case. We usually ask for a safety factor rather than a bare comparison, and we ask to see the calculation written down. It takes ten minutes and it is the single most useful check in the whole design review.
Where Selection Goes Wrong in Real Projects
A few failure patterns come up repeatedly. The first is substitution. A device fails, someone matches the current and voltage from the label, and the replacement goes in. If the original was a fast semiconductor type and the replacement is a general-purpose type with the same rating, the current still flows and the silicon still dies. Ampere rating and voltage rating do not describe speed. The second is ignoring DC. Fuses characterised on AC can behave very differently on a DC link, where there is no zero crossing to help extinguish the arc. Always use data taken at your actual voltage and current type. The third is thermal. Semiconductor protection devices are often mounted in hot cabinets next to the very semiconductors they protect. Derating for ambient temperature is not optional, and a device operated continuously near its thermal limit will drift in behaviour long before it fails visibly.
Summary
Rated current tells you what a semiconductor fuse can carry. I2t tells you what it can save. If the pre-arcing energy is below the withstand of the device you are protecting, and the voltage and time constant conditions match your circuit, the selection will hold up under fault. If you only checked the current rating, you have checked the wrong number.
About Our Semiconductor Protection Range
GFEFUSE is the brand of Dongguan Gongfu Electronics Co., Ltd., an electronic protection device manufacturer established in Dongguan, Guangdong in 2013. Our team covers development, production, and sales, supported by five advanced production lines and thirteen R&D engineers, and our products are exported to customers overseas. For semiconductor protection we supply a dedicated North American semiconductor fuse range alongside our square high speed fuse series, which covers fast DC protection for converters, drives, and power electronics. Customers designing to British practice will find matching options in our British standard fuses, and the wider high voltage fuse portfolio supports the rest of the distribution chain. Products carry UL, cULus, TUV, CE, and CQC certification and comply with RoHS and REACH. Applications span consumer electronics, automotive electronics, power equipment, communication systems, and new energy.
Conclusion
When a converter survives a fault, it is usually because someone did the I2t comparison before the design was frozen. When it does not survive, the post-mortem almost always finds a device selected on current and voltage alone. Read the pre-arcing curve, compare it against the semiconductor's surge rating, confirm the test conditions match your circuit, and the protection will do what it was fitted to do.
Send us your converter voltage, DC link current, prospective fault current, and the I2t withstand of the device you are protecting. We will check the coordination and recommend the right class and rating.
Frequently Asked Questions
Short answers to the questions our engineering team handles most often on semiconductor protection.
What does I2t mean on a semiconductor fuse datasheet?
It is the thermal energy let through during interruption, expressed in ampere-squared seconds. Pre-arcing I2t covers the period up to element melting and is compared with the semiconductor surge rating. Total clearing I2t includes the arcing period and is used for enclosure and busbar checks.
Why is a semiconductor fuse faster than a general purpose fuse?
Because silicon has almost no thermal mass. A general purpose fuse is designed to protect cable, which can absorb far more energy. A semiconductor fuse must limit current and clear fast enough to keep the junction below its damage threshold.
Can I replace a semiconductor fuse with one of the same current rating?
Only if the replacement is the same class and speed. Matching current and voltage alone says nothing about let-through energy. Substituting a general purpose device for a fast semiconductor type is a common cause of repeated IGBT failures.
How is I2t related to breaking capacity?
They are different checks. Breaking capacity is the maximum current the device can interrupt safely. I2t is how much energy passes through before it does. You need both: enough breaking capacity to survive the fault, and low enough I2t to protect the semiconductor.
Do semiconductor fuses need derating?
Yes. Ambient temperature, adjacent heat sources, and continuous cyclic loading all reduce the effective continuous rating. Apply the manufacturer's derating curve for the actual cabinet conditions rather than selecting on nameplate current.