Fuses and PPTCs
A polyfuse is a resistor that is allowed to get hot, and everything good and bad about it follows from that. It resets itself, it never quite comes back, and it is thousands of times too slow to save a transistor.
There is one on almost every board you own. On an Arduino Uno it sits between the USB socket and everything else, a 1812 package marked with nothing useful, and its job is to stop your project setting fire to your laptop's USB port.
Drag the load current and watch what it does — and what it charges you for doing nothing.
It is not a fuse and it does not blow
A glass fuse is a wire that melts. A PPTC — polymeric positive temperature coefficient, sold as a polyfuse, a resettable fuse or a Multifuse — is a slab of polymer packed with carbon grains. Cold, the grains touch and it conducts. Warm it up and the polymer expands, the chains of grains break apart, and its resistance climbs by three or four orders of magnitude in a fraction of a second.
Take the power away and it cools, the grains touch again, and it conducts once more. That is the "resettable" part and it is genuinely useful: a user who shorts something does not have to find a spare fuse.
The three numbers on the datasheet
The part on an Arduino Uno is a Bourns MF-MSMF050-2. Its datasheet gives it:
- Hold current, 0.5 A. It carries this for ever without tripping.
- Trip current, 1.0 A. It is guaranteed to trip at this.
- Time to trip, 0.15 s at 8 A.
The gap between the first two is the part everybody misreads. Between 500 mA and 1 A the datasheet promises nothing at all, because whether it trips depends on how fast the board carries the heat away — the copper it is soldered to, the air around it, the ambient temperature. A polyfuse in still air at 60 °C trips at currents it would happily carry at 20 °C.
What it costs you on the days nothing goes wrong
0.15 Ω, minimum, permanently in series with your supply. At the hold current that is 75 mV the board never sees and 38 mW of heat, on every single day of the part's life.
And it does not come back to 0.15 Ω after a trip. The same datasheet gives R1max — the resistance an hour after tripping — as 1.0 Ω. Nearly seven times what it started at. It recovers further over days, and never entirely.
This is why nobody polyfuses a 3.3 V rail. Half an ohm between a regulator and a board that draws 300 mA is 150 mV of a rail that only has 3.3 of them, and it moves every time the load does.
The thing it cannot do
Press play and watch the clock. It is logarithmic because the two numbers do not fit on a linear one.
150 milliseconds is a long time. A small MOSFET held in a dead short is outside its safe operating area in tens of microseconds — four orders of magnitude earlier. The polyfuse has not begun to warm up.
This is not a defect. A polyfuse protects the supply: the USB port, the wall adapter, the wiring loom, the battery. Those things can take a tenth of a second of overload and shrug. It was never protecting the semiconductor on the load side, and its datasheet has never claimed to.
When to use one anyway
They are cheap, they need no support components, they cannot be programmed wrong, and they survive being the last line of defence. Use one where the thing you are protecting is slow and expensive and the fault is rare: a USB host port, a battery pack's output, a connector a user can short with a screwdriver.
For anything where the thing on the far side is a semiconductor, or where you want to choose the limit, or where you cannot spare a quarter of a volt, the next page is the part you actually want.
Edit this page — content/fundamentals/power/fuses-and-pptcs.mdx
Discuss this article
Ask about this page. The answer stays here, on the page it belongs to, for whoever hits the same wall next.