Polyfuse 8MILLI-ohms post-trip? am I missing something?

[pilotplater]

I'm working on an extremely simple PCB to just add some logic-level shifting to an ESP8266 board before sending the signal out to a strip of 5v 'nixie' style LEDs (I know not always required, but I want to cover all bases)

I thought maybe I could be smart to add a polyfuse on the main power input, never know when you'll come a gutsa with a short or power supply fault etc.

My question has to do with this digikey listing: https://www.digikey.ca/product-detail/en/littelfuse-inc/NANOSMD500LR-2/NANOSMD500LR-2CT-ND/7321653



it seems to suggest that the pre-trip resistance is 2 milli-ohms, and the post-trip resistance is 8......milli-ohms? mega-ohms? I can't imagine Digikey would be so vauge in their listing to have milli-ohms and mega-ohms noted as the same text, but also why would a polyfuse be designed to 'trip' to 0.008 ohms? Seems like a too crazy niche application... Not sure what I'm missing, is this field even relevant?

Cheers
Andy

[maginnovision]

What that means is after it trips the inherent resistance goes up. So it starts at least 2 milliohm across the fuse, if it trips it'll be up to 8 milliohms.

[RoGeorge]

A polifuse "cuts" the current when an over-current situation occurs, then after the situation has restore to normal, the continuity of the fuse is restored.  After a trip, the circuit restore its continuity by itself, it does not remain open like a normal fuse.

When new, the polifuse has 2 miliohms, when an over-current occurs it almost break the circuit, then, if the over-current situation disappears the polifuse "regenerates" itself, but not as good as brand new, 8 miliohm instead of 2 miliohms.

Mega is always written with capital 'M'
mili is always written with lower 'm'

[Zero999]

Yes, as mentioned above, after the first time it's tripped, the resistance is higher than before. This also means it will be more sensitive to over current and will trip at a slightly lower current, then next time. Polyfuses or PTC (Positive Temperate Coefficient) resistors to use the correct term, don't have an infinite life. After multiple trips there may come a point when the normal operating current will cause it to trip, at which point it will need replacing. This is why it's important to not use too lower tripping current rating.

[james_s]

That's interesting, I never realized that. Why don't they pre-trip them prior to shipment so that they're consistent throughout their useful life?

[T3sl4co1l]

If you're depending on a fuse to give "consistent" behavior, you're using the wrong component!

Tim

[james_s]

I just find it odd that it's specifically rated one value pre-trip and a 4x higher value post-trip. It never occurred to me that a single trip would result in such a large change in characteristic. I always just assumed they were similar to circuit breakers which can be tripped many dozens of times and still behave fairly consistently.

[MagicSmoker]

That's interesting, I never realized that. Why don't they pre-trip them prior to shipment so that they're consistent throughout their useful life?

Probably because it typically takes 3-6 seconds to get one to trip and that is way too long to spend in ATE for mass-produced commodity devices. But if you need mil-spec then, yes, they are actually fully tested (so pre-tripped); you pay mil-spec prices, though.

[ejeffrey]

That's interesting, I never realized that. Why don't they pre-trip them prior to shipment so that they're consistent throughout their useful life?

Generally they are never "supposed" to trip.  In the vast majority of cases that would result in a fuse with an unnecessarily high resistance and dissipation for it's entire operating life.

Also it would be an unnecessary processing step that costs money and doesn't have any clear value to most people.

If you want accurate interruption with low and constant resistance use a sense resistor and a FET switch.   Or add fault interruption to your device QC test which will pre-burn-in the switch.  But know that depending on reproducible behavior from a fuse is not a great idea.

Polyfuses are also heavily sensitive to ambient temperature.  If your device needs to operate over a wide temperature range you may need to derate enough that it doesn't provide enough protection in the cold.  In that case you need a standard non resettable fuse.

[Siwastaja]

Also note that pre-trip is MIN and post-trip is MAX value. The actual "typical" difference likely isn't 4x.

For example, pre-trip value may be between 2 and 4 mOhm, and post trip value may be between 3 and 8 mOhm.

But rating for min and max, respectively, is the right thing to do - they matter for your design margins.

Polyfuses require understanding and are not precision devices. They also wear out. After, say, 100 trips, the R may be too much.

[SiliconWizard]

That's interesting, I never realized that. Why don't they pre-trip them prior to shipment so that they're consistent throughout their useful life?

Because that's how they are supposed to be used. Don't you prefer a component that has minimal series resistance in normal conditions, and degrade a bit when shit happens (which should be a rare occurence unless your design is fucked up).

Besides, they will not really stabilize to a given ESR. Every time they trip, they will tend to degrade a bit more.

[Siwastaja]

Besides, they will not really stabilize to a given ESR. Every time they trip, they will tend to degrade a bit more.

This. If they did stabilize after one or two trips, pre-conditioning them would make sense.


Use polyfuses only if you expect handful (tens, not thousands) of trips, and design for a large leeway in resistance (much more than R1). If you can't make it work, then don't use a polyfuse; design an active electronic efuse circuit (ICs do exist), protected by a classical single blow fuse (which only blows if the efuse circuit fails).

[T3sl4co1l]

Yeah, that's the other thing about polyfuses: the self-resetting is a convenience, not a guarantee.  They are good for ~hundreds of cycles and are supposed to fail (permanently) open.

Tim

[ejeffrey]

One more thing: surface mount ones are considerably worse than through hole in most parameters. I haven't paid much attention to the R R1 spread, but the hold to trip ratio in particular is much worse for surface mount devices.

[T3sl4co1l]

THT also more consistent on current rating, since less dependence on trace/plane heatsinking.  But way slower, since they're usually much bigger (and amps+).

Tim

[SiliconWizard]

Yeah, that's the other thing about polyfuses: the self-resetting is a convenience, not a guarantee.  They are good for ~hundreds of cycles and are supposed to fail (permanently) open.

Yeah. And remember they are not a silver bullet. Do your risk analysis. They make sense if an overcurrent situation can reasonably be expected to happen during the lifetime of your device, and if it's better to handle it this way (making the device usable again once the condition is gone), or if it makes more sense that the device stop functioning altogether, then requiring action from its users. In some cases, allowing it to be powered again could be unsafe.

[pilotplater]

Thanks everyone for your detailed info, this is all very helpful. Good points all around!