Talk to me about resettable fuses (AKA polyfuses)

I’m working on an electronics project and I’d like to add a fuse to make it safer. One possibility is to use a resettable fuse like this one (PDF). However I’m confused about the difference between the hold current and the trip current. According to the data sheet they are:
[li]Hold current: maximum current device will pass without tripping in 20°C still air.[/li][li]Trip current: minimum current at which the device will trip in 20°C still air[/li][/ul]
As an example one of the fuses in the link has a hold current of 1.1A and a trip current of 2.20A. This seems contradictory. What happens if the current is, say, 1.5A?

If my project runs safely up to 1A and I want the fuse to trip on anything above 1A what values would I choose for hold/trip current?

At 1.5A, the device may eventually trip, but it will take a long time.
If your device takes 1A, you need a 1A hold current, and then the lowest trip current you can find (greater than any expected surge currents). Note that these devices are usually designed to protect against short circuits - they react slowly to slight over-current situations.

ETA: FWIW, I used these to protect some open-collector output switches against shorted loads. I found that they react much too slowly to protect the transistors unless a small amount of resistance was added to limit the peak overcurrent.

This sounds like what, in the industry, would be called a ‘slow blow’ fuse. It’s meant to be used in applications where it’s going to start motors or other things where it’s going to be called on for a lot of amps at once, but not the rest of the time.
So, for example, my walk in freezer might need a 30 amp breaker, but it might call for 50 amps for a split second when the motor first starts up, just for that first half second when it’s under a lot of load and maybe from time to time when when other things happen in the cold box. But if it sustains more then 30 amps for more then a few seconds, it needs to pop because something’s wrong.

In your application it sounds similar, it’s meant to blow if it sense more than 1.1 for any length of time but can deal with double that for a few second or two here and there, maybe charging up a cap or having a small motor spin up under load.

ETA, from here
A standard fuse may require twice its rated current to open in one second, a fast-blow fuse may require twice its rated current to blow in 0.1 seconds, and a slow-blow fuse may require twice its rated current for tens of seconds to blow

ETA2, on the page you linked to, it says at the top that it has a fast-to-trip time, but looking at the 1.1amp fuse it says it trips in 8.2 seconds at 5.5amps, maximum, that sounds like a slow blow fuse to me, but, then, EE has never been my thing. You want help fixing your furnace, I can do it, but I can’t solder to save my life.

All measurements have an error.
When you say 1 Amp is your cut off, which zone of uncertainty exists ? You specified a single digit, but did you mean 1.0 A, 1.00 Amp, 1.000 Amp ? Well its never clear … so really you should say 1% or 10%.
Well the device has the error margin specified, there is some uncertainty of what happens in between hold and trip currents, and that is all you can be certain of…
hold will not trip the device, trip current will.

They are quite slow and only useful to protect the device from overheating the entire body… eg melting the plastic frame of the device, melting glue or insulation… it doesn’t protect a semiconductor from over-current (as that heats up and overheats in just fractions of a second.)

Fuses in general are not high tolerance devices. If your fuse has a hold current of 1.1 A and a trip current of 2.2 A, at 1.5 A it might trip or it might not. If you buy two identical fuses, one might trip at that current and one might not. You have to get all the way to 2.2 A before they are both guaranteed to trip.

As beowulff said, you need a hold current of 1.0 A. If your device will completely fry at say 1.2 amps, then you need something with a closer tolerance and a much lower trip current.

You also need to decide whether your device can tolerate a certain amount of overcurrent for a small amount of time before it fries, or if it is extremely sensitive to overcurrent and needs to shut off immediately. If you need the protection to trip faster, you may be looking at a fast-blow fuse instead to provide adequate protection.

Why do you want to use a resettable fuse? Why not use a traditional fuse?

In any design, the default is the latter. You would only use a resettable fuse if a traditional fuse won’t meet the requirements.

Those things are PTC (positive temperature coefficient) devices. They’re never 100% “on” or “off,” so at 1.5 A there’s a chance it could simply have some “elevated” value of resistance. It’s hard to say. You’d have to examine the R vs. I curve to estimate what it would be. (And above a certain current threshold, positive feedback forces it into thermal runaway, and the resistance becomes very high.)

At any rate, I would definitely look at using a traditional fuse. But even with a traditional fuse, you can’t determine the proper value - or proper type - without knowing what the waveform of the inrush current looks like.

Thanks guys, I appreciate the help. The project I’m working on is to make an electrically heated scarf which is intended to be used while sitting at the computer. The power supply will be a laptop PS which will heat some nichrome wire that runs through a fire-resistant scarf. Since this is something that will be worn around the neck I want to make it as safe as reasonably possible, thus the fuse. It was suggested that I consider a polyfuse; I like the idea because it’s more compact than a traditional fuse.

The biggest danger is a short of the nichrome wire (which I will try to prevent, of course) and the fuse doesn’t have to be all that fast; it just has to trip soon enough to prevent a fire.

I’ve done some experiments and it appears that 20-25 watts will generate a decent amount of heat. To get 25w from a 12v PS the wire will need to be ~5.76 and the current will be just above 2 amps (the 1 amp from my OP was just an example). So (assuming I use a polyfuse) I want one that has a holding current ~2.1A and as low as a trip current as possible. Does that sound reasonable?

Will this be a commercial product? Mass-produced, or one-off? Prototype? Fun science/educational project?

Where can (or must) the fuse be? In the scarf? At the power supply? In the connector?

This is a project for personal use but if it works well I hope to make few as gifts (my extended family gets cold easily). The fuse can go anywhere that it most convenient. I suppose an inline fuse between the PS and the scarf would be reasonable.

Could I make a suggestion that the connection from the scarf to the laptop be magnetic. No matter how hard you try, people are going to stand up with this on either ruin the cord (or the connector in the laptop), strangle themselves or pull their laptop off the table.
If it’s magnetic it’ll just breakaway.

I’d love to use a magnetic plug. Do you know of any that are available online? In my searches I haven’t found any. I’m probably up to the task of making my own but that does increase the complexity.

Don’t put the fuse in the scarf. A fuse is very sensitive to its surrounding temperature, and will trip prematurely when the temperature is above room temp.

Use a miniature, inline fuse. The closer it is to the power supply, the better.

You should also install a TVS or MOV between the fuse and scarf. It should be wired in parallel with the scarf, and should conduct when the voltage is greater than 15 V to 18 V. (Some people may claim this is not necessary. “Excessive voltage will cause excessive current, thereby blowing the fuse.” True. But a person can be electrocuted from a transient voltage, and the fuse may not be fast enough. Furthermore, what happens if there’s a break in the nichrome wire, and then a transient occurs? It’s cheap insurance.)