Understanding PTC fuses

[bitman]

I've noticed a lot of even small projects that use USB or PSUs with MCUs have a PTC fuse in them, and until a few weeks ago I didn't even know that existed.  I read up on them, but that didn't really help me understand how they work - the details just provided me more questions than I had before.  I am hoping that someone here may be able to ELI5 or similar so I get the idea.

Here's what I've done so far - I found some 0ZCJ0050FF2G "PTC RESET FUSE 8V 500MA 1206" while I was getting some other stuff at DigiKey and added a few of them to "try them out". I understand that they're rated to handle a given "hold current" (actually a given power rating) and once the temperature rises high enough (https://www.belfuse.com/resources/datasheets/circuitprotection/ds-cp-0zcj-series.pdf says above 125 degree C - my Flir tells me it's about 80 Degree C on the surface) the resistance goes up to "virtually remove the current".  Well, that's not what I'm seeing - all the "response time" parameters I see doesn't show up in my little test-setup.

I connected the unit to a bench PSU, providing it with 5V - since the 8V is maximum power. Added a DC load to control the power usage and a voltmeter to show the voltage drop over the PTZ. Unfortunately the 500mA is based on 8V so 800mA would be the limit with 5V if I understand this right. Nothing happens (changes) at 800mA - I need to get close to 1100mA before the temperature slowly gets to the point where the fuse "sets".  But according to the datasheet they're talking sub-seconds ... perhaps I'm reading that wrong, it's absolutely a very slow moving train here. Or do I need to surge with almost 1A for the trigger to be fast? At that point my components will definitely be fried.

The voltage drop is about 0.3-0.5v (pretty high for a 5V circuit) which jumps to 4.7v when the fuse sets. This means there's still current  going through the circuit even when the fuse "pops".  Not sure how that protects the circuit? It takes a while for the fuse to pop, by then damage may already have been done and worse, there's still power going to the units post "pop".  And I seem to only be able to reset by turning off power to let the PTC cool down - even lowing the current consumption to 100mA seems to keep the high temperature fairly steady.

So I'm not sure I understand the "fuse" part. Power is still flowing even when this unit pops, so why is it a good design choice to have one (that's my presumption as I keep finding these everywhere). It seems to be extremely slow reacting and would depend on the type of PCB, thickness and what may be close to the PTC on the PCB?  Let's say I have a design I know will not survive more than 500mA at 9v, how do I pick the right PTC?  If I expect the average load to be around 200mA but it will peak just below 500mA during use for small periods of time. Since I didn't see the temperature go down significant when the current got low, it doesn't seem picking a 250mA "hold current' unit is the right thing? But if I pick at 500mA it will go WAY above 500mA before the PTC reacts, and that's definitely not a good thing. So help me understand the use case and how to pick the right value?

Perhaps this is more an ESD thing than a general input power protection?

[mon2]

IMHO, do not use PTC fuses. They are chemical fuses that will derate over time and are not as reliable (slower to trigger) as silicon protection devices. Consider instead a device like the Diodes load switch AP2331. Just be sure the selected part can support your required current. At the very least, allow for a traditional fuse that can use also the PTC in the footprint but parallel it with the AP2331 device. We do this in volume on all of our USB powered products.

[wizard69]

First; with new stuff these days, YouTube is often a good place to search for educational materials.   In this case there are several videos from supplier or manufactures of such devices.

Second; generally fuses, circuit breakers and such are considered to be fire prevention devices.   That is they are there to prevent wiring from overheating and causing fire.   Even if a fuse is sized for motor overload protection the goal is still the prevention of motor destruction due to overheating the wiring.   If you look at it this way you will then realize that heating is impacted by current over time.   The more current that flows the faster something heats up.   Fuses, circuit breakers and such are generally designed to blow real fast at high currents and much slower around the set point.   Fuses especially are not real accurate as far as to when they blow, this really requires reading and understanding the manufactures data sheets.   In the same way as standard fuses you need to understand how PTC resetable fuses work which requires the data sheet again.   As you may already know there are many things that impact how the devices actually trip.   In any event, according to the pdf you referenced the trip time is based on an 8 AMP current, not 800 milliamps.   At 800mA i'm seeing a trip time closer to 10 seconds - Average!!!   plus you get this disclaimer:

The Average Time Current Characteristic Curve and Temperature Rerating Curve are affected by a number of variables and these  curves are
provided for guidance only

.

Third; I don't really see fuses, of any type, as being a good way to protect electronics.   Use electronics to protect itself as electronics will be fast and more precise.   In a nut shell fuses usually blow after the electronics fail and again this is about safety with respect to fires.   When dealing with power electronics there may be situations where fuses can actually protect solid state power electronics but that electronics must be robust enough to withstand the fault currents needed to trip the fuse.

[chinoy]

First I tried PTC fuses. They were utter crap. Had all the issues the OP had. The worst problem being they dont last.
From PTC I moved to using TVS diodes. Those were a little bit better but when they did fail. The failed shorted. This would burn out the PCB tracks.
In the end I just opted to use a 7812 or 7805 voltage regulator IC.
I found these regulators did a pretty good job of limiting voltage and dealing with noise on the line. Also dealing with shorts.
They seem to be built like tanks. Never had one fail on me. Just an idea.