resettable fuse current choice

[VEGETA]

hello,

If I want the fuse to trigger at 250mA, then this should be the "trigger current" of the resettable fuse.

however, it has also "holding current", say 100mA which means the current that is guaranteed not to trigger at.


my question is what if I ran 150mA or 200mA continuously... what is gonna happen?

it is well below trigger current but also above holding current. It may mean that fuse might trigger at in between currents... but is this correct?

[Ice-Tea]

Congrats, you just found out why resettable fuses sound awesome but are pretty difficult to design with.

To make things worse; these values are often for 25C and are heavily influenced by temperature...

[SteveThackery]

my question is what if I ran 150mA or 200mA continuously... what is gonna happen?

It is what is euphemistically called "undefined", and you will find examples of undefined behaviour in many areas of electronics.

So, all you know is that the device is guaranteed not to trigger at or below 100mA, and guaranteed to trigger at 250mA or more.  What happens between those values is, literally, undefined.  It might, or might not, trigger.  Therefore it is your responsibility to select a device that never has to operate in its undefined zone.

[Whales]

Huh, this explains some problems I had recently.  I thought the trigger was a min, not a max.

[VEGETA]

well, i thought about this "undefined" thing but good thing to know that there are firm boundaries for cutting off the circuit and making sure it operates nicely.

i have chosen this: https://www.lcsc.com/product-detail/Resettable-Fuses_Brightking-SMD1206B016TF_C269110.html

it has triggering current of 370mA and holding current of 160mA.

it will protect the output of LM317M, this one in particular: https://www.lcsc.com/product-detail/Voltage-Regulators-Linear-Low-Drop-Out-LDO-Regulators_STMicroelectronics-LM317MDT-TR_C68136.html

this version of lm317 is 500mA max but i want to limit it to say 370mA, i got other fuses which can get down to 290mA and 250mA but the holing current is 100mA so I thought I should give it a bit more room.

the expected current draw of the circuit is not known unfortunately but main functionality won't exceed couple of milli-amps since it only gives a signal for SCART activation to old CRTs.

I want to test it at near the 370mA range to see how hot it can get, and of course test the short circuit functionality.

when it cut the circuit, then short is removed... will it need to be turned off then on or will return working fine by its own?

[MasterT]

FYI, there are modern voltage regulators like NCV47701, that has OCP feature:
Adjustable Current Limit (from 10 mA to 350 mA) with 10% accuracy

[Vovk_Z]

With current higher than rated workng currdnt it'll heat itself gradually and will tend to break at some point (minutes possibly).

[VEGETA]

With current higher than rated workng currdnt it'll heat itself gradually and will tend to break at some point (minutes possibly).

you mean higher than "holding current"?

so it can generate heat? I actually don't want heat in my circuit. So it is safe to say that I must use a fuse which has holding current more than enough for regular operation.

this particular lm317 seems to have only 500mA so when more is needed i think it will have some drop voltage, but not over current or short protection.

[VEGETA]

FYI, there are modern voltage regulators like NCV47701, that has OCP feature:
Adjustable Current Limit (from 10 mA to 350 mA) with 10% accuracy

due to the low significance of this portion of the circuit and simplicity of its function, plus the cheapness... I chose lm317. this particular model (LM317M) has only 500mA output which is nice.

i have a physical switch to switch a feedback resistor to change the output voltage, this is what the circuit design requires.

your suggested one is nice but very expensive, can't be purchased from LCSC.

[ejeffrey]

Note that metal element fuses are *much* better than resetable fuses.  Except of course for the part where they need to be replaced.

PTC fuses have a much higher ratio between hold and trip currents, and much higher resistance (and therefore heat generation) in the normal operating region.  They are also much more temperature sensitive.

One more problem: PTC fuse resistance is usually a "zero lifetime trip" rating.  The first time they trip, the resistance will irreversibly increase.  This will lower the true trip point and also increase operating resistance and heat generation.

Usually you can go down the following path:

1) Look for a PTC fuse with a hold rating higher than your worst case load.
2) Find out you have to derate it for operating temperature
3) Look up the trip current for a PTC that will actually work, panic.
4) Re-evaluate how big of a problem nuisance trips actually are.
5) Use a metal element fuse and tell people that having to replace it is a learning experience.

[ArdWar]

Don't let the name fools you into unrealistic expectation. PPTC fuses are more like a fuse with second chance (and no more) than actual resettable circuit breaker. Its parameters rarely "reset" to original specifications after a trip, and it only get worse after subsequent trips. They may suffer from aging too, definitely don't ride them close to trip as it makes them age faster.

As general rule:

• If you want dependable protection and faults happen extremely rarely, actual fuse might still be better. It forces the user to acknowledge the fault instead of giving them the chance to ignore it.
• If faults are to be expected during normal operation, other form of resetable protections are often preferred. Maybe pick one of those fancy eFuse or similar if your power devices don't have inbuilt OCP.
• PPTC falls in this no man' land where your product is not meant to be field serviceable while you expect faults to happen more than once, but less than say ten times over the product service life.

[wraper]

my question is what if I ran 150mA or 200mA continuously... what is gonna happen?

Eventually resettable fuse will fail. And I actually mean fail, not just trigger.

[VEGETA]

thanks for your suggestions, really learned more about these fuses.

my device (psu) will be inside a retro game console and won't be opened for any reason. expected faults or high currents are really nothing, i am just adding the fuse to protect the rail since other rails have advanced buck regulators with their own over current protection built in. this rail is either 9v or 12v, user selected but once selected it will stay that way.

i didn't want to add more cost that necessary, efuses are nice really but i believe they can be more expensive, like even at lcsc it will be 0.5$. the idea of using lm317 is that it is cheap and reliable for this very simple insignificant function that is needed from the rail.

EDIT:

I have found this tiny device: TPS25961, price at digikey is 0.61$ and at lcsc is 0.5$. it is smaller that the ptc fuse i am using so it can fit without much changing. can do 0.1 to 2 amps so that i can dial it to 400 mA or 300 mA easily. has 106mOhm resistance but at these low currents won't make much voltage drop... around 32mv at 300mA.

it can tolerate my 9v and 12v rails. it has these fancy Undervoltage Protection (UVP) and Undervoltage Lockout (UVLO) which i am not interested in. so i guess i will hook EN pin directly to Vin, and UVLO pin to ground. this way it uses the fixed values which should be ok. it has quirky I_limit setting resistor values as some in table 7.1 must not be used, meaning it does not cover the full range.

taking that into consideration, picking a resistor value of 147k gives me 340 mA current limit which should be ok. 147k was not used in my design, closest thing is 162k which will give current value of 308 mA.

[VEGETA]

I attached my final circuit in this post, please check it out.

I used the efuse I mentioned above + LM317M. efuse is dialed at 340 mA which should be more than enough.

lm317 portion of the circuit have been verified to work practically, it switches nicely between the voltages with some mV more than required, to compensate for any drop since the device will use wiring harness to deliver power. all tested to see no serious drop using 10 cm awg18 wire... new wire harness is gonna be awg16 gauge with 15 cm length, should be better choice in terms of drop voltage.

[K3mHtH]

Eventually resettable fuse will fail. And I actually mean fail, not just trigger.

Can you give any more info on this? I've used a few in a design. I don't expect them to ever trigger, and I'm just using them as an "in case shit goes wrong".