PCB trace used as a fuse?

[David_]

Hello.

I know that I can use a PCB trace as a fuse and there are a lot of documents and calculators online to calculate such a trace.

But I want to ask:
Do anyone here use such fuses, or have you ever used such a fuse?

If there are anyone here whom has used such a trace I would lie to know what your thought about them are, I at least think that it might be useful and a cool feature to have such fuses in my PCB designs to act as a catastrophic failure fuse.
By that I mean a fuse that isn't suppose to blow ever but in case of some weird and unusual fault could save or spare some components from being destroyed or harmed.

I welcome any thoughts on this and it don't have to be relevant if you have used such a fuse or not, I'm unsure whether I should spend some time designing such fuses or if it is unnecessary.

Regards

- - - Updated - - -

Ok so I see that this topic has been discussed before, but it would be nice if we could discuss it from the point of view that this is meant for fire prevention and such rare serious faults and that we assume that there are also other fuses in the system. Why then another fuse?
Well what if I have a design that may be vulnerable for being shorted by a object failing onto the PCB or any other weird situation.

I am not designing for commercial project but rather my home private projects and as such I am exposing PCBs to the environment of my work desk.

 

[ads-ee]

Let me start by saying...No I have never seen such a thing used before.

Why would you want to add something that basically damages the PCB on purpose, when a failure occurs? If it's to protect a component or a bunch of components, now you have a damaged board, depending on the amount of damage this may mean all those "saved" components get tossed anyway due to the board being damaged, or the need to remove all the parts (perhaps causing some to fail later due to ESD). Even if it doesn't damage the PCB to the point of non-functionality, now you have a board after rework to fix the trace, which then has NO protection, so next time there is a fault, all those parts are now toast.

They do make surface mount fuses, which could be replaced by a technician with a soldering iron. This is something I have seen on PCBs over the years. Usually somewhere in the PS design where if the end user F's up the power into the board and they RMA the board and we replace the fuse for $$$$ and return the board with a note to not exceed the maximum input voltage.

 

[schmitt trigger]

I've seen traces used as fuses, but only in cheap, throwaway Chinese products.
In the particular instance, it was an inexpensive CFL.

The only reason not to use a fuse is because you intend the product to be disposable. And save a few pennies.

 

[pjmelect]

Yes I have seen the PCB trace used as a fuse, and it worked very well. I don't know if it is good practice or not but in cheap equipment that is built to a price it is better than no fuse at all.

 

[c_mitra]

If you ever need a populated PCB to self-destruct for some reason, this may be a decent option...

 

[marce]

In over 30 years I have never seen a PCB trace used as a fuse, I would never do it and all designers I know would never do it...
Basic thoughts, a fuse is deigned to do a job, they are tested and certified, a PCB trace used as a fuse is not tested so you are relying on an untested and unreliable PCB trace to fuse to prevent damage due to some high current incident...
Not a good idea.

 

[David_]

Okey, I couldn't reason about this situation my self but after reading your posts I have no intent what so ever to use such a fuse and I don't know why I got onto that path. Probably because it is a interesting use of a trace, I can only think of 1 situation in which to use it(not counting PCBs manufactured with the express intent to maximize the profit to the extreme or to use it a as a... I don't remember the English term for it but the industry seems to be riddled with companies that purposefully make sure that devices will fail, such as printer ink cartridges with inbuilt counters used to tell the printer that the it has run out of ink when in fact there are lots of inc left, in a documentary I saw they reset that counter and could use the cartridge 3.5 times...
Or that outrageous coalition of the biggest manufacturers of light bulbs who sat down many many many years ago and decided that no one light bulb will be allowed to burn more than 1000hours before it breaks, how long did they burn before this gangster style deal? about 2500hours... That is truly outrageous.)

Anyway the one usage for this kind of fuse maybe would be if you have a circuit that for some reason or another can't be protected by a fuse or if the device is operating in an environment where even after the ordinary fuse has blown there are a little tiny risk of a fire lighting up, if such a situation is even possible(I don't know) then such a one-time PCB fuse could be better than allowing something to burst into flames.

I haven't used them yet but I have carefully picked out SMD fuses from digikey and alike for a project but the one's I choose are a few SMD fuses that actually does have a fuse holder(of course also SMD), they seems to be really neat. Mouser don't have nearly as many as digikey had when I checked a year ago but here is an example of fuse+holder:

Fuse:
https://www.mouser.se/Search/ProductDetail.aspx?qs=8QfiMw8FY5PsEpOJdIg2UA==
Holder:
https://www.mouser.se/ProductDetail...=sGAEpiMZZMtRmoYvq3OwzOfwamFEAnEJiZtramS/hww=

 

[FvM]

Using PCB traces as fuse is a really badly considered idea.

1. Using a PCB trace as fuses means to plan board total loss in case of overcurrent

2. Depending on the expectable current and voltage, it risks fire and arcs jumping over to nearby circuits.

We do however occasionally observe PCB traces acting as unintentional fuse, with similar effects as described above. Therefore it's essential to design PCB trace widths in coordination with fuse I²t values to avoid trace fusing.

 

[David_]

I guess I should continue on this track and learn how people calculate PCB trace fuses, not to use as a fuse but to calculate at around what current I might think that my traces could go up in smoke.

I will look at the documents I have gathered but I find it funny that there are so much resources online for calculating PCB traces as a fuse when it is obviously a bad idea to try something like that.

I for one will take your advice and you have convinced be that I should use traces as fuses... no I mean NOT use traces as fuses

 

[betwixt]

Also consider that fuses are made from low melting point wires and may be held under tension by a spring to ensure a 'clean' and fast break. Copper, as used on conventional PCBs is not going to break easily unless there's a very considerable over-current situation, it is more likely to run hot and char the board beneath it which may itself then become conductive!

Brian.

 

[marce]

I guess I should continue on this track and learn how people calculate PCB trace fuses, not to use as a fuse but to calculate at around what current I might think that my traces could go up in smoke.

I will look at the documents I have gathered but I find it funny that there are so much resources online for calculating PCB traces as a fuse when it is obviously a bad idea to try something like that.

I for one will take your advice and you have convinced be that I should use traces as fuses... no I mean NOT use traces as fuses

https://www.saturnpcb.com/pcb_toolkit.htm

Conductor fusing currents...


https://www.pcdandf.com/pcdesign/

This month an article of conductor heating...

https://pcdandf.com/pcdesign/index.php/magazine/10546-thermal-overload-1601

Fusing of PCB traces....

The whole point of being able to calculate the fusing limit of a trace is so you don't burn it off, its not to create PCB trace fuses but allow us to work out maximum current capacity especially when pulsed currents are used....

 

[Mattylad]

If your making a board that has 100k+ assembly runs per month, and you need to fuse it to prevent a fire on fault then a trace fuse will work and save you the cost of an SMT fuse. which on 100k+ runs does add cost for the potential few that may possibly fail.
If you calculate the trace width etc. correctly then it will work, if you have pads either side you can even repair the board if needs be.

Cheap, not as accurate but it can work - I have seen it work.

 

[FvM]

Cheap, not as accurate but it can work - I have seen it work.

It depends, I think.
- It does not work reliable as overcurrent protection because low overcurrents will burn the PCB substrate rather than fusing the trace.
- It can work as short circuit protection if current and short circuit power are limited by the external circuit (e.g. a regular mains circuit breaker). Even with limited short circuit current, a fusing trace at mains voltage can generate an arc of several 10 mm size, possibly jumping over to low voltage circuits.

 

[c_mitra]

Also consider that fuses are made from low melting point wires and may be held under tension by a spring to ensure a 'clean' and fast break. Copper, as used on conventional PCBs is not going to break easily .

Not all fuses are made from low melting alloys- There are fuses that are used to ignite some chemicals - some explosives, e.g. They are made from corrosion resistant high melting point wires - e.g. platinum. Also many high current fuses are made from copper.

 

[betwixt]

Not all fuses are made from low melting alloys- There are fuses that are used to ignite some chemicals - some explosives, e.g. They are made from corrosion resistant high melting point wires - e.g. platinum. Also many high current fuses are made from copper.

Agreed, but the point I was trying to make is if a design was so cost critical that a 'real' fuse couldn't be fitted, it's unlikely to have platinum traces instead of copper or be intended to ignite the PCB subtstrate!

The trace ratings (Amps per mm width per copper thickness) are there to help a designer ensure the trace can safely carry a current, taking into account the thermal conditions, allowed voltage drop and proximity to other components. The ratings are not intended to be used to calculate the rupture current and to do so would need knowledge and verification of the copper thickness, thermal resistance of the substrate and any chemical changes to the substrate as it overheated and maybe turned itself into conductive carbon. I have seen track fractures in high voltage PSUs, similar to the situation of a fused track, where the substrate took over and continued to conduct, albeit not very well, much like a carbon resistor.

Brian.

 

[FvM]

Not all fuses are made from low melting alloys

True, but that's not the point here. The essential point is embedding the fusing conductor by temperature resistant and arc quenching material.

There's however a large range of fuses with current breaking capabilities between a few 10 and hundreds of kA and also quite different voltage ratings. It may be possible to identify a certain range that can be safely handled by PCB traces.

 

[David_]

This thread has been very intresting, on the subject of calculation of a traces current carrying capabilities I have wondered about the hole trace resistance situation.
I gather that there gave come a new standard for calculating trace resistance that is said to be a big improvement from the old standard, while trying to write a simple Matlab script to use as a trace resistance calculator I got into problems due to the complexity of the newer standards equations and parameters.
The old one was very simple and one could think that such a calculation would be simple, but perhaps there are more things influencing the results than it first appears.

I know that temperature does influence the resistance of copper and thus a higher current equals a higher temperature equals a higher resistance and a higher resistance should equal a lower current.

Let's imagine a process in which a trace is burnt off, if we could look at how much power it takes to fuse it, and if we would look at it as steps of temperature increase.
Would it require more and more energy to create the same increase in temperature each step the higher we would get, in temperature?

 

[marce]

If your making a board that has 100k+ assembly runs per month, and you need to fuse it to prevent a fire on fault then a trace fuse will work and save you the cost of an SMT fuse. which on 100k+ runs does add cost for the potential few that may possibly fail.
If you calculate the trace width etc. correctly then it will work, if you have pads either side you can even repair the board if needs be.

Cheap, not as accurate but it can work - I have seen it work.

These assemblies could not be sold, they break to many standards for electrical safety......

- - - Updated - - -

I gather that there gave come a new standard for calculating trace resistance that is said to be a big improvement from the old standard, while trying to write a simple Matlab script to use as a trace resistance calculator I got into problems due to the complexity of the newer standards equations and parameters.

The old graph was created many years ago and was not the most scientifically conducted experiment... the new spec IPC-2152 was based on more comprehensive experimentation. Worth looking up the history etc...

Let's imagine a process in which a trace is burnt off, if we could look at how much power it takes to fuse it, and if we would look at it as steps of temperature increase.
Would it require more and more energy to create the same increase in temperature each step the higher we would get, in temperature?

Post #11 I posted a link to PCD&F this covers conductor heating go read it and you will learn more.

 

[FlapJack]

I think the correct answer is, if you can get an agency approval like UL or VDE then it is ok. But i do not think you can get one.

 

[marce]

LOL UL get a tiny bit tetchy about burning PCBs, they get tetchy over track temperatures so having a track that is going to deliberately fuse and cause even the slightest ignition will really please them.....
The correct answer is use a fuse in a flame proof housing or a re-settable fuse, you cannot risk a fire however small the chances are so no you cannot use a PCB trace as a fuse unless you are daft.