1

Can 0 Ω resistors be used as fuses? I can't find any fuses physically small enough to fit on my board that can also take the current we're expecting.

In my application, 58 A is the expected continuous current in situations where the software has accidentally turned on every peripheral. That would be rare, but we don't want the fuse to blow in this situation.

The situation where we do want the fuse to blow would cause hundreds of amps to flow through the fuse.

However, the only fuses I can find that fit on my board are 1206 and rated to 50 A which would lead to many false blows, even when the software hasn't accidentally turned on every peripheral. The next size up for fuses is massive.

Is it possible to use something like a 0 Ω resistor in a 2512 package as a fuse, e.g. something like CR2512AJ/-000EAS which is ≤50 mΩ and has a maximum current of 5A? But at what temperature and current will it break?

I'm surprised that I can't find any ceramic 2512 SMD fuses. It has to be ceramic.

ocrdu
  • 8,705
  • 21
  • 30
  • 42
Dan
  • 133
  • 8
  • 13
    I don't understand... You want to use a 0Ω resistor rated for only 5A, to use it as a fuse that isn't supposed to blow below 100A? That resistor will be vaporized when it will see your 58A continuous current you want to be able to handle. – dim Sep 08 '22 at 08:10
  • 5
    Can you split the peripherals so that you have two fuses and there will be less than 50A through each fuse? Or use a single 0 ohm resistor for each peripheral? – Justme Sep 08 '22 at 08:12
  • 22
    seriously? 58 A means that if you tolerate a temperature rise of no more than 60 °C, your trace needs to be around 30mm wide (on a standard 35 µm thickness copper-clad PCB). And you are trying to save space on SMD components? – Marcus Müller Sep 08 '22 at 08:23
  • 11
    A 50mΩ resistance (at 20°C) carrying 58A would dissipate almost 170W, and the resistance would surely increase as it quickly heats to incandescence. 10x rated current means 100x (or more) the maximum heating. – Spehro Pefhany Sep 08 '22 at 08:27
  • 10
    I think a better approach is to fix the part when the software turns every peripheral on – Rahmany Sep 08 '22 at 08:29
  • 3
    If zero-Ohm SMD resistors could be used as a fuse *(regardless of the load current)* I'm sure almost no one in the industry would bother with placing the big chunky glass fuses. – Rohat Kılıç Sep 08 '22 at 08:53
  • 6
    Fuses will blow up accidentally. That is their purpose. It is a terrible idea to disassemble an industrial high power equipment and swap a soldered component every time the fuse is activated. It is not just the additional labour, time and the needed technical personal, but circuit boards are not designed for numerous resoldering. You can get maybe away with 5-10 replacements, but not much more. The heat generated on a res scorches the circuit board as well. Fuses should be easy to replace without the thermal stress caused by resoldering. Also how a resistor blows up is not well controlled... – Horror Vacui Sep 08 '22 at 08:56
  • I think there is consensus that a plain SMD film resistor can be regarded as a fuse, as in when it blows, it burns off and it doesn't fail shorted. If it is suitable to use one as a fuse, well that's another story. Most of the time some trace along the way will burn before that, unless you have a very careful layout with wide traces everywhere. – Lundin Sep 08 '22 at 09:34
  • 1
    Just say no to 58A and a surface mount fuse. Chances are it won’t protect the board anyway at this level. – Erik Friesen Sep 08 '22 at 11:55
  • 1
    How small is "physically small enough to fit on my board"? You can get 70 A fuses that are about an inch long and half an inch wide, and that seems pretty small as far as 70 A fuses go. – Cassie Swett Sep 08 '22 at 12:18
  • 1
    The purpose of a fuse is to avoid being sued by a customer whose building has just burned down. Using anything that isn't UL (or whatever the relevant agency is in your country) listed as a fuse for protection 100% defeats the purpose. When your handling 50 A, pretty much every device that current flows through probably ought to be certified. – The Photon Sep 08 '22 at 14:47
  • Try this: https://www.youtube.com/watch?v=aHimIdiH2Gc – Kyle B Sep 08 '22 at 16:04
  • If the zerohm resistor has 50mΩ resistance rated at 5A it will dissipate 1.25 watts, and may get hot enough to melt the solder and cause it to come loose, and perhaps short out nearby components. At ten amps it will 5 watts, certainly enough to cause desoldering. Of course, the pads and traces to which it is connected will conduct some heat, but consequences are unpredictable and unsuitable for use as a fuse. – PStechPaul Sep 08 '22 at 20:30
  • You may also consider a "fusible link" in the supply cable. Not exactly the same as a fuse, not SMD, but at least designed for the purpose. – david Sep 08 '22 at 21:44

3 Answers3

3

Zero ohm resistors are lousy fuses, and SMD ones especially so: they will burn and possibly sputter metal particles as they die their heat death. They also aren’t calibrated to ‘blow’ above a specific current.

Instead, I think this would be best handled by using over current protection in the power supply itself or just after it, ideally using a fold back current limiter. Such a high current should have a dedicated path back to the supply anyway.

Related: What is "foldback short circuit protection" in a power supply?

Another option is to use a real replaceable or resettable fuse in the circuit. Again, this could be located at the power supply on a dedicated path to your peripherals.

In case you’re missing what I’m getting at, your fusing or current limiting doesn’t need to be right at your peripheral connections. Instead it can be located upstream on a dedicated power route back at the power supply, or in the power supply itself if it’s a separate unit.

Your on-board power bus should be sized large enough to handle the fault. If it isn’t, then use a lower current limit upstream so you don’t burn traces.

hacktastical
  • 49,832
  • 2
  • 47
  • 138
  • 2
    The OP didn't mention the voltage of the supply. When the resistor blows, it may create a sustained arc and burn a hole in the PCB and vaporize the pads. – qrk Sep 08 '22 at 18:24
  • Yes, fuses have a voltage rating for that reason: to ensure that once they’re blown they melt back far enough to make a wide enough gap that the supply voltage can’t jump across. Resistors have voltage ratings for a related reason: their terminal spacing limits the voltage they can handle. – hacktastical Sep 08 '22 at 18:33
  • It's also unlikely that it would be easy to replace a 'blown' 0 ohm resistor – Voltage Spike Sep 08 '22 at 18:50
  • 2
    Assuming there’s enough board and trace left to do so… – hacktastical Sep 08 '22 at 19:01
2

Fuses are safety devices, carefully designed for well defined operation when used according to their ratings.

So the answer is definitely NOT, if you are concerned with the safety of your device (both operator safety and device safety).

If you really need a fuse, you must use a fuse, not another very unreliable poor man's alternative.

LorenzoDonati4Ukraine-OnStrike
  • 24,350
  • 5
  • 62
  • 107
  • I marked the other answer as the accepted answer because it gave more information, but your answer is the one that made my stakeholders stop throwing solutions at me. Thanks for the help :) – Dan Jul 05 '23 at 11:42
1

The fusing current for 19AWG copper wire is about 70 A. Its diameter is 35.9 mils. This size of fuse will not fit into a 2512 space. You will have to move the fuse offboard.

RussellH
  • 12,496
  • 2
  • 9
  • 34