I am experimenting with making a PCB and have created some short circuits when wiring everything up which have fried one of the traces. i was wondering whether it would be a good idea to put a fuse in the next iteration to avoid breaking the whole PCB if i should make another mistake. could anyone recommend whether this is a good idea and if so, how would i choose the right fuse? the circuit runs of a lithium battery so 4.2v -3.3v thereabouts and i am not expecting current draw about 0.5A. Many thanks.

  • \$\begingroup\$ A PTC would be a good idea , get a bunch, you can gang in parallel too digikey.ca/products/en/circuit-protection/ptc-resettable-fuses/… \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Apr 19 '18 at 17:27
  • \$\begingroup\$ You should probably try to explain as many details as possible. Can you define "some short circuits"? How did you short them, and for how long? Can you define "fried one of the traces"? What actually happened to it? Do you actually mean that one of the components failed or that the trace literally melted? \$\endgroup\$ – Bort Apr 19 '18 at 18:11
  • \$\begingroup\$ It was one of the traces actually melting i believe as i rerouted it with a wire and it worked ok. the trace was very thin(too thin, has been addressed for future). \$\endgroup\$ – user2105725 Apr 19 '18 at 19:38

Check out a resettable fuse such as a PTC:


The PTC comes in regular packages like 0805 and through-hole just like your other capacitors and resistors. When the current through the PTC exceeds a certain limit it heats up and the resistance increases, reducing the current to your board.

Once you turn off the power and the PTC cools down then the resistance decreases and after a while you can use it again, making it resettable. It's not perfect so if you keep tripping the PTC repeatedly you may need to replace it eventually, but it's good for a couple of uses.

I'd recommend having a LED before and after the PTC so you can tell when input power is applied (first LED turns on), and when the PTC has tripped (second LED turns off).

To pick your PTC you need to look at:

  1. The voltage used by the circuitry on your PCB
  2. The hold current, how much current can go through the PTC normally (your expected maximum load)
  3. The trip current, how much current is needed to make the PTC start increasing resistance and limiting current to your PCB.

You can do a parametric search from most manufacturers to focus on the products they offer that fit your requirements.

  • \$\begingroup\$ Thanks for this(and every other reply). ive just had a look at the website you recommended and am trying to figure out which smd fuse would be the one for me. since i will be using a lithium battery, max voltage will be 4.2 so i would then choose the 6v max rated fuse presumably as its closest to my max voltage. How might i go about calculating the trip current for my 4.2 voltage which is lower than the 6v rated? many thanks. \$\endgroup\$ – user2105725 Apr 19 '18 at 19:35
  • \$\begingroup\$ (or will the fuse trip at the specified current regardless of the voltage?) \$\endgroup\$ – user2105725 Apr 19 '18 at 19:52
  • \$\begingroup\$ A 6V PTC will work, just pick one that is equal to or larger than your operating voltage. For trip current, choose how far above your maximum operating current you are willing to go before the PTC trips. There's usually a considerable gap between these two; e.g. a PTC with a 250mA hold current may trip at 500mA. That means it's guaranteed to run at 250mA continuously without tripping, but won't start to trip until 500mA. Your peak short-circuit current could be as high as 500mA before protection kicks in. \$\endgroup\$ – user185972 Apr 19 '18 at 20:46
  • \$\begingroup\$ thanks thats very helpful, ill order some and experiment. one thing which threw me with regards to voltage and mA is i have some 250v 500mA fuses which i tried to burn out using 5V and 600mA and it didnt work. it may have been a long burning fuse as the resistor i was using began smoking so i stopped.... (any thoughts?) also, would a traditional fuse (as opposed to PTC) have a smaller gap between the running and trip current? are these available in the smt variety? thanks again. \$\endgroup\$ – user2105725 Apr 20 '18 at 7:35
  • \$\begingroup\$ I haven't had a chance to design or experiment with a traditional fuse yet so I'm not sure how they work. Yep PTCs are in the SMT form factor too, 1206, 0805, etc. \$\endgroup\$ – user185972 Apr 20 '18 at 14:15

A fuse could help, but first the traces on the PCB should be designed to handle whatever the maximum fault current can be. There are trace width / temperature rise calculators you can use to do this. An alternative to fuses are electronic circuit breakers that have the advantage of being more precise and act much quicker than a fuse. Linear Tech, TI and others make ICs that do this.

  • \$\begingroup\$ It's nonsense to make traces that wide so they can handle current the power supply is able to provide. A fuse is device for this task. The traces must be wide enough for proper conditions. \$\endgroup\$ – Chupacabras Apr 19 '18 at 17:51
  • \$\begingroup\$ This is done all the time, in many cases putting a fuse or other protection device on every power bus of a PCB is not practical or necessary. I did not say not to use a fuse and even with a fuse the traces need to be designed to handle the current that can occur before the fuse blows or the traces will blow before the fuse. Even a fast blow fuse can take a long time to blow above its rated current. \$\endgroup\$ – EE_socal Apr 19 '18 at 18:06
  • \$\begingroup\$ I commented on your statement that traces should handle whatever fault current, and that fuse doesn't that matter. Which I considered nonsense. \$\endgroup\$ – Chupacabras Apr 19 '18 at 18:24

You've probably already done this, but just in case: If your using CAD to design it (I say that because I once saw a guy design an etching mask on a blank PCB with a marker), sometimes the auto-route feature on Eagle or other CAD softwares will have issues and route something wrong or have a few traces too close. I remember eagle once tried to auto label all of my wires as ground. always double check and maybe allow more space between traces in the next iteration if in doubt.

I agree with the others,especially raisin-wrangler, I would always use a fuse of some type (maybe add a crowbar circuit in the final iteration for added robustness if you want), recheck your PCB wiring AND I would also figure out a way to check the thickness of the copper cladding that's on the PCB.

I don't know where you got the copper clad boards from, but I know Some Chinese manufacturers often cut corners and have bad quality control. So you could have gotten a PCB that has thinner cladding than you thought, which might throw the PCB thickness calculations out of whack thus causing the PCB traces to not be able to handle the current.

An example of Chinese quality: I once got a Chinese knock off solder station as a gift from a non-Elec. Eng. relative. I took a look inside to find that some of the heatsinks for some finger-burning-hot T0-220 transistors were missing, the soldering iron and heat gun were not ESD safe like advertised and the heat shrink tubing on some connections had not had heat applied yet.

However, I'm guessing a micrometer could check the thickness of the cladding? I've designed boards before, but have usually had a board manufacturer do the manufacturing in the past. Nevertheless its an option to keep in mind when ordering boards on the next iteration. Perhaps you could beef up the cladding thickness or select a different manufacturer of the copper cladded boards to combat a cladding thickness issue if it is present.

That's everything I can suggest. Hopefully this helps. Good luck on the next iteration.

EDIT: wanted to clarify on the copper clad board manufacturer part and put the micrometer part in bold

  • \$\begingroup\$ Thanks for this its good advice. the problem did arrive from me not being scrupulous in checking the paths generated from using a ground plane. nonetheless a fuse would hopefully act as a second chance in the future. \$\endgroup\$ – user2105725 Apr 19 '18 at 20:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.