I am installing a set of recessed lights in a closet, replacing the existing fixtures one at a time. I removed power to the lights while installing each one by using the light switch, not the circuit breaker at the box (after verifying that the the switch was wired correctly with the hot side being switched, not the neutral).

I successfully installed the first light, and had just pulled the second one down. I turned the lights back on to see better (since the other fixtures still worked), but didn't notice that the now bare and exposed hot and neutral wires from the fixture I just pulled were touching.

When I turned on the switch, the lights did not come on and you could hear a deep growling noise somewhere in the house, very similar to water hammer noise. I turned off the switch right away, not knowing the cause. I turned on the switch again with the same result. After turning off the switch, I inspected the wiring from the fixture and discovered the short.

As I understand it, the time it takes a breaker to trip is basically inversely proportional to the current draw. A dead short should be VERY high current, and I would have expected the breaker to trip instantly. It did not do so. Questions:

Shouldn't the breaker have tripped? If so, I can't trust the breaker and should replace it. But, how do I go about testing the other breakers in the box? They are all old... probably original to the 1964 house. What was that noise?

  • You might want to research the make and model panel and breakers. In the past there have been devices that were found to be defective, and require replacement.
    – Tester101
    Sep 26, 2014 at 21:36
  • Could the house have underrated aluminum wiring?
    – sborsher
    Oct 1, 2014 at 19:16
  • 1
    No, everything is 12ga copper. Oct 1, 2014 at 21:10
  • Go look at your panelbox, 1964 sounds about right for you to have a Federal Pacific panelbox. They cheated on their certifications with rigged test equipment. They are responsible for many house fires. If you have one of their panels, it needs to be completely replaced. inspectapedia.com/fpe/FPE_Stab_Lok_Hazards.php
    – Mister Tea
    Sep 26, 2016 at 19:52
  • @MisterTea, thanks for the heads up. However, the panel was an old GE Load Center box as shown in this related post: diy.stackexchange.com/questions/49591/ge-load-center-qs Sep 27, 2016 at 22:38

3 Answers 3


Does it really need (not should) to be replaced: Is there a lifetime failure count for a breaker? -StackEx

Here's some extra Yahoo answers nonsense, from my attempted search in finding a way to safely make a breaker pop.

The other answers to your question here bring valid concern about replacing a questionable breaker. I'd be more concerned with the condition of the wire, which I believe was the noise (rattling in the pipe). Circuit breakers are pretty good about failing in a safe state, I.e, it just won't work anymore. I wouldn't be so concerned that a monetary/slight (crappy) contact didn't trip the breaker; you did not achieve a dead enough short (see wallyk's). It was not a good thing to have happened (although not the worst) and if you're not going to pull the wire it may be safer to replace the breaker with an arc fault.

THQL1120AF, an arc fault for $50, except it's not a tandem. No good unless it happens to be the one single you have there. aplussupply.com sells standard replacement breakers for your panel. I was unable to find a THQL1120 (AF or GFCI) tandem breaker. FYI, I looked up a 100amp AF for that panel, it's $800...

You could swap that single for a new regular tandem, freeing an entire slot (by taking the other wire from the bad breaker (provided that it was on the same leg originally) for a 1" AF, now hooked to the circuits' questionable wire. Request clarification if you do not understand the importance on which of the legs a breaker is on and how to properly relocate them.

Go big or go home. Meaning, as discussed here, replace that panel or leave it alone as much as possible.

  • Mazura, thanks for your thoughts and research. Per the advice of you and others, we're having a couple of electricians bid on panel replacement. I've inspected that branch where I can (all wire and splices at jboxes), but obviously have no way to inspect the wire in the conduit. We'll have new wire pulled on that branch at the same time we get the panel replaced... hopefully. It will come down to $$ unfortunately. Oct 3, 2014 at 18:02
  • Purchase what you can yourself to avoid up-chargers/selling. GL @bobfandango
    – Mazura
    Oct 3, 2014 at 18:26

That noise was a wire or junction/splice somewhere vibrating and heating up. It could have also been the breaker (trying to) trip. There's a good possibility that if you turn the breaker off, it will refuse to turn back on.

At a minimum, you should replace that breaker. You should also inspect that entire length of the wiring on that circuit as you may have melted a wirenut or wire insulation somewhere.

If this were my house (and my family's life) I would call an electrician and have them inspect for damage.


A dead short would have very high current. Fortunately, a dead short is not possible in a typical home.

The wires have resistance as do the connectors, wire nuts, and switch. There is probably nearly one ohm in your configuration. This limits the current to 120 amps—initially. As the point of closing the unwanted circuit, the contact gets hot, perhaps sparking. That considerably increases the resistance which further limits the current.

You probably were hearing the breaker trying to trip. I agree that it should be replaced. If it is not a GFI nor AFCI, it should be inexpensive ($3–$7).

It would be useful and perhaps interesting to measure the short: Disconnect the wire from the circuit breaker, connect the neutral and hot wire which accidentally touched before, turn on the switch, and then use an ohmmeter to measure the resistance of the circuit between the wire (formerly) connected to the breaker and a neutral bar in the service panel. Then measure the meter's own resistance by touching the leads together. Subtract the latter from the former. (Example: suppose the full circuit shows 0.75 ohms and the meter leads show 0.05 ohms. The circuit resistance is 0.75 minus 0.05 which is 0.70 ohms.)

120 volts (in North America) divided by the resistance is the current flow. (For the example 120 / 0.7 = 171.4 amps.) If the current is more than 1.5 times the breaker rating, it should trip pretty quickly, not much more than a second if it is rated for motor starting, otherwise much less than a second.

  • I suppose I'm picking nits here, but 120 amps is very high current in my book. Especially since the house only has 100 amp service. Interesting idea re: measuring the resistance. I'm still trying to ID the breaker box and breakers (it's a very old GE box and the labels are all unreadable at this point), but when I get a new breaker, I will test it then. Sep 30, 2014 at 17:02
  • 1
    Assuming your figure of 120 amps, that is 6 times the current rating for the breaker on that branch. I looked at the Time Current Curves for the breaker (apps.geindustrial.com/publibrary/checkout/… at that current the breaker should take between 1 and 6 seconds to trip. The fault existed for substantially less than 6 seconds, and I would bet it was less than a full second. If that resistance value is correct, then a fully functional breaker of this type should not necessarily have tripped! Oct 3, 2014 at 22:08
  • @bobfandango: Something seems to have gone wrong with your link. The correct one is apps.geindustrial.com/publibrary/checkout/…
    – wallyk
    Oct 17, 2014 at 17:54
  • Thanks @wallyk, not sure what went wring with the link. Still have not had a chance to measure the resistance to the breaker. However, a 12ga conductor should have much less resistance than 1 ohm. Of course, you have to take into account the resistance of any wire nut splices. But I cannot find any data on typical connection tee resistances. Any ideas there? Oct 17, 2014 at 19:35
  • If a 120V circuit genuinely has 100 amps passing through it, then unless there's a large capacitive or inductive load (which have their own rules), something somewhere is going to be dissipating 12,000 watts. Unless one of the devices on the circuit is designed to absorb a lot of heat, something is going to get really hot really quickly.
    – supercat
    Jan 23, 2015 at 17:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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