Short circuit failures with 12/3 NM-B cable and NM clamp connectors

Short circuit failures with 12/3 NM-B cable and NM clamp connectors

Dear Mike,

I would like to bring to your attention a repeatable, hazardous fault arising from the combination of two common products UL listed for use in residential electrical wiring. This does not appear to be an inherent fault in the products, but rather an incompatibility that may require industry cooperation for resolution.

Fault Description

From informal testing it turns out that securing NM-B cable to a metal junction box through a standard metal 3/8” NM cable clamp connector creates the risk that even slight over-tightening of the clamp’s retaining screws will generate an immediate or subsequent short circuit fault between one or more conductors and box ground. This fault is 100% reproducible if the screws are sufficiently tightened, though there is some variability in the pressure required depending upon the alignment of this radially asymmetric cable with respect to the clamp connector.

Issues with excessive clamping force on Type-NM cable are not unknown. John Cangemi, in the UL publication THE CODE AUTHORITY®: ELECTRICAL CONNECTIONS, May 2005 issue writes:
Installer awareness of “securement, without damage” is a concern with regard to the satisfactory installation of Type NM cable, regardless of the outside diameter. Care should be taken when tightening clamps or driving staples – tighter is not always better.
While this is excellent general guidance, it does not appear to be sufficient in the situation I have outlined above. The specific problem is that the margin between insufficient clamping force and excessive clamping force can be small or non-existent with these specific components. Again under informal testing, there is no perceivable force-feedback cue to the installer that the clamping screws are sufficiently tight. One traditional test has been to wiggle the wire outside the box, tightening the clamp screws until the wire no longer moves inside the box. Informal testing shows that at this clamping force the cable has already shorted or been irreparably damaged.


This fault appears to be easily reproducible only with 12/3 cable and this particular style of metal NM cable clamp connectors. Cable with fewer conductors, such as 12/2 and 14/2, is flatter and has a natural alignment to the clamp connector that restricts the maximum force on the cable. Smaller cable, such as 14/3, is not crushed as much by the cable clamp at its maximum closure. We made no evaluation of combinations of cable, such as two 14/2 cables.

The testing methodology was to measure low-voltage short circuits. It appears likely that cable damage occurs even with clamping force below that required to cause low-voltage shorts. Such cable damage may result in a lowered dielectric strength of the insulation, increasing the vulnerability to an arc short-circuit failure from high-voltage transients. It’s also possible that both cable sheath and individual conductor insulations will yield to this clamping force over time, leading to apparently random failures years after the initial installation. In one testing case the failure occurred only after some minutes past the application of clamping force, though this cannot be strictly attributed to plastic creep as opposed to subtle movement of the cable.

The clamp connector consists of a cast, threaded frame with a stamped plate attached by two screws. The plate is embossed with a dimple, apparently designed to provide an increased peak retention force for small-diameter cables. Examination of failed cables shows that this dimple is the source of penetration into the conductors. Experimentally flattening this dimple eliminated the failures, as did bending the plate such that the minimum opening at the fully screwed-down position was increased from approximately 4 mm to about 5 mm.

Some anecdotes of similar failures are available on the Internet using Google, including one at That the problem is not more widely reported may be due in part to common use of alternative NM cable connectors and limited use of 12/3 cable.

From my limited personal experience, this failure did not occur with older Type NM cable using metal clamp connectors. The older cable was of larger diameter and provided more resistance feedback to the installer as the screws were tightened.


My ethical responsibilities as a member of the IEEE require that I report this problem as a potential hazard to the public. However I have neither the authority nor expertise to make a specific recommendation for appropriate mitigation of this hazard. However it does seem obvious that 12/3 cable and standard metal 3/8” NM cable clamp connectors should not be used together. Alternative non-metallic clamp connectors are widely available.

Best regards,

Richard Zulch
Zulch Laboratories, Inc.

I have to wonder what sort of knucklehead is tightening his romex connectors enough to part both the jacket and the conductor insulation on 12/3’s. This is first year apprentice stuff. Guess it gives these guys in labs something to experiment with.

The trick with most things like this (cable retention hardware) is to only twist the shaft of the screwdriver. That is tight enough. You start cranking down on the handle and you WILL overtighten it.

edit; I wonder how many of these are “screwgun inflicted wounds”? That is a great tool if you remember the torque limiter and only use what you need

Thats the thing…how many people greg out their probably even know their is a TORQUE limiter on the screwdriver…they hear the clicking and go…whats that for…how much ya wanna bet…

I think I mentioned it here before. There was connector out two years ago, real cheap made in India, that was notorious for piercing the jacket, and insulation of cables. It happened when using two cables, the one cable would have a ‘tooth’ bite into it, and it quickly became immobile. While the other kept sliding, no teeth on the other side, so an electrician would crank down to get the one cable secure, and in part dig into the other cable.

Bad design, I don’t see them anymore.


You really don’t have to have those cables that tight. That is why they require another support within 12" of the box. It is more to seal up the penetration than to “clamp” the cable. There should not be that much strain on a cable. In fact you can fish them with no support once the wall is closed up. You are really just trying to keep them away from the slings and arrows of outragous construction workers.

I recently happened across this thread in an internet search on an unrelated electrical topic. So, I signed up to post here because I may be able to contribute to this discussion.

As a licensed electrician for 30+ years, I am particularly drawn to the post by Richard Zulch. His findings regarding romex cable confirm my own, which were submitted to UL for study by the Massachusetts State Fire Marshal’s Office, to whom I reported my findings and submitted two samples in evidence of the same problem.

Further research into this problem has lead to some rather startling evidence: In a statistically significant sample of field short circuit tests of installed 120V circuit breakers, which I have submitted to the NEC and to UL, ~10% of the circuit breakers in the average home do NOT comply with the NEC (240.4) because they do NOT provide adequate protection from overcurrent under short circuit or ground fault conditions.

I have proposed changes to the UL and the NEC that need support. Please respond if you are interested in this topic.

What’s there to say. This is a four year old thread and to my knowledge not much has changed in NM squeeze connector design. There has been a design change in some brands of 3-wire NM cable in that now they are almost flat like the 2-wire versions. IMO the entire problem has been due to improper field installation practices as mentioned by a few others in this ancient thread.

How did you confirm that 10% of the breakers did not comply with the standard?


I used a listed test instrument to conduct short-circuit tests at 120V outlets. The instrument records the short-circuit current and the breaker trip-time in milliseconds. With this data reported by the test instrument I calculated the I2T heating that the #14 circuit was subjected to. In 104 out of 1017 tests, the value of I2T that was reached in the short-circuit test exceeded the maximum safe limit for that value in ICEA P32-382-2007. This is the standard cited in a new Informational Note under 240.4 in the 2011 NEC as the appropriate standard for determining the maximum safe limit of overcurrent resulting from a short-circuit.

lol…I just posted it…I did not author it…:wink: