I've got a DC powered ham radio in a spare bedroom converted to an office, with a coax feedline running up to an antenna up in my attic near the ridge. About six months ago, I upgraded my feedline to a low-loss LMR400, and only since then transmitting often trips one of two AFCIs.

The office with the radio and a nearby bathroom are powered by an MWBC. In the panel, the MWBC is fed by two Eaton BR AFCIs next to each other that each feed a phase of the MWBC properly(from each phase of the main). Both have their neutral pigtails to the neutral bar, but the MWBC shared neutral only goes to the breaker for the bathroom. I've checked the GFCI and downstream receptacle in the bathroom with a plug tester, and also verified that the office is not protected by GFCI that is sometimes built into AFCIs. Here's a diagram of the situation, ground not shown:

wiring diagram, ground not shown

When I transmit for more than a second or two, one of two Eaton BR AFCIs will trip with blinker code 6 (self-test failure), once that occurs the other won't trip. It seems random which AFCI trips. If it's the breaker for the office, the radio obviously stops getting power. If it's the breaker for the bathroom, radio continues to work without tripping its AFCI.

Things I've tried that did not help:

  • Replacing the AFCI breakers with two news ones from the supply shop.
  • Replacing the AFCI breakers under warranty with Eaton - the old model the house came with had a known succeptibility to radio frequency interference. The replacements ended up being the same ones from the supply shop.
  • Powering radio power supply through an isolation transformer.
  • Powering radio power supply from extension cord from another circuit on first floor.
  • Powering radio with two different AC->DC power supplies.
  • Adding ferrite beads to outside of coaxial feedline to reduce stray RF coming off the feedline - still trips an AFCI.
  • Doing the above with everything on those circuits but the radio disconnected.

Things that I've tried that did help:

  • Unplugging antenna coax and replacing with a dummy load. Since the breakers do not trip when transmitting, I believe the nuisance trips are caused by radio frequency interference(RFI) with the house wiring.

My locality does not yet require arc-fault protection, but they're a good idea so I'd like to keep arc-fault protection for both phases if there's a way to do so without nuisance tripping when using my radio.

I know that when powering an MWBC with two breakers, they need to be handle-tied so the whole MWBC shuts off at once, so even outside of the radio tripping, I should be doing something about that anyway.

For Eaton BR AFCIs, is the neutral lead only used to power the electronics inside the breaker, or is it actually used for fault sensing like it is on a GFCI? If so, is it expected that installing a 2-phase AFCI breaker will fix my nuisance tripping on transmit? Or should I go to a 2-phase thermal breaker? This is the current state after replacing the originals with the recommended ones from Eaton support:

current panel state, sorry for potato quality due to 2mb limit

  • 1
    I do not enough about AFCIs, but they seem to sense electric arcs/sparks. I think they do this by reading the electrical inference(radio transmission) caused by the arcs. If so then seeing ham radios are not as common in most homes, AFCI breakers might not be made to not see ham/radio transmissions as not arcs/sparks. I wonder if I would have similar problems with a 100,000 watt fm transmitter beside(600ft away) my house(no AFCIs).
    – crip659
    Oct 15, 2023 at 23:39
  • 4
    The Ham radio SE might have some tips.
    – Jon Custer
    Oct 16, 2023 at 0:03
  • @crip659, Doubling tx power only increases field strength by sqrt(2), but doubling distance decreases field strength by 2. I eyballed it, and the field strength at my MWBC (100w transmitter at maybe 10ft) might be roughly double compared to a 100kw transmitter at 600ft.
    – rsaxvc
    Oct 16, 2023 at 0:11
  • You were on the right track with the ferrite bead, but putting one on the coax is pointless. You need a ferrite bead on every conductor that attaches to the AFCI device and as close as possible to the AFCI device to block the RF from your transmitter.
    – MTA
    Oct 16, 2023 at 3:20
  • 1
    Contact Eaton. They know about the issue.
    – user71659
    Oct 17, 2023 at 0:59

2 Answers 2


In the panel, the MWBC is fed by two Eaton BR AFCIs next to each other that each feed a phase of the MWBC properly(from each phase of the main)

AFCIs do not work that way.

One can't just install two single-pole AFCIs on a MWBC and do "whatever" with the neutral wire. It doesn't sort itself out.

NEC 110.3(B) says for a code-legal install, one must to follow installation instructions. That wasn't done here. Searching for reasons amounts to "I installed it improperly, why does it fail?" For what it's worth, most AFCIs, including Eaton's, have built-in GFPE (low sensitivity GFCI) - this is how they detect certain arc faults which are also ground faults. As such, they have the same constraints on neutral sharing that a GFCI does.

The only thing that I can imagine is that you've gotten away with it for awhile because of the way the circuit halves have been used, or because of the wiring in the walls being other than the way you expect. But now that you are putting the circuit to earnest use, reality has caught up.

AFCIs and MWBCs do not play well together, which is why MWBC fell out of favor in the AFCI/GFCI age. If someone chose a MWBC for new construction, then they are fools. If someone chose to optionally retrofit AFCIs onto a legacy MWBC, well, that does not work... and the option in the Eatonverse is to use a 2-pole AFCI. Or, since code does not require AFCI at the breaker for legacy upgrades, simply put AFCI at the first receptacle. Combination AFCI/GFCI receptacles are available.

By the way, for the sake of the installer who wants the labor economy of MWBCs but still needs AFCIs to work, they make "/2/2" cable - black, white-black, red, and white-red. Giving each hot its own independent neutral. Also very useful on post-2011 3-way loops!

  • 1
    I shouldn't've said properly, I just meant not wired to the same phase of the main.I considered the breakers seeing an ground-fault due to shared neutral, but the blink code doesn't agree, and I can run a hairdryer on either branch of the MWBC without tripping. A GFCI tester doesn't trip in the office, but maybe its not enough to trip the level of a GFPE?
    – rsaxvc
    Oct 16, 2023 at 0:46
  • 1
    @ThreePhaseEel, that's where I was going with the hairdryer test. It shouldn't trip GFPE in the bathroom(since hot+neutral are routed through that bathroom breaker), but should trip the office breaker, if it has GFPE, since there is no neutral current to match the 9A on hot.
    – rsaxvc
    Oct 16, 2023 at 2:39
  • 2
    @rsaxvc That is a mystery. However what's clear is you did not install it according to the instructions, violating NEC 110.3(B). There is good reason it should not work. The only remaining question is "why does it work as well as it does?" An academic question for AFCI designers. For a GPFE tester, temporarily miswire a socket to return neutral current on ground, and plug in a load. That should do it. But AFCIs don't promise to provide GFPE, as long as they meet UL tests, UL doesn't care. Oct 16, 2023 at 18:13
  • 1
    @Harper-ReinstateMonica, I'm aware, I didn't wire the panel, and I hope to correct this. I powered a lamp with a 40W incandescent bulb hot->ground in the office, lamp works, breaker doesn't trip, so it must not have GFPE sensing for even 300mA.
    – rsaxvc
    Oct 17, 2023 at 0:58
  • 1
    @rsaxvc sorry, my bad. The bottom line is if it's not installed properly, expecting it to work properly is hopeless. You can try pigtailing both hot wires onto the AFCI with the neutral wire, making it "not a MWBC anymore" and assuring all current returns via the same AFCI. Or if you don't mind violating NEC 110.3(B), you can install the new GE AFCIs, see "picture" 4, video. The main risk point is arcing at the bus stab due to dissimilar contact shapes. Oct 17, 2023 at 21:07

I know that when powering an MWBC with two breakers, they need to be handle-tied so the whole MWBC shuts off at once

That is incorrect. There are two different issues with pairing breakers:

  • Common Trip

This is where a load on one breaker automatically trips the breaker. My understanding is that this is required for combination 120V/240V loads, such as a typical clothes dryer or oven. That's because otherwise you could have a trip in one 120V leg and still have power coming into the appliance. I believe this would also apply to an MWBC that is used to power both 120V and 240V loads, but that is not so common - most of the time an MWBC is used to power a pair of separate 120V loads. Common trip can only be done by design of the breaker - i.e., a double breaker. I don't think you actually need common trip.

  • Common Shutoff

This is accomplished by a double breaker, so if you have a common trip requirement then you automatically take care of common shutoff. But it is also required for a 2 x 120V MWBC, and there it can be handled (pun intended) by a handle tie, though a double breaker is perfectly fine as well. Since a double breaker normally costs about the same as two single breakers, new installations (e.g., the MWBCs found as part of my recent panel replacement) will normally use a double breaker for any MWBCs.

Where it gets interesting is AFCI and GFCI. GFCI, by definition, monitors neutral. GFCI on a 240V-only circuit (which is relatively new) might not have neutral connected since it isn't used. But GFCI on an MWBC or a combination 120V/240V circuit (e.g., clothes dryer or oven) must be a double breaker because it simply won't work properly as two single breakers.

However, my understanding is that while some AFCI breakers do actually monitor neutral, most do not, or do so in a way that doesn't require a balanced circuit like a GFCI requires, unless it is a combination AFCI/GFCI breaker. As a result, two single breakers should be just as effective for AFCI functionality as a double breaker.

All that being said, it might be the case that a double breaker AFCI would perform better, perhaps with fewer false positives, than a pair of single breaker AFCIs. Unfortunately, at least in the case of Eaton BR, a double breaker AFCI currently costs ~ 3x a single breaker AFCI instead of 2x. But I'm pessimistic that this will solve the problem.

I think the key is:

I believe the nuisance trips are caused by radio frequency interference(RFI) with the house wiring.

If I understand things correctly:

  • The problem only happens on particular circuits, which happen to be the ones feeding the room with the radio equipment, but the problem happens even when the equipment is powered by another circuit
  • The problem does not happen if the antenna is disconnected

I suspect the problem lies with the route the wiring takes in the house. Does it run in or near the attic where the antenna is located? If so, rerouting the circuit may help. Alternatively, assuming the circuit uses typical NM cable, running individual wires in metal conduit may solve the problem.

Of course, wires run in metal conduit are also a good way to minimize the need for AFCI! As I understand it, if you have an AFCI requirement and can't install AFCI breakers on your panel, you can run metal conduit from the panel and install deadfront AFCI devices and then run NM cable for the rest of the circuit.

  • 1
    Your understanding is correct. The tripping circuits happen to be the closest to the antenna, only when transmitting out the antenna, regardless of power power. I believe most of the antenna is ~10 feet above the circuits, assuming they run around waist high in the wall. The feedline between radio and antenna does pass within a couple feet, but shouldn't have much RF outside the coax.
    – rsaxvc
    Oct 16, 2023 at 0:20

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.