DIY Home Improvement Forum banner
1 - 20 of 23 Posts

·
Registered
Joined
·
11 Posts
Discussion Starter · #1 ·
With the historic heat wave in the Pacific NW, I plugged in a portable condensing AC unit in my home office . This panel-protected AFCI circuit also serves a switch in another room, which is in a double gang box with a switch that is on a panel-protected GFCI circuit for that room. The GFCI panel breaker has been tripping , coincidentally beginning after I began running the floor AC unit in the office AFCI circuit.

These circuits do not share a neutral - there are no shared neutrals in the house (I know, because I wired this house myself). But they are wired into the same box, with the two circuits as segregated as is practical in a double-gang box.

The GFCI breaker and AFCI breaker (both Siemens) are on the same bus on the panel, but are far apart. I have not yet opened the panel to see if their neutrals are on the same bar.

Therefore, I suspect that EMI from inductive kickback from the AC unit on the AFCI circuit may be leaking onto GFCI circuit, causing the trip.

I really don't want to replace the GFCI panel breaker unless I can verify it's defective, and I'd prefer not to replace it with a plain panel breaker and and add a GFCI outlet downstream.

How could I test my hypothesis here? And if proven , what's the best solution?? an RC snubber on the AC unit? separate
 

·
Registered
Joined
·
30 Posts
Move the ac to another plug in another room that is on a different circuit.

What is the result? Motors cause arcs. arcs cause AFCI's to trip. If it trips again on another known AFCI circuit, my last suggestion would be to try it on a circuit without AFCI/GFCI protection. If it works, then its the AFCI doing it's job.
 

·
Registered
Building my last home
Joined
·
1,643 Posts
Was this posted as a joke? I have spent 50 years in the electrical trade and do not know what
EMI from inductive kickback is . How did you measure this and can you repeat the experiment?
There are electromagnetic pulses from very large and extreme events. think lightning.
Will the electric field get larger under a heavy load, maybe. Problem is the amount larger on wire sizes you have in your home would be really hard to even measure. If the wires are in conduit for example I can almost guarantee that the field would not get big enough to escape the conduit.

RC snubber, what is that?
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #5 ·
Move the ac to another plug in another room that is on a different circuit.

What is the result? Motors cause arcs. arcs cause AFCI's to trip. If it trips again on another known AFCI circuit, my last suggestion would be to try it on a circuit without AFCI/GFCI protection. If it works, then its the AFCI doing it's job.
To clarify ... the trip is on the adjacent GFCI circuit, not the AFCI circuit (where the AC unit is plugged into) .
My question is, can EMI from the AFCI circuit cause a ghost trip on the GFCI circuit?
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #6 ·
Was this posted as a joke? I have spent 50 years in the electrical trade and do not know what
EMI from inductive kickback is . How did you measure this and can you repeat the experiment?
There are electromagnetic pulses from very large and extreme events. think lightning.
Will the electric field get larger under a heavy load, maybe. Problem is the amount larger on wire sizes you have in your home would be really hard to even measure. If the wires are in conduit for example I can almost guarantee that the field would not get big enough to escape the conduit.

RC snubber, what is that?
I haven't yet done any measurements on this yet .. just reporting empirical data points:
  • The GFCI circuit has been trouble-free until I started using the plug-in AC unit on a different circuit, but one that is adjacent to the GFCI circuit inside the same double-gang box.
  • Both circuits are not conduited, both use 14/2+ground NM cable (Romex), but they are stapled next to each other in one room of the house. They use independent neutrals
As a 50+ year pro, I'm sure you know that GFCI breakers are vulnerable to tripping when large inductive loads (think motors) are applied to (and/or removed from) them due to transient current imbalances. Also, I would think you'd know that an RC snubber is a simple RC network placed in parallel to the load to prevent induced energy that is released when a circuit is opened from flowing back upstream (i.e. inductive kickback) . My question is : is it possible that there could be sufficient leaked electromagnetic energy from the kickback (EMI) to be detected on the GFCI circuit as a > 5mA current imbalance ?
 

·
Registered
Joined
·
30 Posts
Move the AC to another room, what happens? This is an interesting situation, but needs more trouble shooting.

What is the draw on the nameplate of the AC?

Are you 110% sure this isn't the same circuit. You know that they don't share a neutral, but you don't know how they're landed on the neutral bar?
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #9 ·
Move the AC to another room, what happens? This is an interesting situation, but needs more trouble shooting.

What is the draw on the nameplate of the AC?

Are you 110% sure this isn't the same circuit. You know that they don't share a neutral, but you don't know how they're landed on the neutral bar?
The nameplate draw is 10.8A .

100% sure they are different circuits - I wired this house myself and meticulously labeled everything. Correct, I'd have to open the panel and trace the neutrals. If they're on the same neutral bar, would that be the culprit, and would the solution be as simple as moving one of them to the other neutral bar?

I also ran across this interesting discussion:

My takeaway from this discussion is that due to the length of my GFCI branch (> 80 ft) , I may have to abandon the panel GFCI breaker in favor of a plain breaker , signal-condition at the home run, and start GFCI protection as far downstream as I can.
 

·
Registered
Joined
·
30 Posts
You plugged in an AC pulling 10A on a residential 15A breaker, with other stuff plugged into what you think is one circuit? Your immediate conclusion is EMI? Any breaker/brand type whatever should trip at 80% so in this case 12A. I have seen Iphone chargers trip temporaries on job sites. If this is in your home office you are looking at an overload problem somewhere.

Let’s eliminate some factors.
1. MOVE THE AC to another room TURN IT ON… what happens?
2. Remove the gfci receptacle fed by a gfci breaker… this is just expensive, replace with a standard receptacle. With ac on and the gfci receptacle removed does the gfci breaker trip? Then you’re on the same circuit almost guaranteed.
3. Are all devices pig tailed or fed thru “never ever backstab” check line and load on all GFCI
4. Is ac on wall behind bathroom? Was this a complete rewire? Could it be this is the one plug you didn’t mess with?
5. Do you have access or own a clamp meter/multimeter/?

This is not an insult to your skills, ability, or intelligence. Based on your verbiage you seem like you know the ropes, but it sounds like you’re looking for a unicorn of a problem. Not talking down to you just trying to help.
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #12 ·
You plugged in an AC pulling 10A on a residential 15A breaker, with other stuff plugged into what you think is one circuit? Your immediate conclusion is EMI? Any breaker/brand type whatever should trip at 80% so in this case 12A. I have seen Iphone chargers trip temporaries on job sites. If this is in your home office you are looking at an overload problem somewhere.

Let’s eliminate some factors.
1. MOVE THE AC to another room TURN IT ON… what happens?
2. Remove the gfci receptacle fed by a gfci breaker… this is just expensive, replace with a standard receptacle. With ac on and the gfci receptacle removed does the gfci breaker trip? Then you’re on the same circuit almost guaranteed.
3. Are all devices pig tailed or fed thru “never ever backstab” check line and load on all GFCI
4. Is ac on wall behind bathroom? Was this a complete rewire? Could it be this is the one plug you didn’t mess with?
5. Do you have access or own a clamp meter/multimeter/?

This is not an insult to your skills, ability, or intelligence. Based on your verbiage you seem like you know the ropes, but it sounds like you’re looking for a unicorn of a problem. Not talking down to you just trying to help.
More data for you:
  • The AC unit is a 20A AFCI breaker - there are NO 15A circuits in the house, by my design - this was a complete re-wire, done by me (removed every inch of aluminum wiring from a 1960s build). To be clear, the trip is on the adjacent GFCI breaker that the AC has never been plugged into, and trips with no known load when the AC on the AFCI circuit is on.
  • Running it in another room on a another AFCI circuit works just fine ... and the GFCI circuit (that was never running the AC) that was tripping, no longer trips.
  • Verified that the two circuits in the box are indeed separate by turning off breakers and checking what devices do (not) turn on.
  • The GFCI circuit (which the AC unit was never plugged into) is protected by a panel breaker, all downstream protected receptables are wired from pigtails. Total length of circuit probably exceeds 100 ft.
  • The AC was in a office/bedroom on an 20A AFCI circuit. The place where the two circuits come into close proximity is the double-gang box and NM cables stapled on the same stud.
  • I do own a clamp meter, but I haven't (yet) done any direct testing for EMI or spikes in the box - I'm not sure if I could reliably measure an inductive spike given typical DMM sample rates of 10Hz or less. I could trap a spike it with a digital oscilloscope, but I don't have one of those - I have other reasons to buy one, so that may be my next purchase.
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #13 ·
It doesn't work that way. A 15A breaker will trip with a load above 15A, not 12A. While a circuit should not have a continuous load of more than 80% of its rated ampacity, that is not the point at which the breaker will trip.
As already mentioned, The AC unit is on a 20A AFCI breaker - there are NO 15A circuits in the house, by my design. And remember, it's a different circuit tripping, not the circuit that the AC unit is plugged into.
 

·
Registered
Joined
·
1,524 Posts
More data for you:
  • The AC unit is a 20A AFCI breaker - there are NO 15A circuits in the house, by my design - this was a complete re-wire, done by me (removed every inch of aluminum wiring from a 1960s build). To be clear, the trip is on the adjacent GFCI breaker that the AC has never been plugged into, and trips with no known load when the AC on the AFCI circuit is on.
  • Running it in another room on a another AFCI circuit works just fine ... and the GFCI circuit (that was never running the AC) that was tripping, no longer trips.
  • Verified that the two circuits in the box are indeed separate by turning off breakers and checking what devices do (not) turn on.
  • The GFCI circuit (which the AC unit was never plugged into) is protected by a panel breaker, all downstream protected receptables are wired from pigtails. Total length of circuit probably exceeds 100 ft.
  • The AC was in a office/bedroom on an 20A AFCI circuit. The place where the two circuits come into close proximity is the double-gang box and NM cables stapled on the same stud.
  • I do own a clamp meter, but I haven't (yet) done any direct testing for EMI or spikes in the box - I'm not sure if I could reliably measure an inductive spike given typical DMM sample rates of 10Hz or less. I could trap a spike it with a digital oscilloscope, but I don't have one of those - I have other reasons to buy one, so that may be my next purchase.
Is the other AFCI breaker (that doesn't cause GFCI trips) on the same bus leg as the GFCI breaker? I have my doubts that there would be significant interference between the two circuits because the switches are in the same junction box.

I've seen a case where a Siemens AFCI breaker would trip due to some type of interference from another circuit. But I haven't seen/heard of similar problems with a GFCI.

Are you able to tell if the GFCI trip happens exactly when the AC compressor starts or stops? If so, does it happen every time, or just some times?
 

·
Registered
Joined
·
4,340 Posts
GFCI breakers are vulnerable to tripping when large inductive loads (think motors) are applied to (and/or removed from) them due to transient current imbalances.
My question is : is it possible that there could be sufficient leaked electromagnetic energy from the kickback (EMI) to be detected on the GFCI circuit as a > 5mA current imbalance ?
Well, think about how those transient current imbalances happen. Just as a capacitor will flow unlimited amps to resist changes in voltage, an inductor will kick unlimited voltage to resist changes in current. The voltage increases until insulation breakdown occurs somewhere... and then the inductor's current flows until it runs out of energy.

If the insulation breakdown is hot->ground or neutral-> ground, that is indeed a ground fault. Once the arcing starts, some normal current may flow as well until the voltage falls and the insulation is able to interrupt.

If that insulation breakdown happens in an area surveilled by a GFCI, it will trip.
 

·
Banned
Joined
·
328 Posts
No it cannot. It's illegal to have a black "hot" wire (say 14 guage) combined to make (say a single 12 guage equiv.) using two different breakers. No one would do that and it's illegal.

Hot must run to ground to trip a breaker. check your panel: there is only one line in, one line (a bus) out, and all are on the bus. The bus is INDIRECTLY connected to the white lines - that depends on the circuit diagram of your municipal power provider and you did not mention your locality or power company, NEC.

If it could do that one another, then it would do the same on all the breakers! They are all on the same copper rail.

You could have faulty grounding: likely - most buildings are not grounded by code, they are built by contractors who cross grounds everywhere without a care. (this is one big reason why installing generators to a panel requires a separate panel, btw, the ground paths are never what they are supposed to be). But bad grounding causes LATE TRIPPING or NO TRIP, not early trip.

YOUR PANEL / GFCI: they are famous for being loose. you can "wisely" check wire tighness on it. the blades on some devices get bent. (meaning, if there is one tightening screw that won't be enough). You have to take those off, adjust the blade, and click them back in (if they slide, then it's "not in", depending on the model completely, just a rough statement).

YOUR DEVICE. To work properly it must have the right ampherage and good grounding. If your electrical plug has a loose ground (or any wires loose): your device will pull or arc too much when starting, and set off a GVCI or whatever contraption they put in your box. If it's a 240V then the two hot line must be frequency balanced obviously, we all knew that.

YOUR DEVICE: if you have too sensitive a GFCI (infact you must choose one appropriate for the appliance, not just plug a major appliance to any plug). If I have a sensitive GFCI and plug in my vacuum it will trip off at the plug (has one built-in, outdoor plug). If I turn off the vacuum, plug it in, then use the vacuum switch it doesn't trip. Why: the vacuum switch sparks less and the GFCI doesn't see that switch.
 

·
Banned
Joined
·
328 Posts
I had EMI from inductive kickback once that I fixed. Some harry homeower had run a SINGLE CIRCUIT out to a plug/lamp 15 ft, back again, then up to a flourescent lamp with a large ballast. Whenever I turned on the lamp at the other switch (comes before the big loop and lamp): I'd see a spark come from behind the switch plate.

HOWEVER - that spark goes to the panel with the rest of the white wires not on ground (ground is not preferred for high voltage and amperage return, you remember). the EMI kickback wouldn't kick off a a gfci on another separate circuit. nope. no way.

(the electric code says that wire loops are not allowed, btw, they must terminate at the distance not come back)
 

·
Banned
Joined
·
328 Posts
"This panel-protected AFCI circuit also serves a switch in another room, which is in a double gang box with a switch that is on a panel-protected GFCI circuit for that room. The GFCI panel breaker has been tripping , coincidentally beginning after I began running the floor AC unit in the office AFCI circuit."

Obviously, NEC, we cannot see how that mess is all connected. The electrical company insures grounding is proper between boxes and doesn't allow non-certified persons to make connections. If your area ignores laws and codes: then your on your own. Don't live in a place that makes illegal wiring.
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #19 ·
No it cannot. It's illegal to have a black "hot" wire (say 14 guage) combined to make (say a single 12 guage equiv.) using two different breakers. No one would do that and it's illegal.
Sorry , mis-typed here ... all 20A circuits are 12/2 + ground. No 14/2 wire and no 15A breakers anywhere in the house. I wired the house myself and the inspector was impressed by the quality DIY work.
 

·
Registered
Joined
·
11 Posts
Discussion Starter · #20 ·
I had EMI from inductive kickback once that I fixed. Some harry homeower had run a SINGLE CIRCUIT out to a plug/lamp 15 ft, back again, then up to a flourescent lamp with a large ballast. Whenever I turned on the lamp at the other switch (comes before the big loop and lamp): I'd see a spark come from behind the switch plate.

HOWEVER - that spark goes to the panel with the rest of the white wires not on ground (ground is not preferred for high voltage and amperage return, you remember). the EMI kickback wouldn't kick off a a gfci on another separate circuit. nope. no way.

(the electric code says that wire loops are not allowed, btw, they must terminate at the distance not come back)
No loops, I can guarantee (I did my homework before embarking on this wiring project). The fact remains that when the AC unit is removed from the AFCI circuit, the GFCI circuit does not trip.
 
1 - 20 of 23 Posts
Top