6

I recently moved into a house that has multiple outdoor outlets eventually feeding an outdoor lighting transformer and a small waterfall pump. There is an outlet on the interior of the house first before the wiring goes outside, originally with two GFCI outlets at the end wired in parallel, like so:

Breaker -> Outlet (indoor) -> Outlet (outdoor) -> Outlet (outdoor) -> GFCI (lighting) LINE -> GFCI (pump)

My plan was to replace the first (interior) outlet with a GFCI, and then all others connected to the Load side, with modern WR rated receptacles.

Breaker -> GFCI (indoor) LOAD -> WR Outlet (outdoor) -> WR Outlet (outdoor) -> WR Outlet (lighting) -> WR Outlet (pump)

The outdoor wiring is direct burial (not in conduit), and based on the condition of the outdoor receptacles I'd say is about the age of the house (c. 1975). The markings are "E30445 (UL) AWG 14 CU 2 CDR WITH AWG 14 GROUND TYPE UF-B 600 VOLTS SUNLIGHT RESISTANT". Most, if not all, of the outdoor outlets were plain old 1970's indoor-spec (non-WR), which were rusted, corroded and in some cases cracked (and yes, they were all still hot!). Note: the first two outlets are typically unused, and used only for the occasional leaf blower or some temporary outdoor device.

I have two separate but related questions/issues:

  1. Is it fine, better or worse to have all of these outlets protected by a single upstream GFCI? The installation instructions for both lighting transformer and pump explicitly say they should be plugged in to a GFCI receptacle directly, which I assume is why it was done the way it was. What I've read (link, link) the issue with GFCI's in series is the nuisance of having to determine which load caused the fault. I'm not clear if having a second GFCI (in parallel) with the lighting/pump plugged directly in, if there is a GFCI upstream, is required or adds any additional safety.

  2. When I replaced the first/indoor outlet with the GFCI (a Leviton GFTR1) as shown, it immediately faulted. I disconnected each outlet up the chain and once I disconnected the last outlet (pump), it stopped faulting and was fine for the past 4 months. Last week I had turned off the breaker, and when I turned it back on the GFCI faulted and wouldn't turn back on for two days (though everything was fine until I turned power off and then back on a few minutes later). I had just had some landscaping done, and so the ground had been super-watered, not sure if that would have had anything to do with it. Once it came back, it has not faulted again. My question then is... is this a problem or concern, or is this more characteristic of the high sensitivity of a modern GFCI with older outdoor buried wiring? As in, so long as the GFCI remains on is it safe to keep in operation?

Thanks!

640KB
  • 161
  • 4
  • The buried wiring, what is the cable type? I.e. what markings? – Harper - Reinstate Monica Jul 08 '21 at 14:42
  • I'm mainly interested in if it's NM or UF cable. UF is very wide/thin and the wires are embedded in the plastic jacket. NM has an outer sheath that merely wraps around the wires, and paper packing around the ground wire. – Harper - Reinstate Monica Jul 08 '21 at 15:16
  • 2
    @Harper-ReinstateMonica wait, found it: "E30445 (UL) AWG 14 CU 2 CDR WITH AWG 14 GROUND TYPE UF-B 600 VOLTS SUNLIGHT RESISTANT". – 640KB Jul 08 '21 at 15:25
  • 1
    That cable should be fine for another decade at least, probably two. It has no priority compared to the rest of the situation. – Mast Jul 09 '21 at 08:11

2 Answers2

9

Is it fine, better or worse to have all of these outlets protected by a single upstream GFCI?

Definitely better inside. A GFCI contains sensitive electronics that will normally last much longer inside in controlled temperature and away from rain, snow, dust, wind, etc.

The instructions are either unclear (extremely common) or it may be that they do specifically refer to a "GFCI receptacle" because with many older houses that is by far the easiest way to add GFCI protection. But (a) if you have multiple GFCI in a circuit with one feeding off the Load connections of another then you do have the multiple trip problem and (b) the protection is identical whether at every receptacle, first outside receptacle, last (or any) inside receptacle or breaker. General recommendation is "last inside receptacle" because that protects the outside, is generally (based on typical circuit layout) easiest to "match" to the outside for reset when needed and is usually less expensive than a GFCI breaker.

When I replaced the first/indoor outlet with the GFCI (a Leviton GFTR1) as shown, it immediately faulted.

Perfect! The GFCI is doing its job as designed. It protected from a faulty pump. It later protected from typical water problems. "Older outdoor buried wiring" is not the problem - though a problem in that wiring would indeed trip the GFCI. You have the far more typical "pump gone bad" and "water got into stuff" problems. The water dried up and that particular problem went away. The GFCI is doing what it is supposed to do. Yes, newer GFCIs are often more sensitive than older ones. But given the scenario - outside with water - that is a very good thing.

There are exceptions. Typical is a refrigerator or a fire alarm control panel. Both have life-safety consequences (unknown food spoilage, unannounced fire) and so there are some exceptions with respect to AFCI and GFCI for those devices. Safety is always a balance - can't make the entire plane as durable as the "black box" because it would (literally) never fly. But outdoor stuff on GFCI? Hard to see any downside - if an old pump or wires need replacement, that's better than people needing replacement.

manassehkatz-Moving 2 Codidact
  • 115,769
  • 11
  • 130
  • 326
  • 1
    Got it, so I'll go ahead and use the plain WR receptacles for the pump and lighting. – 640KB Jul 08 '21 at 14:54
  • 1
    One other thing to note though, the pump is switched and would have been off when the GFCI tripped the first time. Also, it was actually unplugged when the GFCI tripped the second time (and wouldn't reset). It was only after I waited 2-3 days and tried it again did it stay on, but nothing had changed... – 640KB Jul 08 '21 at 14:56
  • That's because you had two different problems. Pump - worn out insulation or some other problem resulting in a small current leak, and if that happened when switched off, problem must have been before the switch, which happen. Water ingestion elsewhere - one of the other receptacles or a junction box somewhere that got soaked and dried out after a couple of days - which definitely happens. – manassehkatz-Moving 2 Codidact Jul 08 '21 at 15:04
  • 2
    On the pump switch, is both hot and neutral switched? Because the GFCI is monitoring both hot and neutral. Most of the time switches are wired to only disconnect hot, which does not isolate the equipment at all. – Harper - Reinstate Monica Jul 08 '21 at 15:11
  • 1
    @Harper-ReinstateMonica In conclusion then, It sounds like this is the kind of thing to keep on watch, if the GFCI continues to fault in the future (with the pump unplugged) then the issue is wiring related. If not, then it was just a fluke or I replace the pump someday (not nearly as expensive as trenching up the lawn!). Thanks! – 640KB Jul 08 '21 at 15:42
  • 3
    "open hot" sounds like something one of those Magic 8-ball testers would say. Those things are not made to troubleshoot old wiring, and are worthless for that task. They are for checking new construction for wiring mistakes, and are by no means conclusive (they don't list every fault). However on a classic magic 8-ball tester, "open hot" means no lights at all, which simply means hot is disconnected. – Harper - Reinstate Monica Jul 08 '21 at 15:58
  • @Harper-ReinstateMonica just to throw one more stitch into it, I just went out to plug in the waterfall pump again, and noticed the plug is only two prong, no ground pin. Is that... bad? – 640KB Jul 08 '21 at 16:31
  • Actually, if GFCI protected then lack of ground pin doesn't matter so much. In fact, GFCI can actually be used to protect circuits that don't have ground available. The G in GFCI, a bit counter intuitively, doesn't need a ground wire to do its job. – manassehkatz-Moving 2 Codidact Jul 08 '21 at 16:53
  • @manassehkatz-Moving2Codidact Got it, as in, the same way you can install a GFCI receptacle in an older home that doesn't have a proper ground (so long as you put on the little stickers). Hypothetically would that make it less likely that the pump could cause a ground fault since it's not directly connected to ground? Or wait, I suppose it is connected to ground by virtue of the fact that it's sitting in water in the ground... – 640KB Jul 08 '21 at 17:14
  • 1
    Yes on the first part (stickers for No Ground) etc. No on the second part. In a perfect world you want that ground because the classic situation is: appliance has metal case, short of hot to case trips breaker, short of neutral to case trips GFCI. But in both of those cases, if you do not have a fully connected ground path (appliance case to ground pin on plug to receptacle all the way back to the main panel) then no current flows despite the fault. Which is OK until a person gets in the path. If a person is in the path (touches metal case, feet on the ground and not super-insulated) – manassehkatz-Moving 2 Codidact Jul 08 '21 at 17:17
  • 1
    as in, we really want to prevent this. – 640KB Jul 08 '21 at 17:25
  • 1
    then the circuit completes through the person. Now you have a big dangerous problem. GFCI comes to the rescue by detecting the imbalance between hot & neutral. However, in those same situations if you have a proper ground then either the breaker or the GFCI would have tripped before, likely before anyone touched the metal case, avoiding the situation getting so close to fatal. – manassehkatz-Moving 2 Codidact Jul 08 '21 at 17:27
  • Yeah, that happens a lot. Often takes out the first wave or two of rescuer, also. There was a fountain incident in Texas where it took out a girl, took out the 2nd girl who tried to rescue, took out the 2 additional who tried to rescue them, and then the firemen were getting nailed until one figured out how to fire-axe the supply. Lost all 4 girls. – Harper - Reinstate Monica Jul 08 '21 at 19:25
5

When several GFCI receptacles are in a circuit, all daisy-chained off "LINE", and with the "Load" terminals not used... that means nothing is protected except things plugged into those sockets.

Whereas anything hooked to the "LOAD" side of any GFCI device will protect everything - the wires, the sockets, and the things plugged into the sockets. We prefer that because of elegant economies: It means protecting more with fewer expensive GFCI devices, and it means getting the GFCI out of the rain. And protecting more stuff (e.g. the wires and sockets) is always better. Right?

Of course, when we protect more stuff, we increase our risk of a trip. You are adding a bunch of stuff (mainly: wiring) that was never GFCI protected before. This stuff may have a pre-existing ground fault, that had been present all along. That will cause the new GFCI to trip and be unable to reset.

So it often necessitates a "bug hunt" into the existing circuit to find this ground fault that has always been there. Just like any other detector: if you get a radon or CO detector and it immediately goes off, you know you had the problem all along.

Harper - Reinstate Monica
  • 300,628
  • 26
  • 286
  • 734
  • 1
    If a new detector immediately goes off when installed, it doesn't necessarily mean there's a preexisting problem (although it usually does); it could mean that you bought a faulty detector. (Just being pedantic.) – Vikki Jul 09 '21 at 00:21
  • Occams razor always has a sharp blade. – manassehkatz-Moving 2 Codidact Jul 09 '21 at 00:24
  • 1
    @Vikki and if you want to be pedantic, "we haven't conclusively ruled out vaccines causing autism". But the problem with both those statements is they are extremely improbable... but due to cognitive bias, we have a population who is *spring-loaded* to believe them. Hang out on here awhile- the vast majority of GFCI trip-ees have that inexplicable and irrational bias toward "faulty GFCI". So your pedantic assertion is not a lie, but it has the same popular effect as a lie. – Harper - Reinstate Monica Jul 09 '21 at 15:45