We have a little beach hut, and everything metal is oxidizing like crazy because of the salt water atmosphere. The electrical installation is very "artisan", so to speak.

I've noticed that the lights dim way more than they used to do a couple of months ago when we use high amperage appliances.

I suspect that some of the connections have oxidized quite a bit, and may be the source of these voltage drops.

How can I isolate the cause? Would it be as simple as measuring resistance? How many ohms could I expect to see on a slightly oxidized or lose connection?

Also: does "contact cleaner" spray help in those cases? Is there some product (maybe a conductive spray or gel) that I could apply to contacts so that they don't oxidize as quickly?

  • 1
    I dk enough about your "artisan" electrical setup so this is just a comment, but ANY resistance is a big problem. Connections with resistance tend to heat up, which is def not acceptable. All you can really do is re-do the connections. I'd start with the main feed, which is probably aluminum which is more prone to corrosion than copper. You'll have to work with the power company to pull the meter. Then pull the main wires, check for corrosion....if there is any and its not too bad you could clean them up with a wire brush.....continued below Feb 28, 2023 at 15:25
  • You don't want a conductive gel or spray. Electrical joint compound is electrically non-conductive. It coats the wires and prevents oxygen from reaching the connection. Conduction is direct wire to wire with the wires being pressed together by wire nuts. Better is the AlumiConn block connector if the wires are aluminum. Do you have aluminum wires? If you have copper wires, then a first try that might work is simply to tighten the screw connections. Get good joint compound like Burndy Penetrox and put some inside each wire nut before screwing it on. Feb 28, 2023 at 15:26
  • 2
    if the corrosion is bad, you'll have to cut it off, re-strip, hopefully there is enough slack that would make this possible. In either case be sure apply the NOALOX goop to prevent more corrosion. Feb 28, 2023 at 15:28
  • 1
    Burndy Penetrox is far superior to NoAlOx according to authority Jesse Aronstein. I used NoAlOx 40 years ago with seemingly acceptable results, but switched to Penetrox 20 years ago based on Aronstein's online reports. Feb 28, 2023 at 15:33
  • 1
    If the wire ends are overheated so as to damage the insulation you should remove the damaged ends to get to undamaged conductor. If a lot of conductor is damaged you might have to pigtail. If the wire is corroded but not overheated, then you should burnish with fine emery cloth or paper with joint compound on the wire, then use a new wire nut or other fastener. Feb 28, 2023 at 15:38

3 Answers 3


The amount of work you'll have to do to measure it is greater than the amount of work to clean it up. So just take all the relevant points apart and visually inspect.

Well, that's not quite true. The resistance that you are imagining exists, makes A LOT of heat. You will see it on a FLIR or feel it on equipment. Not that I would necessarily be touching equipment in an "artisan" installation LOL.

If you're using backstab connections, you can't inspect those, and they're unreliable in the best of conditions, so I would advise not using them.

Also if any electrical work is "artisanal" I would consider making it fully up to Code. Certainly miss no opportunity to make the most of water-resistant kit.

One thing you may be able to resolve by measurement is a "Lost Neutral". On overhead service wire drops, it's 2 insulated wires wrapped around a bare "carrier" wire. The carrier wire carries the weight of the whole shebang, and it whips in the wind for 30 years, and the carrier takes the brunt of the metal fatigue. It often "snaps" often visually obviously.

When you lose neutral, the two hot wires are still 240V apart, but the two phases are no longer 120V to neutral. They are ??? and ??? to neutral, and the two numbers add up to 240V. So you might have 110V and 130V... 90V and 150V.. that kind of thing.

Anyway, Lost Neutral creates that same bouncy voltage effect when you turn on a large 120V appliance. It reduces voltage on its phase and raises it on the other phase. This can damage appliances on the high phase. I mention that, because it's super easy to test for.

So this isn't an oxidized connection, this is a broken connection with neutral current "failing over" to an alternate path... via your neutral-ground equipotential bond to your grounds, to your ground rods, through the earth, to the transformer's own ground rod, or the ground rod and N-G bond of a nearby house that's on the same transformer.

95% of the time, the Lost Neutral is in the utility's wiring to your house, which is in their bailiwick not yours. And good news here - they fix it for free.

  • Yes, I had a half-floating in that same hut once (before knowing about this), and almost tore my hair out trying to figure out the crazy measurements I got. That was fixed, and I now have relatively normal electricty - apart from the lights dimming when turning on heavy loads. The 700W water pump's inrush current is nasty (I'm looking for an NTC inrush limiter, but they aren't available here), and the 4000W water heater just dims things slightly. The whole thing is hooked up via a 400 meter, palm-tree mounted primary aluminium cable, so some drops are to be expected. Mar 1, 2023 at 2:23
  • But I guess you're right about fixing the connections being easier than measuring them. So I'll do that. Any tips how to protect these salt water atmosphere exposed connections from further corrosion? Mar 1, 2023 at 2:24
  • 1
    @Moritz Good enclosures! I would get the best water-sealed enclosures you can get. If you have to, build a little hut around it, so it is functionally indoors. Note the working space requirements of 30" wide, 36" standback room and 78" tall. And use the Noalox paste intended for aluminum wires. Note that most lugs are aluminum, so you usually actually do have mixed metals. Yes, you can hop on a voltage drop calculator and ask it what voltage drop to expect from 400m of 4/0 cable for 120V @ 12A (a typical large 120V load). 120V is the worst, 240V has less voltage drop. Mar 1, 2023 at 3:52

Resistance should be very low between any two points on your neutral and/or ground wires throughout the building. So low that a meter set to beep on items that are connected should always beep. Specific readings may help figure this out, but the real fix is to go through everything to check for bad or loose connections.

Typical for the US is:

  • Screw connections from each circuit to the breaker panel (hot to breaker, neutral to breaker (e.g., if GFCI/breaker) or neutral bar, ground to neutral or ground bar.
  • Wire nuts or lever (Wago) connectors any place wires need to be joined together, which can include splits in a junction box that is "only" a junction box, pigtails to connect multiple items in a junction box (e.g., two switches powered from the same hot), connections to light fixtures or switches that have wires attached instead of screws, etc.
  • "Backstab" connections

Backstabs are the worst. On better quality receptacles and switches there is a way to put a straight wire under a piece of metal that is tightened by a screw. But on the cheaper ones there is a hole where you stick a wire (generally only 14 AWG but people stick 12 AWG in there too...) and those wires can get loose.

Screw connections are generally pretty solid, but if they were not done properly (e.g., wire hooked the wrong direction, insulation under the screw) they can go bad too.

Wire nuts and lever nuts are generally very reliable. But a wire nut that is not tightened properly or that has wires not stripped properly (insulation inside the wire nut) can have problems as well.

Methodical checking/fixing of everything is the best thing to do, but it takes time and effort. Assuming your setup uses wire nuts (open a couple of boxes and you'll find out), I'd grab a bag of assorted common sizes of wire nuts and have at it. Basically:

  • Any backstab - remove the wire, clean as needed, move to screw
  • Any screw connection - if it doesn't look right and/or any evidence of arcing, pitting, burning, etc. remove, clean as needed and replace
  • Any wire nut - unless they seem really, really good, remove wire nut, check the wiring, replace with a new wire nut
  • Any receptacle or switch that shows signs of problems, replace.

Replacements of standard receptacles and switches are inexpensive, so go a step up to "spec. grade", which gets you "screw to clamp" instead of backstabs and also (for receptacles) gets you "self-grounding", though that only works if you have metal boxes.

$100 in materials will go a long way. The time to do it all is the hard part.

  • 1
    Torque to spec as well…
    – Jon Custer
    Feb 28, 2023 at 15:31
  • 1
    To add to this: the way to test a connection with common tools is by testing voltage drop under load, not resistance. If your wall heater (for example) is on and you see 240V at the panel, 239V at the thermostat, and 218V at the heater, there's a poor connection somewhere between the thermostat and the heater. Not that it matters much in this particular scenario, because this answer absolutely lays out a reasonable approach that needs no testing.
    – KMJ
    Feb 28, 2023 at 15:48
  • Thank you - but how low a resitance is "VERY low" for a connection point? Low as in basically unmeasurable with a normal multimeter? Or A couple of Ohms? Since my voltage drop issue (lights dimming) mainly happens while the 700 Watt water pump is running, can I measure this resitance (carefully!) while the cables are energized? Feb 28, 2023 at 15:51
  • Measurable but pretty low. Quick search found: One thousand feet of 14 AWG solid wiring has a resistance of 2.53 ohms Feb 28, 2023 at 15:59
  • 1
    A couple of ohms is too high (bad). A good joint's resistance should be down in the several milliohm region, IMO. Resistances this low are difficult to measure directly, and usually require a 4-point measurement system, which eliminates errors caused by the probe and contact resistance.
    – SteveSh
    Feb 28, 2023 at 17:04

Directly measuring Ohms on power conductors is tricky, the resistance may not be ohmic - it may change depending on the voltage being used to measure it, also you have to turn the power off to make the measurement.

The safest way to measure the source resistance is to treat the source as a Thevenin source, and observe how much the voltage dips when you turn known loads on and off. (powerful heating appliances like hair dryers and toasters work well as test loads)

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