Rules and Guidelines
Want to participate in the discussions? Please read our Community Rules and Guidelines. Need some help? Visit Support and Answers.

  • State Helpful Answer
  • Replies 21 replies
  • Subscribers 42 subscribers
  • Views 452 views
  • Users 0 members are here
Options
You may also be interested in

Shower tray acting as capacitor? Measured 28v!

Asli

First post as I'm hoping someone here might be able to give me a good technical theory (aware without more points of data a true explanation might not be possible) which explains the phenomenon I have experienced recently.

It started with me receiving what I immediately recognised as a small shock from the metal tap in my shower, a small sharp tingle. An important detail is that I could not feel this shock with my bare hand (even soaking wet) only by chance a tiny cut on the edge of my thumb had touched and there, without fail I could feel it. Upon stepping out of the shower the shock disappeared, back in, it returned.

To make sure I wasn't going crazy, I grabbed my multimeter and took a measurement from the waste water to the metal tap, 20V. My gut feeling was some sort of conductive path based on the waste pipework as the voltage dissipated as the water in the waste drained away with the tap off.

Next thing I did was obviously call a qualified electrician out who tested my installation and could not find anything wrong, a real head scratcher for him. At the end of troubleshooting day one, the working theory was that the water and biofilm in the pipework combined with the water at my feet were creating a path to the cast iron waste pipework much further down in the building, that might not be bonded to the building's main earth point as the waste pipes into each property (its a block of flats) are all plastic so no need to bond.

Purely by chance I did some more testing later myself, and noticed that the voltage jumped up before the water had even had a chance to run down the waste for any time at all really, it seemed to occur as soon as there was a decent pool of water in the tray, this is when a ridiculous idea popped into my head, and I thought the plastic shower tray could be acting as a giant dielectric‽

Anyway I plugged the shower and filled the shower tray, so now the waste was completely out of the equation and the water was much more consistently covering the whole base of the shower, we're up to 28v! I also took a resistance measurement of 290kilohm to give a current of just under 1mA. Tiny, and explains why I can only feel it precisely on the tiny point where my skin is broken and nowhere else.

As for the source of the voltage, as no faults were found I am at this point assuming it is all leakage current, I have a few PCs, all smart lights, so both the socket circuit and the lighting circuit had an influence on the measured voltage.

With that in mind, is what I am proposing: a capacitively coupled pool of water in the shower tray technically feasible as a mechanism to introduce a different potential into that scenario to the earth in my flat, bonded to the shower taps via pipework, that is carrying some leakage current?

I still need to investigate under the shower of course, perhaps some structural metalwork, or other cabling? OR a very damp floor with an unknown slow leak that is conductive enough to join to some other structural point that eventually joins true earth independently of my electrical installation.

Fascinated by figuring this out, but also very confused currently.

EDIT:

Got access to underneath the shower, there is metal studwork on 3 sides of the shower. The metal base bar of the studwork runs around the sides of the shower tray base. The left wall to earth reads 33v and the right wall 28v. There is no damp, cabling, screws poking through the floor into the shower tray that might have broken the surface.

Is it technically possible for current induced in the metal studwork from live cables, to be capacitively coupled to the thin sheet of water pooled in the tray, which then passed to the grounded shower tap via me?

  • 0

    Hey folks, thanks so much for the responses, lots of interesting stuff.

    I used this as an excuse to buy a slightly more feature-full meter (not that I need much of an excuse to buy new tools anyway ha!) My current multimeter (used for wiring guitars mostly) a Fluke 101 did not have a Low impedance (LoZ) mode. I got one that did, that also happens to be a clamp meter capable of measuring AC current at a resolution of 0.001A. I unfortunately did not have the budget to stretch to something more sensitive than that like some of the earth leakage meters I'd seen.

    With the new meter I took new readings:

    Stud to shower: 36v

    Stud to shower (LoZ mode): 0.5v

    I then ran some crude temporary 'supplementary bonding' from the stud work to the shower out of curiosity to see if this would help:

    Stud to shower: 0v

    AC current with clamp around bonding cable running to the shower: 0.000A (likely outside of the range of this meter I assume, if it were 0.1mA it would not read this but sounds like it is in the perceptible range from what you say @Mike? Definitely possible with broken skin I would assume

    To answer some of the questions, the voltage disappears gradually as circuits are switched off, one of the first tests the electrician did was isolate circuits one by one, we found that the sockets contribute 10v, the lighting contributes 12v, smoke alarm 8v and a few others a few volts each, there was no single obvious source that might indicate a circuit that had a nick in the live against some metalwork or anything. It was this gradual buildup of the voltage and no obvious problems and nothing alarming on the insulation resistance tests he did (on every circuit) that makes it seems like an induced voltage, it was just the mechanism of transfer, and crucially what was coupling to the water that was baffling. I now know that to be the studs.

    The readings were taken with one probe on the stud, and one probe on the shower metalwork, from which the electrician already checked the resistance back to the CU to see if the bonding was sufficient and he was happy with the reading. The original reading was taken with one probe submerged in the plughole directly and the other problem on the shower.

    I could not feel it at all when standing on the shower with my normal, unbroken (but wet) skin, stood in the pool of water. It was exactly as @kfh mentions, a small crack on the side of my finger where the skin was broken AND wet. In my old job we used to repair computers and sometimes needed to troubleshoot them whilst on with exposed non enclosed power supplies, so we had some electrical safety training and I remember the resistance of broken skin being massively lower than other scenarios but don't remember the ranges given! So this made me think back to that.

    Anyway my conclusion so far is that:

    • Bonding helps, will discuss this with the electrician
    • Regs don't require bonding of metal studs and the like if they are completely enclosed in a wall and insulated that way
    • The current is tiny, the voltage appears to be an induced voltage on the studs and this is probably safe?
    • The mechanism of the coupling of the water to the studs is possible, albeit incredibly unusual?
  • 0 in reply to Chris Pearson

    As a teenager I suffered a very bad shock,  during which I was not able to breath while a current that left lovely entry and exit burn scars on my hands flowed until I was disconnected by  friend who saved my life.

    Apparently while the current was flowing I was making a loud 50Hz growling noise, and then dropped to the floor when the power was cut. I dont actually remember that but I do remember looking up at a lot of worried faces.

    Based on this, I can say that 'lungs' are indeed parts that do not work well while current flows, and as we die without them working, I'd put them on the 'essential 'list.

    The start and stop of the defibrillator is indeed a lethal level, but the time shape of the pulse is critical, and  it is also critical not to apply the shock if there is a healthy heartbeat, so the machines check for that.

    https://www.ahajournals.org/doi/10.1161/01.cir.97.16.1654

    Mike

  • 0
    Asli said:
    metal tap, 20V.

    Is there any reason why there cannot be 20-volts on the tap when measured against true earth?

  • 0 in reply to Sparkingchip

    No there isn't, that wasn't tested, he did however test the tap against the supply earth with the main bonding conductor disconnected, and there was no voltage there, just against the studs. Do you think it is worth asking him to do this? Perhaps the wander lead out the window and a screwdriver in the ground as mentioned above

  • 0 in reply to Sparkingchip

    Is there any reason why the cannot be 20-volts on the tap when measured against true earth?

    For s steady state 20V is a bit more than I'd hope for. With TN-C-S there can be some significant difference due to voltage drop along the supply PEN - but usually that isn't that high - with 3-phase distribution in the street N currents tend to cancel out so the overall voltage drop should be limited, even with the worst case of a long single-phase system the N voltage drop shouldn't be more than half the total allowable - 16% of 230V = 36.8V - so 18V say at worst.

    TT system with a lot of earth leakage (or worse an uncleared L-PE fault) could be (a lot) higher - but I suspect that's not the case for a block of flats.

       - Andy.

  • 0

    Is this a power shower?

    Things that may have 'electronics' in them often have EMC/RFI filters which explicitly divert supply/neutral currents to the CPC earth terminal via a "Y" capacitor, typ 0.1uF on each leg.

    These apparent low voltage leakages are tricky because of confused (conflicting) ideas about what makes things safe. On the one hand we have hard 'earthing' that maximises 'fault' conduction, then on the other hand is (double?) 'insulation', all  sandwiched around stray capacitance and path resistances, and alternate paths [and the possible red herring of 'True Earth' - you are not in that loop Wink ..] 

  • 0 in reply to mapj1
    mapj1 said:
    Based on this, I can say that 'lungs' are indeed parts that do not work well while current flows, and as we die without them working, I'd put them on the 'essential 'list.

    I see what you mean. However, it was the muscles which work the bellows rather than the bellows themselves which were affected - the same phenomenon as cannot let go.

    In any event, one can quite safely hold one's breath for a few minutes.

  • 0 in reply to Asli

    I'm relieved that in the low Z mode, the voltage falls away. It is a bit odd that the voltage rises the more circuits are active, but perhaps there is wiring from each circuit clipped  or tied to the metal studs and so capacitvely coupled L_E  
    Regs dont need the studs to be earthed but, like foiled plasterboard they often are, kind of fortuitously ,by the back boxes for light switches and so on. If it eliminates a risk, then a length of wire held by a self tapping screw into the stud at an exposed end, and then connected to the CPC ('earth') of the nearest light switch or socket would remove that voltage difference.
    Personally I'd suggest do the earth in the garden test at a convenient moment, as it is always reassuring to know that the company supplied 'earth' is still connected to the real thing somewhere.
    Mike

  • 0 in reply to Philip Oakley

    Not a power shower no, fed by a PulsaCoil electric water heater, as is the rest of the flat. Voltage is still present with the circuit for that switched off however.

    Ah yes I had read this about modern electronics! As I understand it, this is why the electrician asked me to unplug everything in the house before doing the RCD tests? He noted that any leakage from appliances like this could skew the result.

    The only reason I was thinking about leakage current at all was when trying to theorise the source of this mystery voltage, I had initially thought that it must be leakage on my installation's earth, flowing to a different earth (metal waste pipework, structural steel etc.). But as others have pointed out calculating the 'direction' is not so simple. And it now seems that induced voltage on the studs flowing to my earth is just as likely.

    I think the second explanation is increasingly more likely though, as the electrician was verifying my readings with all the appliances disconnected after having done his tests. If it were leakage from appliances one would expect the voltage to have dropped off somewhat after unplugging them all?

  • 0

    there is metal studwork on 3 sides of the shower.

    One (rare) possibility, but one it might be good to eliminate, is the "live wall" situation. A damaged cable can connect sections of steel studs to live conductors (could as easily be neutral as line). The connection may be be pretty poor (e.g. conductor not quite in contact with the steelwork or one bit of steel not in contact with the next, or part of the path going via plasterboard (which contains salts). Unless the steel studs happen to be connected to the electrical earthing system somewhere, the fault can be invisible to a conventional insulation resistance test.

    There have been one or two nasty incidents with live steel studwork - which is why soft sheathed cables concealed in such walls now need 30mA RCD protection regardless of depth - and the situation can change over time (e.g. what was a high resistance can become much lower if the wall becomes damp).

    An insulation test to the studs, rather than the MET, might be enlightening.

    Alternatively it might be just capacitive coupling between perfectly healthy installation cables and the studwork - but you would have thought that it would tend to cancel out on average as you'd have the N/PE between L and stud as often as L being closest.

       - Andy.

Community Rules & Guidelines