Brand new. Granny charger trips RCD when charging. Replace how?

[WannabeOwner]

Thanks.

Easy to get and try to see if it is a solution to this problem?
Affordable?

 

[arg]

But what are chances of 2 houses being so close to 30ma limit

Note that the limit is not in effect 30mA but some lower figure - the standards for RCDs specify "must trip at 30mA within 250ms", "must trip at 150mA within 40ms", "must not trip at 15mA no matter how long" (figures from memory but I think they are right). So any given RCD will trip somewhere between 15mA and 30mA depending on manufacturing tolerances, and will usually trip faster the bigger the current.

What is the implications of this? Not tripping wise but non compliance wise?

A long and twisty tale of legislation.

The UMC itself must be CE-marked, which in turn means it must comply with the Low Voltage Directive (which regulates electrical safety of appliances). And indeed the UMC userguide has the legally-required "Declaration of Conformity" that confirms it does claim to comply with the LVD. The LVD itself just has broadly specified "essential requirements" about avoiding electric shock, fire etc. but there is then a long list of "harmonized documents" (mainly international standards) that you can use to show that your equipment meets the essential requirements. One possibility is EN60335 "Household and similar electrical appliances": that has limits on maximum leakage currents, ranging from 5mA for cookers to 3.5mA for washing machines, 0.75mA for some other classes of appliance. But it's not clear in which category Tesla will have placed the UMC for the purposes of their testing. Note that this lines up with what @Jeremy Harris was quoting from BS7671 - that is saying fixed wiring should be designed for max 3.5mA from 13A plugs, this then tries to ensure that the appliances you buy with a 13A plug on them are compatible with the fixed wiring.

The car on the other hand is exempt from CE marking, and instead is regulated under UN ECE. I know those regulations now have EMC requirements on plug-in EVs comparable to what's required under the EMC directive for CE marking, but I have no idea what (if anything) it says about leakage currents.

Then you have the fact that they could legitimately have tested the UMC on its own, or in conjunction with a "representative" vehicle such as a Model S, allowing it to be sold without regard to the different behaviour when used with a Model 3.

So if this all does turn out to be a real problem there is in theory a legal stick to beat Tesla with, but it's a rather long and unwieldy one.

 

[ianto999]

Will a PV inverter still 'leak' at night time? Would this vary with PV change?
I'm trying to find a pattern in my rather random RCD trips.

 

[RealEvil_UK]

FWIW I drove to Tesla West Drayton. It doesn't trip their RCD nor my brother's. But it trips mine and my parents. So that's 50% failure rate on a small but significant to me sample size.

I had a ongoing service chat via SMS and they stopped replying as soon as I said it worked at my brother's house despite the implication being that the UMC is leaking (who knows if it's more or less than 3.5)?

All in all I now have no clue if the UMC will work when I need it. I expressed this to Tesla and they have not replied. Their last suggestion was for me to contact an electrician to inspect my house. Which is fine if this was a fixed charger .. but this is the "emergency" charger.. maybe they expect me to call a sparky every time I go on a road trip in anticipation of it 50% tripping a RCD?

 

[bridgstfer]

Hi All. If it helps I had this same issue and it just turned out to be cumulative leakage. I had 3 x MCB for different groups of sockets in the house all on the same RCD. If I turned off 2 of them and plugged granny lead into the 3rd the car would charge fine but switching another back on with throw the RCD. Ended up replacing all of them with RCBOs which have individual RCDs inside them and the issue went away. The usual culprits for leakage are Solar PV interters, white goods and TV aerial boosters in the loft which get damp so try unplugging these.

This was a real issue for us for a while before we got Wall charger installed as we couldnt charge the car at all without having most of the house electric off!

 

[Glan gluaisne]

Just measured the earth leakage on my UMC when charging. Well within the limit for a portable appliance, between 3 mA and 3.1 mA, with no noticeable high frequency noise on the supply.

3 mA could be enough to trip an already marginal installation though. Although the nominal trip current of RCDs* is 30 mA, it's not uncommon to see them trip at around 25 mA or so on test, as the 30 mA figure is the maximum. I measured our house total earth leakage last week, and it was about 17.5 mA. It could easily have been just a bit higher, say 22 mA, in which case an additional 3 mA of leakage from the UMC might be enough to cause it to trip (if it wasn't for the fact that I installed an all-RCBO system, so every circuit has it's own separate leakage protection).


* 30mA is for RCDs fitted to a non-TT installation. A house with a TT installation would normally have a 100 mA RCD on the incoming supply.

 

[arg]

Just measured the earth leakage on my UMC when charging. Well within the limit for a portable appliance, between 3 mA and 3.1 mA, with no noticeable high frequency noise on the supply.

Did you also measure the UMC on its own (ie. without the car)?

 

[Glan gluaisne]

Did you also measure the UMC on its own (ie. without the car)?

Yes, nothing at all as far as I could see. Mind you, my clip-on leakage tester's lowest range is 200 mA, with a resolution to o.1 mA, but I suspect it's not really up to measuring, or even, perhaps, detecting, leakage that's below maybe 1 mA.

 

[MrBadger]

How do I test this? I have a clamp meter but need to check what things it can measure.

 

[Glan gluaisne]

How do I test this? I have a clamp meter but need to check what things it can measure.

Not many clamp meters will be able to measure this low; I have a dedicated leakage current tester that has a more sensitive lowest range.

Easy place to make the measurement, without interfering with anything, is on the earth tail in the meter box. That will give the whole house leakage current, so you can see an incremental change when something is switched on or off.

I have a very short extension lead that has all three conductors broken out in the cable, so I can measure anything that's plugged in through it.

 

[MrBadger]

My lowest AC current range is 2A with 0.001A resolution. It then goes on to specify accuracy, <1A(4%+40) not sure what the +40 means.

I'm not looking for exact readings, but a relative indication what circuits are highest would be useful. For instance, if turning off my inverter drops it significantly, then its potentially a relatively easy fix.

 

[Glan gluaisne]

My leakage current tester's lowest range is 200 mA, with a resolution of 0.1 mA, and it seems to reliably measure down to about 1 mA AFAICS. I doubt that you'll get a particular accurate measurement on the 2 A range, TBH, as the 4% accuracy figure alone could be as much as 80 mA of error. The plus 40 may refer to the error band in digits, perhaps, so +40 of the lowest digit?

The spec for mine on the 200 mA range is

• Resolution: 100µA
• Accuracy: +(5.0% of reading + 8 digits)

 

[MrBadger]

Well that was interesting and, for what I wanted, useful.

A big assumption, but lets just say that meter was reading in the right ball park.

Whole house earth block going back to the main incoming supply - around 19mA.
Earth block to gas main, iirc 6mA - not really sure what I am seeing here.
Earth block to dedicated cooker supply, 3mA.
Couldn't get earth block to main consumer unit .

Turning off/on a number of individual circuits showed a - swing typically 2-3mA. Some showed no measurable swing. Didn't try lighting circuits. Inverter was around 2-3mA.

Forgot to try the fridge/freezer - but last test I did it made no difference to UMC tripping whether it was plugged in or not.

So, what are the chances of what I am seeing being about right?
How important is the earth block to gas main reading?

That 19mA is whole house, plus whatever the effect of the gas main reading. Take off dedicated cooker circuit, then its 16mA.

I doubt the readings are correct as, even with an over sensitive RCD, 3mA for Tesla UMC is taking the consumer unit over its tripping point.

But I do know that, if I disconnect inverter (2-3mA) and turn off utility room/boiler cupboard (3-4mA), I should probably have enough headroom for it not to trip, even if readings were half what they should be. And if they were half what they should be, then the RCD would be even further from its tripping point.

So my conclusion is, if the UMC trips when utility and inverter circuits are turned off, then UMC is not working as it should.

Obviously a schoolboy approach, no need for any lectures...

 

[Glan gluaisne]

Those numbers seem to be in the right ball park to me, pretty much what I'd expect, and surprisingly good for the current range on your clamp meter.

I've seen 30 mA RCDs trip at ~20 mA before now, so it may be that you have a particularly sensitive one. If so, then swapping it out for a different type may help.

The 19 mA figure is the whole house leakage. The current in the bond to the gas pipe is probably not significant (although it'd be interesting to find out where it's coming from, not that it helps your problem).

How long does it take the RCD to trip once the car starts charging through the UMC? Any chance you could get a quick leakage current reading on the whole house before it trips? Might show if there's a fairly hefty spike in earth leakage, or whether there's something else making the RCD trip (they can trip on transients sometimes).

 

[arg]

The current in the bond to the gas pipe is probably not significant (although it'd be interesting to find out where it's coming from, not that it helps your problem).

It's presumably diverted neutral current - the neutral current will split between the neutral conductor in the supply cable and the alternative path (gas pipe->mass of earth->earth electrode(s) at transformer plus extra PME electrodes) in the ratio of their resistances.

So if the supply cable resistance was 0.05 ohm and your house happened to be drawing 10A in total at the time you measured the 6mA diverted current, that would suggest that the gas pipe viewed as an earth electrode had an impedance of 0.05*10/0.006 = 83 ohm - which is not unreasonable (especially as all these numbers were plucked out of the air).

 

[MrBadger]

How long does it take the RCD to trip once the car starts charging through the UMC? Any chance you could get a quick leakage current reading on the whole house before it trips? Might show if there's a fairly hefty spike in earth leakage, or whether there's something else making the RCD trip (they can trip on transients sometimes).

Now you have your car this description will make a bit more sense...

charge port goes blue, then just when you expect it to go green, it trips. So I guess blue is the hand shake and trip when it starts to pull current.

 

[Glan gluaisne]

It's presumably diverted neutral current - the neutral current will split between the neutral conductor in the supply cable and the alternative path (gas pipe->mass of earth->earth electrode(s) at transformer plus extra PME electrodes) in the ratio of their resistances.

So if the supply cable resistance was 0.05 ohm and your house happened to be drawing 10A in total at the time you measured the 6mA diverted current, that would suggest that the gas pipe viewed as an earth electrode had an impedance of 0.05*10/0.006 = 83 ohm - which is not unreasonable (especially as all these numbers were plucked out of the air).

That makes sense, I wasn't thinking in terms of the gas supply pipe being a reasonable earth, as the pipe supplying our old house seemed to have a plastic sheath over it, where it came up the wall from the underground run.

 

[Glan gluaisne]

Now you have your car this description will make a bit more sense...

charge port goes blue, then just when you expect it to go green, it trips. So I guess blue is the hand shake and trip when it starts to pull current.

So it's tripping as soon as the car onboard charger turns on. Probably going to be hard to see any spike in the leakage current, I suspect, as digital meters sample and may well not show a short duration increase (something my old Avo 8 was better at doing).

 

[MrBadger]

ill try and see if peak hold works

 

[MrBadger]

I'll put this one to sleep for me now. Thanks to @NorfolkMustard we were able to try a second car and UMC a known good combination. As expected, any last glimmer of hope being a dodgy UMC evaporated when the RCD tripped as with ours.

I did manage to get a charge going, as predicted, by turning off inverter and utility room - inverter alone was not sufficient. Once charge started, turning these circuits back on allowed car to continue charge.

I did take a leakage reading at 22mA. I don't think I specifically caught the peak.

So unfortunately, a little more to be done than just moving the inverter on to its own dedicated RCD - was never happy that it shared an RCD. Also involves swapping out the main RCD, and hoping its not so trip happy.

Question, will an RCBO fit in the space of an MCB in an MK consumer unit? They look taller to me.

Or finding a bloody electrician willing to install a dedicated EV charging consumer unit (with type B RCD) and some 13/16A circuits. Maybe I will have more luck if I keep that a separate job from installing the TWC.