EmonEVSE tripping RCD

Got a slightly strange issue here, currently trying to make sense of it but can’t find anything.

Installed the charger just over a week ago with a Matt:e O-PEN projection device upstream. Since then I’ve been charging exclusively off solar divert with great success and no issues at all

The last two nights I’ve been attempting to charge at full beans with off peak power, starting circa 0030 (±600 sec as per latest regs). Each night at roughly 0105-0115 the Type B RCD in the Matt:e has tripped, and I can’t figure out what would possible be causing that. The EVSE reports a no ground fault (possibly because the Matt:e isolates earth at the same time if the RCD trips first and some residual power logs the failure before turns off, but this is a guess)

The full wiring is as follows:
Matt:e (C32 MCB > Type B RCD > O-PEN detection contactor)
Hyundai Kona

If anybody has any ideas on what could be causing this it would be greatly appreciated.

I just ran a supervised test to see exactly when it was happening, started at 1155 and heard the contractors trip off at 1228, meaning a runtime of 33 minutes. Again the EVSE reported “No Ground” before it lost all power.

Curiously, I went straight out to the garage to trip it back on and almost as soon as the O-PEN contractor energised the RCD tripped again, so whatever the fault was hadn’t cleared. Left it a minute or two, tripped it back on and everything was working as expected, charging started at the full 32A (until I put the solar divert back on).

The timing on this seems to be too consistent for it to be a random fault causing it, but the fact it doesn’t occur when trickle charging is even more confusing to me as if it was a self test I’d expect to see it at other times.

I sounds, from what you’ve written and depending on where you are in the UK, to be neither a genuine earth fault nor ambient temperature related, but it does seem to depend on charging current, so temperature as in I²R, where something is warming up and that’s generating the trip. It is

which makes me think ‘thermal’.

It could still be a genuine high resistance earth causing the problem - for example if you have enough earth fault current to dry out the soil around your water pipe or earth rod, and it takes half an hour to do that, but not enough to do it when you’re charging at the lower rate. It’s quite a stretch to believe this is happening, but it’s not to be discounted until proven otherwise.

I think it’s much more likely to be something more mundane, I’d double-check and make sure all terminals and connections are good and the installation instructions haven’t been mis-interpreted.

1 Like

All wild possibilities, I’ll open everything up later and double check all terminals. I don’t see how the earth resistance would cause the RCD to trip though, unless that’s how the EVSE deals with a no earth fault condition, or if it gradually ramps up leakage to measure earth resistance to such a point it trips?

Do you know how your

(did you really mean ‘projection’?) actually works? From what I can tell, it might be assuming that a voltage dip L–N means the neutral has fallen off, or it might be measuring a voltage somewhere else. The website seems very good at extolling the need for their device and hence its virtues, it’s rather less clear on just how it achieves it all.

One thing made clear on their website is the Matt:e O-PEN does NOT do residual current protection.
I have a plug-in RCD that will trip when I plug an extension lead into it. The lead’s insulation is fine. I put it down to stray capacitance.

And I’ve been in the profession for long enough to have encountered enough wild possibilities never to discount one. I’ve also learned that neglecting the simple explanation is also very risky.

Sorry, yeah that is a typo, it is a meant to be protection, specifically Matt:e SP-EVCP-B as suggested in the install instructions, which does included a Type B RCCB.

Unfortunately I can’t think of any way to assert the O-PEN is functioning 100% as expected, but the voltage drop measured under full load (at night when no solar from neighbours either as that minimises the drop in the day) is well within spec (from ~245 to ~230). The O-PEN also wouldn’t explain the RCD tripping as that just controls the contactor and would reconnect once it thought the fault has cleared.

I’ve just checked all the terminations and everything is correct according to installation instructions, but now I’m thinking I might have to come up with a full test plan to measure leakage whilst under full load to examine if there is enough to cause that. I guess there could be a total leakage from the EV, EVSE and O-PEN electronics that is just pushing it over the RCD threshold or the RCD could be faulty and oversensitive.
I was primarily interested in asking if there were any known scenarios that could cause this to investigate first before I break the meters out everywhere and so I know where to start looking.

It would be worth checking the MCB and RCD terminals also to see if there’s a loose live or neutral. I had a similar issue with my Zappi a few weeks ago, the RCBO was tripping at 7.2kW charging at night time after an hour or so of charging but not while slower charging. In my case I had a loose live in the isolator which was getting hot and then tripping.

I’m guessing you have very warm weather over in the UK so this is a less likely possibility but is there moisture inside the charger?

I think you will need to get hold of an RCD tester to test the RCD is functioning correctly i.e tripping at the correct leakage current and also an earth leakage meter to measure how much current is leaking to earth.

Are you able to charge from other AC charging points at full power without them tripping?

Thanks all for the advice, I managed to bribe my sparky friend out on Saturday evening with the promise of a hearty dinner, and we spent a good few hours troubleshooting this, here’s the observations we found.

  1. RCD checked out all fine, 24mA on ramp test, good values in 1x 5x 0.5x.
  2. We’d left a service loop of cable at each end for future use, however the earth had been pushed up next to the O-PEN module which had quite high RF output from its relay and was inducting onto that.
  3. When removed from the enclosure, leakage sat at 11mA and after 40 minutes at 32A nothing had tripped.
  4. Thinking we had solved the issue, we shortened earth to be the minimum practical length, closed everything up, and I started another test thinking I’d just let it run for an hour to confirm we’d fixed it.
  5. After about 45 mins I went on to check, but it was offline, MQTT revealed it had died around the 40 minute mark, RCD had tripped again.
  6. Now confused we opened it all back up and noted the RCD was very warm, almost uncomfortably so, and another RCD test revealed it had dropped sensitivity to 21mA, not terrible, but a suspicious change. As it cooled the sensitivity returned to normal.
  7. Final test time, 32A, timer set for 25 minutes so we’d definitely be there. Let it run and after the 25 min mark we returned to observe it and measure the temp every 5 minutes. We also checked cable temperatures outside the enclosure which were barely above ambient.
  8. With lifting the flap it must have cooled slightly because it reached 45 mins before it tripped. Instantly we took the cover off, measured the RCD temperature at a toasty 47ºC and as quick as was safe ran an RCD test.
  9. Finally a result, it now measured 19.5mA, but more surprisingly/concerningly it did not trip at all up to 33mA on a 180º test. Sticking it on a 100mA test it reached 45mA before it tripped (no exact value as the tester just flicked to fail <50mA due to it expecting a 100mA RCD). Given this took a minute or two to run and it would have cooled, I highly expect it may have been even more sensitive when it tripped. I think it’s fairly safe to say the RCD is no good, and I will be getting onto the supplier today to get it replaced.

Thanks again for everyone’s help, although if you have any comments on our methodology/observations they would still be appreciated. I’ll update further once the new RCD is in to confirm if that fixed it.

That’s some good diagnosis. Out of interest what brand and model is the RCD?

It’s an IMO B10R2040-30-B, as supplied with the SP-EVCP-B. To their credit Matt:e said the temperature change is expected, but the readings aren’t, so they are sending a replacement out for me to fit and see if it resolves the issue.

1 Like

I am not a big fan of using one wire for earth and neutral combined. It has safety issues.

However, if your RCD trips, do I understand correctly that the EVSE then also looses ground? I believe in that situation a local earth rod at the EVSE would be required in Belgium. Of course PEN is as far as I know not used in Belgium, we think it is unsafe. If the RCD opens here, the PE is still connected.

A. Do you have PE locally at the EVSE? Have you tried if the RCD also trips with a separate PE wired to the EVSE?

B. On #6 in your test, if the RCD is passive, it has a lot of fine copper winding and resistance will increase with temperature, that may explain the drop in sensitivity with temperature.

C. For electrical installations, 47ºC is not “toasty”. The OpenEVSE here reports temperatures over 60ºC in summer. That is limit for consumer grade electronics… an EVSE should be built with automotive grade components. Not sure if OpenEVSE is… electrical installations can withstand 47ºC easily, but not recommended of course.

D. In general, “RF noise” will be picked up a lot more than 50Hz by the RCD windings, it should filter these out, otherwise it may also be on the suspects list for the weird behavior.

I agree with you that in certain circumstances, a combined PEN conductor might not be the best. In my city, there was a fire in the offices of the electricity company, and a significant proportion of the records giving details of earthing arrangements across the whole city were destroyed. The only practical solution was to treat the whole city as if it were a TN-C-S system, so invoking more stringent safety rules. The result locally is modern installations will be true TN-C-S, but old ones (e.g. mine) will actually be TT, with the protective earth being provided by the buried lead water supply pipe.

I can understand why the RCD supplying an EV standing outside the house should disconnect the earth as well as all the other conductors when it suspects that the PE conductor has become detached from the true earth - it will reduce the possibility of the vehicle being at an elevated potential with respect to the local earth on which it stands, therefore reducing the shock risk to an individual who might touch the vehicle whilst standing on wet ground. But in every other circumstance that I can think of, the vehicle should remain earthed.

No, the RCD does not disconnect the CPC conductor. However, if the O-PEN fault detection device e.g Matt:e detects a possible PEN fault by sensing over or under voltage then all conductors including CPC are disconnected.

Ah! Thanks for the clarification !!

Yes and there is another scenario I encountered.

I had lost PE on the outside outlet I was using for the EVSE. Because my electrician messed up in a splice box. Nothing alerts you that this is the case. So you are in danger. With PEN you know because you get no power from that outlet.

The situation was discovered by the OpenEVSE. It disconnected and said “No Ground”. Deep bow to the guys from OpenEVSE and OpenEnergyMonitor, you may have saved my life!

1 Like

He didn’t do his job properly, and should be told. And if you feel the lapse was sufficiently serious, the authorities need to be made aware.

My neighbour had an electric shower fitted by a plumber who thought he knew enough about electricity. He illegally broke the seal on the supply fuse (obviously to pull it out so that he did not have to work on live connections), cut the cable between the meter and distribution board, inserted a pair of connector blocks so that he could have a second smaller distribution board for the shower, and in connecting it up, swapped line and neutral. So in the event of a fault to earth, ALL the wiring in the house was only protected by the 80 A supply fuse - none of the individual circuit breakers/fuses would have seen the fault, because they were in the neutral conductor. And bear in mind, the cable we use for lighting circuits is rated at 6 A.
I put it right immediately I found it.

In Belgium, changes like the one at your neighbour’s requires certification by a reviewing authority. Like when I added solar panels last year and swapped out my inverter, that required a review with certification. The installer cannot submit an invoice until the certification he has to request has been completed.

The slip up in my home by my electrician has been fed back. Certification was not required since it did not include changes in any panel, it was on one fused circuit only. Something I could have done myself.

I’m no longer employed now, retirement (kind of), so now I have time and do everything myself, including the panels. Being a masters in electronics engineering I am allowed to but of course my work also must pass certification. The work for the solar change I did myself and passed with flying colors :-).

I am adding a small panel on a fused circuit behind the existing fuse in one of the main panels. That does not need to be certified as long as I keep to the max number of outlets/branches (eight). This is so I can add a RCD type B to a branch that feeds OpenEVSE. The RCD in the main panel is type A. And I did not swap it for a Type B, which I could have. Type B for things like bathroom, washing machine and so on is unusual and puts off the reviewers, so I decided to keep with what they are comfortable with and add the Type B in a branch, they did not mind.

Here, the installer must be registered with a trade organisation, and then they can certify the installation themselves; but that didn’t stop this fellow, he most likely wasn’t registered by any competent body. I’m professionally (not trade) registered as a Chartered Engineer. We write the rules the trade must obey. :smile:

Why is that? I thought a type-B RCD will protect you from everything a similarly rated type-A RCD would, but with an additional protection against DC leakage.