When i have a grid power failure the RCD and MCB trip off in our V1 gateway covering the AIO IF we are generating solar.
any body else had this and know what the fix is ?
the Solar RCD and mcb stay On...
After Dark if there is a grid outage all is fine.
Any thoughts Regards Rob c
rob c I had something similar, where when the grid dropped the system would trip. In my case it turned out that there was a loose connection in the breaker for the AIO within the gateway. Once the GE engineer had tightened the connections the problem never re-occurred.
One thing I do get since, is if the battery was full and there was loads of solar (so exporting) the system would "shut down" for 20 seconds or so whilst the inverter was powered down and then all came back - so not the 20ms cut-over in this case, but of not a full battery, or at night, or little solar then the cut-over is pretty instant
thanks CW I've power it all down check the connections..
regards r
rob c If all the connections are tight when you checked, there was one other possibility that was being considered for my setup and that was that there was more than 30ma of earth leakage, so there was talk around replacing the breaker in the Gateway with a 100ma one - you would need an electrician to look into that
Post a picture here of the RCD that trips. In addition to it's leakage rating, it may be that it's also the wrong type of RCD and is sensitive to dc leakage coming from your inverter.
We are also getting this, with a v1 gateway, A013 firmware.
We know the 30A RCD is too small for our twin string inverters so we are getting a larger one very soon - but this is really to fix a different issue. Gateway was fitted with the wrong busbar - which was sorted right before GE went bust - might be worth posting a pic of the inside of your gateway here (unscrew top cover). Our installer is really unfamiliar with gateways and does not read instructions so it is not easy to know where the problem is. Our Earth rod hasn't been tested but apparently that wouldn't cause this issue.
CW My installer had to replace mine with 100ma to resolve tripping when switching to island mode. The theory being that my CU has 2 30ma RCD's feeding back to the singe 30ma in the gateway meant a potential for overload.
Previously there was a problem where the CU RCD's would trip but that was before the Earth Bond Relay fix came about.
There are still some situations, as yet unidentified, that cause a trip or a 20s reset on switchover to/from island mode though.
Neither high solar or over 80% SOC are the unidentified issue. Its such a hard thing to reproduce or diagnose I am with the live with it stage and have put my UPS's back on the vital kit in the house!
We are having the same issue -
what is the CU? And what did the earth bond fix? We are in the same situation although installer planning to fit 50mA RCD, he said the cabling he put in wouldn't be sufficient for higher. We have frequent power outages here and whether it is light or dark out what happens is we lose power then it starts a restart cycle and power comes back but cuts out within about 2-4s (part way the through the doorbell's chimes), after a few seconds the restart cycle begins again etc... does that sound like your issue? I don't know if we have DC leakage... inverters are Solis and we don't have a monitoring dongle but they don't come up with an error.
AnonymousHooman CU => Consumer Unit.
If you have an AIO/Gateway combo and the grid fails, there is a changeover relay in the Gateway that will firstly disconnect your house Live and Neutral supply from the grid, and then, very shortly afterwards connect them to the Live and Neutral coming from your AIO inverter, assuming that your battery is charged up enough to power the inverter. As the reason for the grid failing may be due to someone/something completely chopping through your supply cable, your incoming Earth cannot then be trusted either, so a local earth rod is fitted when the AIO/Gateway is installed. Very shortly after the Gateway switches your supply over to the AIO inverter (ie Island Mode), your house Neutral is connected to the installed Earth rod via a N-PE bond relay to clamp L and N voltages so that insulation max voltages are not exceeded, and so that protective devices (ie CU RCDs) continue to work. The grid supply Earth is not disconnected as it provides a useful parallel path if it exists (ie the supply cable is still intact). All well and good. When the grid supply is restored, the N-PE bond relay opens, and the L/N changeover relay operates to reconnect your house to the grid supply. However, if the N-PE bond relay is slow to operate such that your local Earth rod is still connected when the changeover relay reconnects you to the grid, the AIO RCD device in the Gateway will see this as significant earth leakage and will trip, cutting power to the AIO and your house and because of this my Gateway was replaced not too long after it was installed.
I'm writing this from memory, so I might have got some of the details wrong, but you can still get the gist of what I am saying. Be aware that the supply earthing arrangement I have described may well be different for people living in the countryside with overhead supplies rather than urban dwellers like me.
AnonymousHooman Before the relay change our CU RCD's used to trip on changeover. This was mostly fixed a couple of years ago, soon after install, when the whole N-PE relay issue blew up for GE AIO. I believe since then GE ensured the relay was set up at commissioning.
After that the Gateway RCD became the problem there was a lot of investigation done including swapping out the gateway but it seemed that the combined earth leakage across my 2 CU RCD's was over 30ma but not sufficient to trip the CU RCD's 30ma each. My installed said things like USB chargers and the like can contribute to the leakage and we have them all over the house. We also have 2 large Solis inverters.
GE then recommended the 100ma RCD in the gateway. So as to match the same manufacturer as the original, my installer had to source this from abroad as they were like rocking horse poo at the time but preferred to do this rather than use one from another manufacturer.
The upshot of all this was that when testing, by dropping the grid from the gateway switch, things worked fine as long as solar was below the levels the gateway could distribute and SOC was below 80%.
However, on very odd occasions we have had the odd blip where either gateway or one of the CU RCD will trip, or the system will so the 20s reset with no trips. This may or may not be related to short blips on the grid but varying degrees of SOC and solar are involved.
In the end we just accept it as its impossible to fault find without debug level monitoring of everything involved in what is a pretty complex scenario. T-M explained the N-PE relay much better than I could have!
I dont believe we have ever had the restart loop you describe but if it is repeatable and you have a split CU you could try taking each side of the CU out the equation to narrow down. Does it actually trip the gateway or CU RCD anytime during the looping?
Hopeful things worked fine as long as solar was below the levels the gateway could distribute and SOC was below 80%.
I understand the need to the solar output to be able to go somewhere in the event of island mode, but why the 80% SOC level? Is this just leaving a big enough "gap" for solar power to go into until the solar output is ramped down, or is there another reason?
CW The SoC can easily be out by 20% so the cells could in fact be nearly full while the SoC shows 80%.
Its (apparently) not quite as bad as the LV batteries & hybrid inverters but it is entirely possible for an error of that magnitude to be present if the battery hasn't been regularly charged to max/discharged to min.
Vestas OK, that makes sense thanks. Having never had a DC coupled battery, do you get the same behaviour in similar circumstances (loads of solar, full battery etc) or is the system able to curtail solar output quick enough to prevent problems?
Hopeful Because of the slow N-PE bond relay problem, I believe it was common practice at the time for installers to disable it.
This meant that when in Island mode (ie during grid failure) the L/N supply from the AIO inverter was "floating" with respect to Earth, typically giving 120vac L-E and 120vac N-E, though as they were floating, there was theoretically nothing to prevent them floating up to voltages that would exceed the max voltage rating of the cable insulation, possibly leading to insulation breakdown.
Also, if you plugged a socket tester in and pressed the test button, nothing would trip. This is how my AIO/Gateway was originally installed, and contravened the IET Codes Of Practice. That was why I complained about it and got it fixed by having a new Gateway installed.
People buying AIOs/Gateways on Ebay and elsewhere really need to be made aware of this when purchasing. The Serial Number of my original Gateway was a GW2330xxxx device (Week 30 of 2023), so anything before and up to this is likely to be suspect, but I'm not sure where the cutoff between Gen1 and Gen2 Gateways lies.
T-M splitting hairs but it wasn’t a slow relay, it simply wasn’t there in the gateway at all as they’d forgotten to fit it.
They fudged it eventually using a N-E relay in the AIO itself whilst they made the later gateways compliant, however the ‘fudged’ ones then start tripping the RCDs in the gateway so you end up fitting 100mA ones to stop it.
Here my RDC's twin 30mA
so the general consciences is to change the 30ma AIO RDC to 100mA is that what we are saying ?
If the above is correct, i believe the part number is: GSH202 A S-C32/0.1 any one know I can buy one ?
Regards rob c
hoggy Haha. So it's even worse than I thought. Have to ask whether it was the fault of someone in the UK not knowing what the relevant CoP was, whether the introduction/update of the CoP caught the UK on the hop, or whether the fault lay at the China end?
So glad I forced GE to fix it. Not only did I get a new Gateway, but I also had to have a main board changed in the AIO when an update failed and bricked it. All happened within weeks of having the damn thing installed too. Felt like jumping off a bridge at the time!
rob c Maybe a 100mA Type AC RCD would be better (anyone else wish to confirm?), which would be GSH202 AC S-C32/0.1. IF you can find one. Try contacting ABB directly maybe.
Type AC RCDs are not "blinded" by pulsating or smooth DC currents, which I understand some/most inverters put out along with the 240vac.