4 minutes ago, anotherbrownbear said:

Quite expensive for what you bought.

Jip, it is ... but that bullet is through the church ... cannot return it as I opened the box ... 

6 minutes ago, The Terrible Triplett said:

Jip, it is ... but that bullet is through the church ... cannot return it as I opened the box ... 

Where did you buy it?

4 minutes ago, anotherbrownbear said:

Where did you buy it?

From LiveCopper. 

ABB does not want to take it back, after we debated for blue murder colours that they gave us/me the wrong one.

Lady we spoke to kept on saying her manager says this, says that ... she and her manager had no clue ... thanks to Plonksters kind assistance to prove them wrong.

I lost this round.

There, tripping of DB that is on AC_out1, sorted. 

@plonkster it was the blerrie stupid red plug all this time, the one I wanted to keep to show the DB is on. 

The very expensive ABB breaker can stay, don't need the next level as you have.

1 minute ago, The Terrible Triplett said:

 it was the blerrie stupid red plug all this time, the one I wanted to keep to show the DB is on. 

Chop the plug and put on a new one. 'tis what I did to the one I had. I am still amazed at how a MOV (which I thought should pass just about no current below its breakdown point) can pass enough current to trip an RCD. Of course it is of the transient kind, so it is an absolutely microscopic spike on a chart. I should research that more.

Well here you go, from Wikipedia. A MOV has a capacitance of between 100 and 2500pF. Towards the higher end you're in the range of a typical capacitive dropper (for reference), which passes enough energy to switch small relays (this is how almost every daylight switch powers itself). So that explains why it passes a small transient spike when you mess around with the neutral bonding (which is what it does during the relay test): The darn thing is a capacitor!

Edited January 15, 20197 yr by plonkster

Must be dud though, 'cause at those values its not going to pass 30mA @ 230Vac.

Maybe its responding to a much higher frequency harmonic?

2 hours ago, phil.g00 said:

Must be dud though, 'cause at those values its not going to pass 30mA @ 230Vac.

The earth leakage clamp was showing peak values around 15mA to 20mA, which is enough to trip many 30mA RCDs. The meter's sampling rate isn't particularly fast. But after replacing the RCD the problem went away, which suggests that this peak is incredibly short lived.

I think (but didn't prove yet) that because I have devices/appliances with such surge protection devices between L/N and earth (little triangle connection), it forms a natural voltage divider and without a N/E bond the earth tends to float around 110V. So as the Multi opens and closes its bonding relay to test things, it creates a square wave modulation (0V to 110V) of sorts on top of the AC. The relay test sounds like around 10Hz if you listen to it, so you have a 10Hz modulation on top of a 50Hz AC signal... with things on there that passes AC. I don't have the right equipment to test it unfortunately, specifically... no storage scope.

6 hours ago, The Terrible Triplett said:

There, tripping of DB that is on AC_out1, sorted. 

$(%*#&!!!! And it is back. Tripping again.

Will have to start again, line for line, to find what it the new cause is. 

Everything was perfect, so tonight I though, hey, lets show the wife ... should not have done that. She jinxes everything! (facepalm)

@plonkster - what can I ask my electrician to test when they come back again?

1 hour ago, The Terrible Triplett said:

 - what can I ask my electrician to test when they come back again?

First thing to test is that the RCD itself works properly and at what residual current it trips. The tester for that is cheap, costs around R200 at the local electrical place. Though I doubt it, just maybe the darn thing trips at 10mA or something ridiculous. But if experience is anything to go by, it will trip between 15mA and 20mA.

Second thing is to get an idea of the peak imbalance, the item that is tripping the inverter. For this you put a clamp meter around the live and neutral wires where they enter the RCD, put the meter on peak-hold (it's kinda useless without that function), and then you let the inverter trip the RCD and note the current. Of course it is better if you have an actual earth leakage clamp meter, those are designed for this sort of thing. If your system is anything like mine, you're going to get a 20mA reading or thereabouts. You can then remove the RCD and put in a normal isolator (temporary, for testing) and repeat the test. Mine spiked to 40mA.

After that I started doing divide and conquer, turning off breakers and unplugging appliances to find out how each circuit contributes to this total. It is time consuming and a PITA. Eventually noticed that almost the entire peak is contributed by the garage door opener and the gate motor. For you that is going to be different.

Once you know what appliances they are, you can attempt to split the DB and use more than one RCD for small groups of circuits. In my case that didn't help, the two mentioned appliances on their own was enough to trip the RCD.

So I took the plunge and ordered the breaker. It was a massive and expensive chance, that much is for sure.

But for starters, getting an idea of the magnitude of the leakage is a good starting point.

9 hours ago, phil.g00 said:

Must be dud though, 'cause at those values its not going to pass 30mA @ 230Vac.

Yep, I thought so, that may have been contributing a wee bit but it's not the big baddy boom.

And now you've got a half decent earth on the place ( I thought I saw 2ohm on a photo somewhere?), problems are showing up that you have been blissfully ignorant of.

Do you see inside your original orange box, (also on a photo somewhere) there has been arcing to earth from what I assume was the neutral?

That's a sign that your neutral voltage has been high and it means that it has been looking to get to earth wherever it could.

So the insulation has been getting stressed all over the kip.

I expect you have a neutral-earth fault, this is supported because you haven't mentioned any coincidental MCB trips, ( and I think you would've)

This could get tricky to find by the "divide and conquering the MCBs" technique.

Why? because of red herrings:

"Borrowed neutrals"  say you narrowed it down to 2 MCB's. Either one trips still trips the common E/L. Hell they even trip their own individual dedicated E/L's. 

So it may very well be it can only be remedied as @plonkster says, but first exhaust "Occum's razor".

It could be that the live heads out from different MCB's,  but the far side of the appliance, the neutral has just been tapped in to the closest available neutral. In other words, only one wire back. Let's say there's an earth fault on that neutral.

Hence both appliances will appear to trip the E/L and even separate E/L's when a dedicated one is put on each circuit.

Another scenario lets say wiring is running together along the same route, maybe the assymetrical enough that the magnetic effect of a loaded circuit is sufficient enough to induce a 60mV in the neutral wire of an adjacent circuit.  ( your new earth spike is now at 2ohms, yeah) Well, 60mV/2ohm's = 30mA and your E/L will trip, but it will appear to trip when you turn on a healthy circuit.

Another, scenario you have multiples or combinations of the above. It has the potential to be head-wrecking based on incorrect assumptions.

On the other hand, it could be as easy to find as playing a tune on the MCBs. Try it, but before you spend big money do this:

Next time your E/L trips, keep everything just how it is.

Note this down, (for future Cluedo, if necessary).

Trip all the MCB's in your DB, ensure (test) everything is dead.

Disconnect all the neutrals from the earth bar. One by one separately megger (500v setting) the neutrals down to earth, returning clear neutrals one at a time to the common earth bar - don't stop when you find a neutral down, there maybe more than one.

A multimeter is not good enough to do this job, everything can seem clear until you overcome the initial insulation breakdown.

Might save you a bob or two.

Edited January 15, 20197 yr by phil.g00

I am making some assumptions here, be aware of that,

One is you have a single E/L and all neutrals are through it.

Two, the whole solar side is still disconnected awaiting the DB approval.

6 hours ago, phil.g00 said:

Disconnect all the neutrals from the earth bar

Just to be a pain here but neutrals go to a common neutral bar not the earth bar (  I know what you meant but this might confuse other members)

After the earth leakage neutrals and earths never touch, or obviously the earth leakage device will trip

.

  @Jaws   , of course yes.

I think I wrongly named a "neutral" somewhere else in the post as well.

My gist to other members is, there is an order of probability to employ when electrical fault-finding.

Start with the assumption that the earth leakage is doing exactly what its supposed to do.

Namely, trip for an earth fault.

Rule that out first, and rule it out properly.

Edited January 16, 20197 yr by phil.g00

7 hours ago, phil.g00 said:

"Borrowed neutrals"  say you narrowed it down to 2 MCB's. Either one trips still trips the common E/L. Hell they even trip their own individual dedicated E/L's. 

 

7 hours ago, phil.g00 said:

On the other hand, it could be as easy to find as playing a tune on the MCBs. Try it, but before you spend big money do this:

Right. I did something similar during the test though for me it had no effect.. but Phil is right. You start with just disconnecting the live side (ie opening breakers), and if you make no headway you also disconnect the neutral wire (on the neutral bar) to the same circuit. This is where the earth leakage clamp meter helps you: It gives you a yay-or-nay indication as to whether the one you just disconnected is involved in the situation, especially where you have multiple possible causes. The RCD is pretty binary (either you are over 20mA or not), but the clamp meter can tell you "you're still over... but that was better!".

You can even (this is potentially dangerous, only do it as a test!) disconnect the neutral and leave the live connected. If the circuit doesn't go dead, then it is getting a neutral from somewhere else (ground fault, or shared neutral). Be very careful when you do this. The disconnected neutral wire (which you conveniently pulled to the side in the DB) becomes live when you do this (because power goes via any loads on that circuit and raises the potential on the little loose black wire).

In my case I eventually found that the disconnecting the breaker was enough. The leakage spike immediately dropped. Removing the neutral had no significant additional effect.

4 minutes ago, plonkster said:

the earth leakage clamp meter helps you

I'd recommend the Megger, it's an intermitent fault,  maybe pecking, at 230V, -- 500V will make the difference between a maybe and clear answer.

This morning the APC UPS was off the other inline UPS that powers the internet connection, front door camera, on.

So I switched the Multigrid off at the DB, see 1, simulating a Eskom failure.
That breaker disconnects live and neutral, correct?

Switched it all back on, no problem. The ABB expensive breaker in the separate DB works perfectly.

Then with the APC on, the ABB breaker trips. Yesterday it did not trip.

Here is the Ac-out1 DB, that powers lights, computers, TV's, router / fibre etc with the ABB breaker that keeps on tripping when the Multigrid syncs with Eskom.

Edited January 16, 20197 yr by Guest

I wonder if the surge-protection circuit in the APC UPS may be causing/contributing to this problem?  I would try connecting the UPS loads directly to AC-OUT1 and see if this changes anything. Then repeat this test with APC UPS completely disconnected from AC-OUT1.

35 minutes ago, NigelL said:

I wonder if the surge-protection circuit in the APC UPS may be causing/contributing to this problem? 

Mmmm, that could apply to all UPS'es re. the surge protection?

Re. the APC, let me get the guts together for that. Behind the APC is my server ... let me see what I can do.

@The Terrible Triplett, 

I feel your pain, but really this thread could go on for months if you start guessing like this.

You'll get round to a methodical approach like I suggested sooner or later.

Or you may get very, very lucky first.

And I genuinely wish you luck.

@The Terrible Triplett See the point made at circa 7:25,  and his comment about what is on at the time is an unreliable clue as to where the fault is.

 

47 minutes ago, phil.g00 said:

You'll get round to a methodical approach like I suggested sooner or later.

Yes, I agree wholeheartedly. When the electricians come back I will give them what you and Plonkster suggest to test, copy/paste, not my understanding of it. :-) 

This entire sage, the Multiplus, is driving computer related stuff and house lights. It is a very "contained" scenario.

A device or combination of devices, are causing the issue. That is my suspicion as I pulled all the wires from that separate DB board and I connected all the plug boxes at the ends myself. Testing to confirm that I did it all right. But if you think about it, their is live/neutral and earth, all plugs.

Further to the wires. I can put down extension leads to bypass the existing wires. It is a mission though!

The part I hate the most is when the PC's / devices are rebooted. When all is off, just the lights are on, there is no tripping. The lights are not the cause either, as they used to be on a plug with a changeover box. So unplugged, it still happened.

I have a very sad feeling, as per Plonksters experiences, that it can be an ongoing issue. Can find the device/s, but as one adds a new device later, it can always come back one day.

Why it is just him and me, that I don't know.

@The Terrible Triplett, I would advise you not to use extension leads for safety reasons, if they are just two wire leads.

I remember visiting an old lady whose washing machine was plugged into an extension lead in the kitchen from the far wall. I inquired politely as to why she did not use the adjacent plug to the machine. I was told when she used that wall socket ( the one she didn't need the ext. lead to reach), it tripped the earth leakage.

So she never used that wall socket for anything anymore because it was obviously faulty.

Unknowingly she had hidden the problem, not solved it.

That why I am a bit wary of this desensitized E/L solution without exhausting all other possibilities.

What she now had was the same fault but she had made sure that the only earth return was through her and a wet floor.

So watch yourself the overloads are to protect equipment, the E/L is to protect you.

 

Edited January 16, 20197 yr by phil.g00

5 minutes ago, phil.g00 said:

... if they are just two wire leads.

Extensions leads are always always 3 wires and the thicker ones and I tend to cut off the cheap ends and replace them with good ones.

If a wall socket is dicey, it gets fixed.

 

Right, I tried bout 8 times. Sometimes the breakers stays up, sometimes it does not - no pattern whatsover that I can notice.

This is the part where I start thinking and that always ends up to be very very bad for my pocket. Something is going to blow.

Electrician need to take over and do the test as suggested by you @phil.g00 and you @plonkster

4 minutes ago, The Terrible Triplett said:

Sometimes the breakers stays up, sometimes it does not

Are you saying more than the E/L trips? That's relevant information.