SMA 400TL tripping RCD

[keefy]

Just switched on and inverter goes through the motions. Just at the generating stage and it trips out the RCD.

Replaced RCD, checked ALL connections 10 times each, run as 2 strings as many times before, all connections are correct.

Any ideas?

I don't really want to take this out now and send back to SMA, could do without the hassle.

Only other thought is the wiring in the house is litter, possibly an issue with too high earth fault already existing and now this is tipping it over the edge. Marginally proven by 1/2 30mA RCD test tripping the 30 mA RCD out. Also don't fancy tracing trillions of niggly faults right now... too much to do.

Any suggestions would be helpful, apart from replacing RCD with isolator!

 

[bumcrack]

Don't SMS spec minimum 100ma?

 

[SolarCity]

It is not a good idea to have an inverter on an RCD for a couple of reasons. First of all you have a high possibility that it will trip out and secondly you will fail to get adequate disconnection times on any circuit relying on the RCD itself. If you can, put it on it's own seperate consumer unit and ensure that the wiring of the circuit does not require RCD protection.

 

[moggy1968]

4000TL does require a 100ma RCD, and it needs to be a type B (which cost a fortune)
It will require an RCD as you don't have simple seperation of AC and DC circuits, as it's a transformerless inverter (according to the big green book).
Sibert can do you a seperate garage board with the appropriate RCD.
they have an excellent PDF on this issue, and others on specifically SMA inverters and other relevant stuff
which is a real education!!

see here
Type B RCDs - Sibert - automated equipment manufacturer and supplier of process consumables

 

[bumcrack]

The type B RCD requirement is debatable if you look at the dc declaration by SMA, for now anyway !

 

[SolarCity]

4000TL does require a 100ma RCD, and it needs to be a type B (which cost a fortune)
It will require an RCD as you don't have simple seperation of AC and DC circuits, as it's a transformerless inverter (according to the big green book).

You don't require an RCD as the 4000TL prevents leakage of DC onto the AC side by design.

 

[moggy1968]

I challenged SMA on this, and they said the regs were wrong. I explained that the regs are the regs and it's up to them to comply. The regs can't be wrong, they are the regs, they may be nonsensce but they're the regs and thats that.
They then said there was going to be a change to the regs to allow for this so it was alright. I explained that whats in the pipeline is irrelevan, it's whats in the regs now I have to comply with., so I don't think I would take SMAs word for anything to do with compliance. They just want to sell you product, your the one responsible for making sure it complies. Imagine the fallout if it became known SMA inverters can't be mde to comply with the regs under some circumstances!!!

 

[moggy1968]

You don't require an RCD as the 4000TL prevents leakage of DC onto the AC side by design.

The regs do say specifically simple seperation, which means a transformer. I'm not sure SMAs inbuilt protection satisfies that requirement.
my interpretation may depend on my needs on the day!!

 

[moggy1968]

do you happen to know if the fronius IG3.6 has the same soret of protection?

 

[BruceB]

The 4000TL instructions do NOT say you need an rcd.
They say IF you have one then it must be at least 100mA.
SMA have also issued a letter that states that their inverters meet the exemption in the regs so that they do not require a type B rcd (if there is an rcd). There is a more complex discussion to be had on that.
The best solution is to design your installation so you do not need an rcd.
Regards
Bruce

 

[bumcrack]

The regs do say specifically simple seperation, which means a transformer. I'm not sure SMAs inbuilt protection satisfies that requirement.
my interpretation may depend on my needs on the day!!

The next line down in the regs from that states that if the inverter is incapable by design of injecting DC into the AC side ( or words to the effect of ), a B RCD is not required. The declaration from SMA satisfies that.

 

[moggy1968]

cool, thanks for that!
any info on the 3.6 fronius, save me going and opening the box up!!

 

[moggy1968]

my discussion with them related specifically to burying cable in walls and the need for a 30ma RCD. They were arguing you could use a 100ma for that, which you can't.
As big solar says, we just try and work around it so we don't need one, but for some reason all our installs at the moment seem to be on TT systems which rather puts the mockers on that!!!

As I have said elswhere, I don't tend to use TLs anyway, I prefer the HF and for a 4kW install will tend to use fronius or power one 3.6s rather than a 4000TL

 

[BruceB]

A point to remember with TT installs is that although the regs do require an rcd on all circuits, they do not specify any value so you can make it as high as you like provided the calcs for touch voltage remain below 50V

 

[moggy1968]

yep, the RCD is only there to provide disconection within the required 0.4 secs, not for personal protection so you could use a 100ma RCD so long as all the other requirements for not using a 30ma RCD are met.
I used this when I rewired my folks house which is TT with the board in the loft. I fed the garage supply in SWA via a 100ma time delay RCD so the garage board, with it's 30ma RCDs will trip first, so my dad doesn't have to trail up to the loft every time he trips the RCD, which I reckoned would be a lot!!

 

[bumcrack]

SMA have also issued a letter that states that their inverters meet the exemption in the regs so that they do not require a type B rcd (if there is an rcd). There is a more complex discussion to be had on that.

Bruce,

Surely this discussion must focus on the likelihood for an IGBT to fail in such a way that it passes DC through. In other power electronics I use, IGBT's do not fail in this way. They either work, or there is a lot of carbon, and just a heat transfer plate where the IGBT used to be.

If an IGBT did pass DC it would be over in less time that would be required to trip any form of breaker...they go with a right old pop.

 

[SibertSolar]

A point to remember with TT installs is that although the regs do require an rcd on all circuits, they do not specify any value so you can make it as high as you like provided the calcs for touch voltage remain below 50V

Ah yes, I missed that valid point on my response to the other thread, thanks Bruce. And you're quite right about the best practice being to design out the requirement for an RCD in the first place. (Or don't use a TL inverter...)