Confused with RCD protection

[saint]

I have a few questions on this hopefully someone can help my confusion.

1. As i understand it SMA and Fronius TL inverters do not require type B RCD protection because they cant feed DC into the AC side of the installation. We use these makes and also Power One and Samil. I have looked on their websites but cant find any info on these makes. Where can i look on line to check this?

2. I have read other threads saying the PV shouldn't be protected by an RCD that also protects other circuits. i.e. shared RCD. Why is this and where is the regulation on this??

3. Provided the TL inverter can not feed DC into the AC installation due to its design, then providing RCD protection is not required to comply with other RCD requirements of BS7671 (i.e. 30mA in for cables in wall etc), then no rcd is required at all? Correct???

thanks in advance.....

 

[JulianC]

1) Yup. Ditto for Power One - can't remember where I saw it, probably in the manual. Dunno about Samil

2) Inverter allowed to take up to 5 secs to shut down in the event of loss of mains under G83. Therefore, RCD trips in the event of a fault and inverter potentially keeps circuits live for up to 5 secs. Same applies to an RCD just protecting the pv circuit as well, if a fault occurs on that circuit.

3) Almost. Regardless of TL inverter's ability to inject DC onto mains, if RCD is not called for by BS7671, no need to fit one. If 7671 calls for an RCD, then you need to decide whether type B is required or not, depending on inverter specs.

 

[Gavin A]

1 - yes for Aurora, no for samil - they don't make the statement, and somewhere in the paperwork it alludes to the higher DC feed in levels.

2 - ALso, why risk the call backs for nuisance tripping?

3 - as per Julian

 

[saint]

An email from Samil....

[FONT=Verdana, Arial, Tahoma, Calibri, Geneva, sans-serif]Due to our inverters being transformer less, as you quite rightly stated, with no separation, all of our inverters will let a small amount of DC current out on the AC side, known as “DC injection”. In order to pass the certification tests (G83 / G59) our inverters must NOT allow above a certain mA amount to leak down the AC line, they must detect and shut off when abnormally high levels of DC injection are present, (DC INJ high – fault) However, they are all transformer-less which means that there will be some DC injection but it should be very little. We have found that a type A RCD is fine. Only on rare occasions have we recommended installing type B. You are not legally obliged to install type B as our inverters are designed and tested to operate according to the UK standards for such devices, and should keep the DC injection levels suitable for use with a type A RCD. You can find the G83/G59 test results on our website. It is also worth noting that It is perfectly normal to install our units without installing an RCD, provided that the cabling is not concealed, as you may already be aware.[/FONT]

 

[Gavin A]

An email from Samil....

Due to our inverters being transformer less, as you quite rightly stated, with no separation, all of our inverters will let a small amount of DC current out on the AC side, known as “DC injection”. In order to pass the certification tests (G83 / G59) our inverters must NOT allow above a certain mA amount to leak down the AC line, they must detect and shut off when abnormally high levels of DC injection are present, (DC INJ high – fault) However, they are all transformer-less which means that there will be some DC injection but it should be very little. We have found that a type A RCD is fine. Only on rare occasions have we recommended installing type B. You are not legally obliged to install type B as our inverters are designed and tested to operate according to the UK standards for such devices, and should keep the DC injection levels suitable for use with a type A RCD. You can find the G83/G59 test results on our website. It is also worth noting that It is perfectly normal to install our units without installing an RCD, provided that the cabling is not concealed, as you may already be aware.

doesn't really cut it does it.

for reference the G83 limits on DC injection are well in excess of the levels at which a type A RCD wouldn't be suitable - I'm sure Andy Sibert will fill in the blanks on that at some point.

 

[BruceB]

The wording the Samil email would have to satisfy is this (712.411.3.2.1.2):

"..... Where the PV converter is, by construction, not able to feed d.c. fault currents into the electrical installation, an RCD of type B to IEC 62423 is not required."

The email to my mind does NOT satisfy that, so in this litigious age I would say that IF you use a RCD with a Samil, then it should be type B or you leave yourself exposed to civil and criminal penalties if something went wrong.

 

[saint]

I agree Bruce...

Can anyone point me in the right direction for SMA, Power One, Fronius?

 

[BruceB]

For SMA:
http://www.sma.de/fileadmin/content...ents/Medium_Power_Solutions/RCD-UEN110430.pdf
http://files.sma.de/dl/22701/SB-SMC-ZE-EK-en-18.pdf

Detailed discussion in this thread which includes the Power One document as a pdf
http://www.electriciansforums.co.uk...forum/43974-power-one-ac-side-protection.html

Do not know about Fronius

 

[SibertSolar]

All good info above, agree with Gavin and Bruce. I do have my own personal view as to the validity of the 712.411 section regarding an inverter's design and DC fault currents etc but let's not go there right now.

Type-A RCD trip coils will saturate if >6mA pulsed DC leakage current exists. Any pure DC component will saturate a Type-A RCD. Type-B RCD will function under AC fault conditions regardless of presence or amplitude of pulsed or pure DC leakage current component.

Compliance with DC injection limits under G83 (20mA) has no relation to compliance with the 712.411 section that Bruce referenced (VDE 0100-712 / IEC 60364-7-712) regarding DC fault currents.

The question that Samil should be asked is "can you guarantee that any DC leakage current component exhibited by the inverter under either normal working, or fault, conditions, consists of no more than 6mA pulsed DC?" If the answer is no then you would be wise to fit a Type-B RCD rather than Type-A (if, of course, it is determined that an RCD is required)