This is beginning to do my head in.

Power-One state that an AC type 300ma RCD should be fitted on the AC circuit when using an Aurora transformerless Inverter. This we did on an install that was assessed for our MCS accreditation. However we were pulled on the fact that the AC cable ran behind plasterboard for 2 metres and may not be more than 50mm from the surface. We were told that we should fit armoured cable as there was no 30ma RCD on this circuit and this would then meet the requirements of BS7671 17th Edition.

Correct me if I am wrong, but I think every new circuit has to be protected by a 30ma RCD under 17th Edition. This then conflicts with the recommendations from Power-One who state that the reason for fitting the 300ma RCD is "This will aim to avoid wrong protection trips, due to the capacitive leakage current related to the PV modules and that will be present because of the transformer less inverter’s topology."

Fitting a 30ma RCD in the circuit would negate the usefulness of the 300ma unit and lead to nuisance tripping which has been raised in another thread.

To meet the Assessor's requirements we will simply fit armoured cable for that part of the AC cable which is behind the plaster board.

Can someone put me right or suggest how we can connect in the AC circuit and meet both the requirement for the 300ma RCD and the requirements of 17th Edition.

If you run armoured cable or put the ac in trunking then no need for RCD protection

There is no regulation in the 17th edition that says every circuit must be protected by a 30mA rcd.

There is a requirement that where a cable is buried less than 50mm deep in a wall that it is protected by a 30mA rcd or one of the other measures listed, eg SWA cable. Also there is a requirement for socket outlets to be protected, with some exceptions.

Thanks Guys. Maybe being misled by my Spark!! Dug out 17th Ed and re-read 522.6.101, 102 and 103. Would have to agree with you. (and my assessor!)

In the multilingual installation manual the instructions say an rcd can be fitted optionally and then they give a value of 300mA. But they do not say one should be fitted.

Bruce
Decision based on this statement from Power-One which seems a little bit more than optional/. (hope it attached).

I see where you are coming from, but I still read that as saying if you are going to put an rcd in the circuit for any reason, then it should be 300mA

It should and for other reasons not stated by Power-One it will need to be a Type B

What they have said is true, it is what they haven't said that creates the requirement for Type B.

More info to follow on why Type B later. (Whole paper being written at the moment)

I am interested what you are going to come up with Worcester because the Power One letter says:

"The detection of leakage currents that could be present in AC lines requires an additional external device.
For the AC line protection, with reference to the German standard VDE100-712, if the PV or Wind inverter
isn’t able to inject direct current to ground (i.e. isn’t able to leak direct current), it is not required to install a
B-type RCD. For this reason, it is not required to install a B-type RCD at the output of an Aurora Inverter."

which is pretty unequivocal about NOT needing a Type B.

Regards
Bruce

You are quite correct Bruce. Specifically regarding the inverter's ability to inject DC onto the AC circuit, they are correct to state that a Type-B RCD is not required. The same goes for SMA and Fronius TL inverter statements. What they don't consider is the PV installation as a whole (the PV design/installation engineer is responsible for the WHOLE installation, not just the inverter) and, particularly, the scenario where a pure DC ground fault might occur on the DC side, and the potential impact of that when using a TL inverter. The below statement and JPG image might help to make sense of what I'm trying to say.

"In the event of a restricted earth fault on the DC side of the installation, the DC and AC components of the residual current depend upon the topology of the PV inverter and level of DC voltage (PV array). The residual current will be the algebraic sum of the AC component (50Hz in the UK), DC component and a high frequency component determined by the PV inverter's switching frequency and EMC filters (typically 16-22kHz). The form of the residual current will thus vary along with the DC content, which means it would not be safe to use a Type-A RCD, hence the reason for recommending the use of a Type-B RCD - Reg 712.411.3.2..1.2"

If the pure DC component of such a leakage current is greater than 6mA then a Type-A RCD cannot be guaranteed to perform within its standard scope of operation (see Reg 133.1.3). SMA themselves have acknowledged that the smooth DC content of a residual current can affect the tripping characteristics of a Type-A RCD by as much as 30% based on their tests/findings.

There are other considerations/discussions to be had regarding 30mA "human protection" requirements and conflicts with the TL inverter manufacturer's recommendations to use higher trip-current level RCDs, to avoid nuisance tripping, but I might have to come back to that later on.

I'll step back out of this for a while as I don't want to appear "sales-oriented", my aim has always been to provide technical input/support surrounding this topic (as various previous threads on this subject have shown hopefully), which I will continue to do so if people are interested or require it.

OK thanks. I'll wait to see a full discussion although 2 quick comments:
- for a ground fault on the dc side, from what I understand, SMA and Power-One inverters (and probably them all) monitor that continuously so one ought to be able to rely on the inverter picking that up and cutting off.
- in your (or Doepke's) diagram, I would normally consider the steady dc component to be to the level through the middle of the sine wave, not to the bottom of it (top right of diagram).
Regards
Bruce

Hi Bruce,

Yes, but the RCMU (residual current monitoring unit) fitted inside most TL inverters will typically only react to residual current "jumps" (ohmic) from 30mA upwards. Slowly changing residual currents (as would be found with, say, a DC leakage current to ground by damaged/broken-down wiring insulation within the DC installation), both AC and DC though, would only be detected between the 100mA-300mA range.

An RCMU is not a residual current protective device (RCD/RCCB) and therefore is not an AC-DC sensitive residual current circuit breaker. It is not intended to replace an RCD needing to be fitted at the AC side of the inverter for fault protection (or fire protection) according to the relevent installation/erection specifications for automatic disconnection (eg, VDE 0100-712 / IEE Reg 411.1).

Regarding the DC component illustration, I think the position of the output wave, and transposed DC component (ignoring any HF component) are dependent on the topology of the thyristor/firing circuit in the inverter and which transistors/thyristors are firing. Not 100% sure on that but I think the diagram is symbolic/representative without being specific for the reason that topology/firing circuits are varied.

The linked file here might help a bit more perhaps,

Andy

Hi Andy,

Thanks for the link to the presentation which I have read and I think understood, notwithstanding the Germanic English.

No criticism of you, but I have in my mind that Doepke's business in part is selling rcds. Yes, I am getting old and cynical.

If we have to protect persons at 30mA (rather than say protect against fire) for a fault on the dc side as described, then I can see the reasoning. But I am unclear why we might wish to do so. We do not do so on the ac side except under specific circumstances (eg sockets & buried cables). Everything is double insulated on the dc side and as far as I am aware we normally consider that sufficient protection. Is there a good reason to go beyond that? If so then it is probably more economical for the inverter providers to change the sensitivity of the rcmus and make them compulsory so they will provide the necessary detection and protection.

Regards
Bruce

Hi Bruce,

Yes, you are right about Doepke - they are specialists in this field. I am aware of the need to avoid overly sales-oriented input but I can assure you that the aim is not to promote sales for the hell of it, it is much more to try to help installers/designers to understand when certain types of products should be used, under certain scenarios. It's a safety issue and is a highly technical subject. Being an electrical engineer by trade myself, I am more interested in understanding it correctly myself and getting it "right" than pushing one particular product over another, to be honest.

The info about DC-side earth faults is not intended to make people feel that there needs to be additional protection "in case" of a DC fault. Indeed, as you rightly say, the DC circuits should consist of class2 devices and double-insulated wires etc. Whether or not there is a requirement to fit a 30mA RCD on the AC side has nothing to do with potential DC-side faults. It's more to make sure that, IF an RCD is required under the current regulations then the CORRECT type is fitted. Fitting the incorrect type of RCD, under certain scenarios of installation methods, equipment and fault conditions, can in fact compromise the performance of the "incorrect" RCD being fitted and therefore compromise the level of protection expected from fitting such a device in the first place. Ie, Type-AC trip coil saturation by pure DC leakage currents, Type-A trip characteristics moving outside of their required performance criteria, etc, etc.

It's simply an info sharing/educational exercise that's being performed here and, as you know, I have been working on this for some time now. Part of the reason for this is that the inverter manufacturers have muddied the water somewhat with the statements that they have issued regarding their product's ability to inject DC onto the AC circuit. We still see/hear comments about "TL inverters having built-in Type-B RCDs" or "an RCMU is a built-in RCD" or even "I need to fit an RCD to every new circuit" etc and, frankly, it's a little worrying. Clarity and wider consideration of potential scenarios is the aim really. Doepke sell bucketloads of a wide variety of circuit-protection and control products with Type-A/Type-B RCDs being only one area within their portfolio. However, they are one of only 2 Type-B RCD specialists globally who have technical design patents etc (the other being Siemens) and they produce a huge amount of product for other brands currently in use in the marketplace. They have been in the RCD design and manufacturing business for well over 50yrs and they really do know their stuff. I would be more than happy to provide more info about Doepke's history/pedigree and even share some interesting stories with you if you would like to take that off-line maybe?

My aim really isn't to make people feel that "you must use Type-B!", rather it's to try to help understand the reasons for using Type-A or Type-B in the right circumstances, according to the various wiring regs/standards etc when using certain types of inverter under certain types of electrical/earth supply and potential fault conditions.

It is certainly viable for the TL inverter manufacturers to build-in Type-B RCDs into their products, in addition to the existing RCMUs etc but I suspect that such a solution might be a long way off yet. In the meantime, we are here to try to help people avoid situations where circuits/systems are left inadequately protected.

Hope this makes sense,

Andy

PS: One last thing - Doepke didn't develop the Type-B RCD technology as a attempt to create a market for themselves. This was very much a pull from the German industry following some serious accidents where frequency converters/inverters were in use and inadequately protected. It is actually law in Germany to have a Type-B fitted where required under their regulations, not just a recommendation/suggestion of best practice.....

Thanks for taking the trouble to do the long version. I look forward to it all being translated into straightforward rules and guidance in due course. I am happy to try to get my head around it being an electrical engineer by background, but I suspect that for those that spend their time fitting PV in straightforward domestic and light commercial situations they just want simple tabulated rules to go by that will not increase cost unnecessarily.

Regards
Bruce

Yep, you're spot on there Bruce, I think some form of decision-tree/flowchart might be useful.....