RCDs: Selection, Types and Testing webinar questions and answers

Your questions answered.

What type of RCDs should be selected for Comms Cabinets when you don't know what active equipment will be installed? Prior to the 18th Edition, we used to install 32A supplies and commando outlets to cabinets with just MCBs.

The use of such socket-outlets is only one design solution, and any design solution may provide constraints.

Often, the only reason we use 32 A socket-outlets to BS EN 60309-2 for this application (load currents are usually below 16 A), is that prior to the 18th Edition, there was a requirement for 30 mA RCD protection on socket-outlets rated at up to 20 A. Before that, equipment was often connected, in line with international product standards, using 16 A socket-outlets to BS EN 60309-2. In cases where high inrush currents were experienced, direct-wiring was often used.

Since many of those 32 A socket-outlets may be used to supply unfused power distribution units with appliance outlets to BS EN 60320 series. Are we comfortable that adequate fault protection is provided for the appliance flexible cables, if the 32 A socket-outlets are protected by 32 A protective devices?

According to BS 7671:2018, an exception to Regulation 411.3.3 is permitted if a risk assessment determines that RCD protection is not provided.

The current Draft for Public Comment for Amendment 2 to BS 7671:2018 proposes to remove that exception. If that remains a requirement when Amendment 2 is published, an alternative solution for connection of this type of equipment may be necessary.

For clarity, when considering the level of earth leakage likely to trip a device. Did you mention that 30mA for example, could trip with half the earth leakage of its rating? so 15mA would be enough to trip this device?

According to the relevant product standards for RCDs, RCDs must not trip with a current of up to half the rated residual operating current (15 mA for a 30 mA RCD) - however, they may trip at higher currents. In practice, through, instead of tripping at just over 15 mA, a typical 30 mA RCD will trip with residual currents of between 18 and 23 mA.

To prevent unwanted operation ("nuisance tripping") Regulation 531.3.2 recommends that standing protective conductor currents in circuits protected by an RCD are limited to 30% of the residual current rating of the device (9 mA for a 30 mA RCD, or 3 mA for a 10 mA RCD).

What happens to the RCD when you have DC components in the AC waveform. say from electronic components in equipment and how is there guidance for selecting to mitigate against it?

DC components can saturate the core of a Type AC RCD, which can stop it energising the trip coil. Table A53.1 of BS 7671 contains guidance to help select the RCD Type required for particular applications.

Are S-RCDs and FCU-RCDs (to BS 7288) recognised for additional protection in BS 7671?

RCDs to BS 7288 are not recognised for use as an RCD by BS 7671:2018. See Regulation Group 531.3.4.

Clause 1 of BS 7288:2016 states "SRCDs are intended for use in circuits where the fault protection and additional protection are already assured upstream of the SRCD."

Further, Clause 0 of BS 7288:2018 states the devices are only suitable for additional protection against direct contact, and therefore cannot provide fault protection (e.g. where disconnection times cannot be met in a circuit), nor would they be recognised for additional protection against fire due to short circuits in appliances or flexible cables connected to the socket-outlet or connection unit.

However, there is nothing to stop accessories containing SRCDs to BS 7288 being fitted in electrical installations complying with BS 7671, as they comply with the relevant standards, although as stated the RCDs within them cannot be recognised for the functions of fault protection or additional protection for the purses of BS 7671.

When looking at an RCBO in a distribution board, how do I know if it's fitted for fault or additional protection?

There is no way to tell directly. If the RCD is rated above 30 mA, it cannot provide additional protection against electric shock, but RCDs up to 300 mA may provide additional protection against fire.

If you have a split load consumer (db1)unit with 30ma RCDS protecting circuits with less than 50mm2 cabling in wall and you require garage feed (db2) with 30ma and then down stream a garden office, do you require only protection from on (db 1)? As two-way boards have 30mA coming with it.

There is no need to provide a further 30 mA RCD downstream of a 30 mA RCD. If you do, there is no guaranteed selectivity (in a fault, either may operate first).

However, there are occasions in which BS 7671 mandates this (for example, a caravan pitch outlet installed in accordance with Section 708 of BS 7671 must have 30 mA RCD protection, and there must also be 30 mA RCD protection within the caravan according to Section 721).

What do you recommend for an office ring final circuit? RCD protection because of cables in walls, but IT equipment leakage may exceed 30mA.

Regulation 411.3.3 requires 30 mA RCD protection for all socket-outlets rated 32 A or less, unless a document risk assessment determines RCDs are not necessary.

The risk assessment would necessarily have to consider whether, particularly standard BS 1363-2 socket-outlets were available for ""general use"" by those using, cleaning or maintaining the space.

Suitable division of circuits, and/or the use of isolation transformers, can help address unwanted operation of RCDs.

Can you risk assess out the use of an RCD for critical safety equipment in an unmanned area? Such as fire alarms or battery tripping unit in a substation?

That depends on a number of factors, including the reason the RCD is provided in the first place.

Examples provided in the IET Electrician's Guide to Fire Detection and Alarm Systems include:

(a) the requirement for an RCD in Regulation 411.3.3 can be avoided by not supplying such equipment from a socket-outlet, but from suitable connection units.

(b) the requirement for RCDs in Regulation Group 522.6 for cables concealed in walls or partitions may be avoided by the use of alternative wiring systems and/or alternative cable routes.

Where RCDs are used for fault protection, such as in TT systems, it may not be possible to design around, or ""risk assess out"" the use of RCDs.

If you have the ‘wrong’ type (A, AC, F, B etc.…) does that mean the RCD is not detecting the leakage current that you would like it to be detecting, and therefore potentially not tripping out when it should, OR by having the ‘wrong’ type does the leakage current cause the RCD to nuisance trip?

Either is possible, but of course from a safety perspective we are usually more concerned with the first case - potentially not tripping when they should.

As a rule of thumb, how many computer workstations could you protect with a 30mA RCBO protected circuit, if each workstation were provided power for 2 monitors and a desktop computer tower?

That depends on the protective conductor currents of individual products. There is no "one size fits all" solution.

Strictly speaking, Class I pluggable Type A equipment (connects to a BS 1363-2 socket-outlet in the UK) to BS EN 62368-1 is permitted to have a protective conductor current of up to 5 mA, and on that basis, 1 product per circuit would be the limit according to Regulation 531.3.2! However, experience tells us this is not usually the case.

Would the proposed restriction to Type AC RCDs to resistive loads rule them out for most lighting circuits (LEDs) and most socket rings (chargers, PCs etc.)

Yes, for this reason an amendment to Regulation 531.3.3 is proposed in the current Draft for Public Comment of Amendment 2 to BS 7671:2018.

Would it be always safer to then specify Type B RCDs?

Yes.

Would it not be better to just phase out Type AC RCDs to avoid potential issues in the future where the DC component of installed equipment is unknown?

Yes, for this reason an amendment to Regulation 531.3.3 is proposed in the current Draft for Public Comment of Amendment 2 to BS 7671:2018.

When will the market bring itself more up to date to produce Type B RCD's? Currently, from my experience, they are not as readily available as other types, and it has caused me some issues in the past.

When RCDs were first introduced, at that time Type AC, they were very expensive. It is expected that the cost and availability of Type B RCDs will improve over time.

There are now many pieces of equipment with DC motors used for energy saving and speed inverters?

Agreed. Appropriate RCDs should be selected.

Where can 30mA RCD complying with BS7288 be used? As they do not provide additional protection unless protected by a 30mA RCD to BS 61008 or BS 61009 and if they are, do not provide selectivity.

RCDs to BS 7288 are not recognised for use as an RCD by BS 7671:2018. See Regulation Group 531.3.4.

Clause 1 of BS 7288:2016 states "SRCDs are intended for use in circuits where the fault protection and additional protection are already assured upstream of the SRCD."

Further, Clause 0 of BS 7288:2018 states the devices are only suitable for additional protection against direct contact, and therefore cannot provide fault protection (e.g. where disconnection times cannot be met in a circuit), nor would they be recognised for additional protection against fire due to short circuits in appliances or flexible cables connected to the socket-outlet or connection unit.

However, there is nothing to stop accessories containing SRCDs to BS 7288 being fitted in electrical installations complying with BS 7671, as they comply with the relevant standards, although as stated the RCDs within them cannot be recognised for the functions of fault protection or additional protection for the purses of BS 7671.

As there are many pieces of equipment with DC motors and inverter speed control it is difficult to find the leakage, manufacturers say it varies?

It does vary, and may depend on the installation earthing arrangement. Alternative means of installation of VDSs include the use of isolation transformers to control leakage currents. This is recommended in, for example, BS EN 60204-1.

Why would an RCD test be carried out at The Distribution Board with the cover removed. It would be safer to test at a socket outlet using a plug in lead. No exposed live parts.

It is safer to test the RCD at a socket-outlet. However, there are occasions where this is not possible (for example, circuits where there is no socket-outlet, or even connection unit). Of course, we would recommend the safest possible test method is used in every case.

Max disconnection time for RCD trip test at 100% rated Idn per BS 7671 is 200ms while under IEC 60364 is 300ms. In my experience, I cannot recall ever seeing any tripping time in the order of over 100ms. My question is why the inconsistency in the specified tripping time, since the two standards are supposed to be harmonised since 16th Edition or so?

There is no disparity between BS 7671 and IEC/HD 60364 in this case, see Table 3A in BS 7671:2018 (page 363). BS 7671 specifies that RCDs are to comply with either BS EN 61008-series, BS EN 61009-series, BS EN 62423, or BS EN 60947-2.

200 ms disconnection time is required in both BS 7671 and IEC/HD 60364, for automatic disconnection in TT systems, so a prospective fault current of at least twice the rated residual fault current, and a non-delay RCD, are necessary.

It is true that older RCDs to BS 4293:1983 were required to trip in 200 ms, but this standard was formally replaced by BS EN 61008 series in 2009. This standard is not recognised by BS 7671:2018.

Similarly with SRCDs, although versions of BS 7288 prior to 2016 were required to trip in 200 ms, the current version, BS 7288:2016 aligns with the trip times of BS EN 61008 also. This standard is, however, not recognised by BS 7671:2018.

If, during an EICR inspection and test, we discover an RCD socket outlet fed by a single pole MCB what is the C code to use?

That depends on whether the installation was installed according to BS 7671:2018. The installation may comply with an earlier version of BS 7671.

The socket-outlet may comply with an earlier version of BS 7288, which did not have the same constraints requiring upstream additional protection (which were only introduced into the 2016 version of the standard).

As with BS 4293 RCDs, SRCDs complying with earlier versions of BS 7288 would simply be a suitable product for the time.

It would be up to the inspector to determine if there were a safety concern with the particular installation, or whether the arrangement simply didn't comply with the current Edition of BS 7671.

Could you confirm that when testing a type A the type AC test should also be carried out so as to test both segments of the RCD also when testing the type B that you test on AC followed by A then followed by B tests?

The instructions of the test instrument manufacturer should be followed.

Does the Electricity at Work Regulations 1989 allow us to test trip times at the board? My company's Electrical Safety Rules and Procedures don't allow us to do this. All tests need to be done out in the field.

The safest approach should be followed for that particular conditions. For socket-outlet circuits, that is an easy decision.

There are situations in which a test from a socket-outlet is not possible without modifying the installation.

Most RCD test equipment comes with features of 0 deg and 180 deg. Can you please elaborate when these settings are applicable when carrying out the RCD tests?

The normal test sequence is to carry out each of the tests, at least the 1x and 5x test, with each polarity in turn. Not all instruments have a polarity option on the 0.5x test.

On a 4 pole 3phase RCD can this still be used when there is no load neutral? Will the RCD sense that there is no load neutral connected?

Yes, installation instructions will tell you how to install the RCD in these cases.

If sockets and connection units with integral RCDs are not recognised by BS 7671, should these be used?

RCDs to BS 7288 are not recognised for use as an RCD by BS 7671:2018. See Regulation Group 531.3.4.

Clause 1 of BS 7288:2016 states ""SRCDs are intended for use in circuits where the fault protection and additional protection are already assured upstream of the SRCD.""

Further, Clause 0 of BS 7288:2018 states the devices are only suitable for additional protection against direct contact, and therefore cannot provide fault protection (e.g. where disconnection times cannot be met in a circuit), nor would they be recognised for additional protection against fire due to short circuits in appliances or flexible cables connected to the socket-outlet or connection unit.

However, there is nothing to stop accessories containing SRCDs to BS 7288 being fitted in electrical installations complying with BS 7671, as they comply with the relevant standards, although as stated the RCDs within them cannot be recognised for the functions of fault protection or additional protection for the purses of BS 7671.

Testing the RCB for a ring main at the DB doesn’t really measure your possible fault path. Isn’t better to test at the mid point of the ring main?

Testing an RCD on a ring final circuit cannot be a substitute for carrying out the ring final circuit continuity test - especially where 30 mA RCDs are used, a broken cpc will not be detected.

How many computers would you recommend on a single RCD circuit as they have a nasty habit of earth leakage?

That depends on the protective conductor currents of individual products. There is no "one size fits all" solution.

Strictly speaking, Class I pluggable Type A equipment (connects to a BS 1363-2 socket-outlet in the UK) to BS EN 62368-1 is permitted to have a protective conductor current of up to 5 mA, and on that basis, 1 product per circuit would be the limit according to Regulation 531.3.2! However, experience tells us this is not usually the case.

What frequency should you carry out 'routine' testing of RCD's by operation of test button? Between formal inspections?

Regulation 514.12.2 of BS 7672:2018 states six-monthly, although older installations will have labels stating quarterly.

It's not unknown for manufacturers to state "Test Monthly" on the devices themselves.

If the earth impedance test exceeds the value given in BS7671 then do you still carry out the RCB test?

For initial verification the tests need to be carried out in the correct sequence. If any test fails the testing has to stop and any defect rectified and the test repeated. Earth continuity should be verified before getting to the loop impedance test and any RCD testing. See regulation 643.1.

Is it best to install Type A RCDs from here on on ALL final circuits?

In general, yes, and for this reason an amendment to Regulation 531.3.3 is proposed in the current Draft for Public Comment of Amendment 2 to BS 7671:2018.

Would it be common to specify a type B RCD for welding supplies?

No. Welding transformers tend to have high inrush currents. So a Type C or maybe a Type D may be required.

When recording the RCD times, surely the slowest at 5X would be the value to record as this is the additional protection of the RCD. You would have verified the disconnection time for fault protection when doing the Zs!

There is no guidance on this. It is up to you to chose and justify your decision. If you record the longest trip time then you cannot be wrong.

I didn't understand the three wire test, with earth probe connected to upstream neutral. is there somewhere where the rationale for this is described in more detail?

The instrument acts a bit like the test button. It connects the upstream side of the RCD to the downstream side and injects a test current to unbalance the RCD and then measure the disconnection time.

Section 710 0f the IET regs says you should not use type AC RCDs in medical locations, only type A or type B.

This is correct.

Are these the same procedures for all types of RCD, and does table 3A display the correct disconnection times?

Yes to both.

I have encountered issues testing circuits involving drives-mostly in residential projects (VFDs, Elevators). How is the testing of RCDs carried out in such scenario?

As stated during the webinar any load should be disconnected before testing the RCD.

There seems to be a two different opinions on this, but can an RCD be used when a final circuit has a high ELI? This doesn't look to be the case in BS 7671 when looking at the tables for max Zs, however cable calc software packages seem to enable you to do this to clear a high ELI error.

BS 7671:2018 makes it very clear that RCDs can be used for fault protection (ADS) in TN systems, except where there is a PEN conductor on the load side of the RCD - see Regulation 411.4.5. Similarly, Regulation 411.5.2 clearly states that RCDs may be used for fault protection (ADS) in TT systems.

In all cases, the RCD cannot protect against fault current between live conductors, and overcurrent protection is still required.

Some RCD's require 250mA or greater to trip within 40ms is any information available centrally on which manufacturers require this?

The manufacturer's devices must meet the performance standards set out in Table 3A of BS 7671. If they do not trip at 40ms or less at 5 x Idelts n them they are not compliant.

Does this just apply to power circuits?

RCDs are not required on SELV and PELV circuits, and thereby by definition, not required by Ethernet. They are also not suitable for DC power circuits at this time, for any voltage.

The BS 7671 sample RCD test form requires the tripping times to be recorded. In actual field tests, it is almost certain that the trip times are never the same figures when repeated. Hence, what is the purpose of recording these figures whereas seemingly one would reason that as long as the trip times are within the permissible limits, an indication to record "test passed" would suffice, rather than recording a trip time which may be different each time the same test is repeated?

Test times will vary slightly but should not vary significantly. We advise recording the highest trip time.

What type of RCDs are required for electric car chargers (7kW to be installed at domestic properties) to avoid the use of an earth rod?

RCDs disconnecting all poles are required for all EV charging equipment. The RCD must be Type B, unless protection against DC fault currents is provided in the charging equipment (in which case the RCD has to be Type A or Type F).

Because there is no selectivity with upstream non-delay RCDs, any non-delay RCDs must also disconnect all poles.

I have seen when testing a dc motor down stream of type AC RCD, go from normal to 600mS.

Loads should be disconnected before carrying out an RCD test.

RCBO has a 10KA fault breaking capacity. If usually the DNO supply electrical incomer TNCS 80 & 100A with fault current at point of installation 15KA, how can it comply that RCBO 10kA fault breaking capacity at the consumer unit, If the fault current at point of installation is 15kA?

This is addressed in the National Annex of BS EN 61439-3 for relevant assemblies. Similarly, many overcurrent protective devices in domestic consumer units are rated only 6 kA.

Type AC RCDs are currently abundantly in use, I have read that these are to be phased out as more and more electronics based equipment are now being installed. Is there a specified date for these Type A to be phased out?

An amendment to Regulation 531.3.3 is proposed in the current Draft for Public Comment of Amendment 2 to BS 7671:2018.

The webinar used an A-Type RCD upstream of a B-Type in your example for a EV circuit. Bema states this should not take place, please can you confirm?

RCDs disconnecting all poles are required for all EV charging equipment. The RCD must be Type B, unless protection against DC fault currents is provided in the charging equipment (in which case the RCD has to be Type A or Type F).

Because there is no selectivity with upstream non-delay RCDs, any non-delay RCDs must also disconnect all poles.

Could you clarify the requirement to use a Type B RCB for an EV charger? Is it only when there is no in-built fault current protection?

RCDs disconnecting all poles are required for all EV charging equipment. The RCD must be Type B, unless protection against DC fault currents is provided in the charging equipment (in which case the RCD has to be Type A or Type F).

Because there is no selectivity with upstream non-delay RCDs, any non-delay RCDs must also disconnect all poles.

How can you design to not exceed 30% of leakage current of an RCD when it is unknown what a customer will plug into socket outlets? Also why is there no requirement to complete a ramp test or use an earth leakage clamp meter to establish any leakage currents?

That is a matter for the designer. A similar situation exists in accepting that a 20A or 32A socket-outlet circuit is unlikely to be subject to overload.

However, in the case of RCD we should only be talking about "nuisance tripping" not the fire or damage risk of overload. Where equipment is safety-related, or business critical, other options should be considered for connection of equipment that avoids the use of RCDs.

I have heard several speculations about how to make correct a RSD test - from fuse board, connect the leads to RCD or load point, i.e. socket. Which is correct?

It does not matter where on a circuit the RCD is tested. Only that any load is disconnected.

On a new build domestic installation where an EV charger and PV system require RCD protection, should they each have a dedicated device, or can they be installed alongside other circuits in a split load consumer unit?

Both PV and EV chargers require their own dedicated circuit. EV chargers usually have the RCD built in to the charge point. PV cables may require RCD protection depending on cable type and their installation method. It would be advisable for PV and EV circuits to have separate RCD protection to avoid nuisance tripping due to cumulative leakage currents or locking up due to DC leakage or nuisance tripping due to DC where an RCD DD device is fitted.

We often encounter unwanted tripping. Is it good to use 2 pole RCBOs rather than single pole RCBOs in office installations to reduce unwanted tripping?

It does not matter if a double pole or single pole RCD is fitted as both will be subject to unwanted tripping due to cumulative leakage currents.

RCD testers need calibrating from time to time. Is there a simple check to indicate that the tester is accurate that can be carried out by the user between calibrations?

Yes. Instruments should be subject to on going accuracy tests between calibration. Proprietary test boxes are available to carry out these confidence tests.

What are your thoughts on use of RCDs in commercial catering installations which have 3 phase catering equipment such as cookers, dishwashers?

If they are connected to sockets up to 32A then 30mA RCD protection is required unless there is a risk assessment in place to say it is not required. If hard wired without a socket then RCD may be omitted if the manufacturer does not specify RCD protection is required. Very much for the designer of the installation to decide.

Are RCD affected by changes in temperature? Should this be taken into account when testing?

The British Standard specifies 2 ranges of operating temperature. -5 to 40 degrees C and -25 to 40 degrees C. You may get small differences in test times but they should still meet the requirements of the BS. Check with your manufacturer for individual devices as to what range they are in.

Is it legal to put a relay switch with series resister connected like test button so it would work as a shunt trip?

BS 7671 does not permit this (Regulation 531.3.1.202).

The answer to a question related to the test button potentially tripping an upstream device. Please can you explain how? My understanding is the test resistor is connected L-N, across the internal coil - any upstream device would see this as a balanced load?

The test resistor connects the outgoing line to the incoming neutral to unbalance the RCD. It should not trip any upstream device as there is no test current to earth. The upstream device will only see a line neutral current just like a load.

Three phase - the test button wont work without a neutral?

This is correct.