Current sensing in an EV wall box

[Gyro]

what does it matter to you? my electrician was sweating badly when he tested mine and got dodgy figures. it reported an insane earth path resistance that we both know was wrong. He then as randomly as he got a fail got a pass and put those figures down. You tell me, was he right or wrong? I am happy, I watched him and helped him install it and knew very well that the install was fine. The meter was one side of the wall and the charging point the other side, total cable length was no more than 2m.

There are rules beyond your original topic that govern the installs. So you get the design right, leave the electrician to get the install right and everyone is happy. There are people all over the place that break rules in all sorts of things. The idea is that there are sufficient checks in place to catch the persistent wrong doers.

Ultimately I could have put that charging access point in myself with no obligation to show a future tenant or buyer the install paperwork, then what? That is why your job apparently is to design a device so safe that even if mis-installed would protect itself, the car and people. Mine has cut out on an RCD fault.

The difference is that you bought a hopefully legit charge point, and therefore had some reasonable confidence that the install was actually safe and acceptable!

[Faringdon]

Mine has cut out on an RCD fault.

Thanks, thats interesting....the RCD is a bugbear for EVSE's.........if you buy offtheshelf you have to buy the RCD (30mA/6mA) and relay in one package, and it costs a fortune.....so people want to make them themselves...and designing a 30ma/6ma RCD needs a fluxgate and isnt a quick cheap job.

Do you know if its possible to buy a type B RCD (30ma/6ma)  which doesnt contain the relay/contactor?.....but rather just a signal for you to use to actuate a relay/contactor?....i am pretty sure such a thing doesnt exist on the market?
................-----------------------......................
Incidentally, what would you think if an EVSE tripped out on 10mA of AC imbalance (instead of 30mA)?......thats a standards fail!....even though its safer as it trips out on less current.
............---------------------....................
Also, is there any reason that you cant use a cheap type B RCD, with a cheap relay in it...but just use it to instigate  a signal to turn off a more high quality relay/contactor?
Cheap Type B RCD (£20)
https://www.toolstation.com/axiom-incomer-devices/p90500?store=I6&utm_source=googleshopping&utm_medium=feed&utm_campaign=googleshoppingfeed&mkwid=s_dc&pcrid=515847200330&pkw=&pmt=&gclid=Cj0KCQjwvaeJBhCvARIsABgTDM4osmCd5dA_6O6J_ZpILg38mJGsz5GgxGBm0qkjN-WYnVQY5aEzKKoaAqE7EALw_wcB&gclsrc=aw.ds

[Simon]

As most of us are not privy to the standards we can't answer what you do and don't need.

Your link does not clearly explain what type the RCD is, which is part of being cheap!

Get the standards, read them, understand them because they also will not tell you how to design something only what the end result must be. Note that meek words like "shall" and "should" in standards mean: "YOU MUST DO THIS OR YOUR SORRY ASS WILL WIND UP IN JAIL AND IT'S YOUR FUCKING PROBLEM TO WORK OUT HOW TO ACHIEVE THIS AND THEN DEMONSTRATE HOW IT ADHERES TO THE STANDARD, ELSE YOU GO TO THE AFFORE MENTIONED JAIL".

My house earthing appears weird. It is both connected to the neutral and the conduit from the substation. My electrician has suggested I contact the distribution network to get them to confirm which it should be. My neighbours house is the same.

[dmills]

No access to the standards, but fluxgate sensors for DC measurements like that seem to be an off the shelf Part, Kemet FG-R05-3A for one example among others.

In sufficient production volumes it is probably cheaper to do your own, but certification costs make this sort of thing a better buy in at more modest volumes.
I was actually not aware that fluxgate was an off the shelf technology, but it appears to have very much become a thing.

Seriously, this shit is just not that hard, I found that part in distribution in < 30 seconds.

In standardese "Shall" is usually mandatory to comply, "Should" and "May" are quality of implementation things, "Should" generally being stronger then "May", but there will probably be a bit at the start of the document which explains this.

Thing is, there are PLENTY of standards around street furniture and lighting and I know you have played in that space so this shit should not be anything new.

[nightfire]

Mine has cut out on an RCD fault.

Do you know if its possible to buy a type B RCD (30ma/6ma)  which doesnt contain the relay/contactor?.....but rather just a signal for you to use to actuate a relay/contactor?....i am pretty sure such a thing doesnt exist on the market?
................-----------------------......................
Incidentally, what would you think if an EVSE tripped out on 10mA of AC imbalance (instead of 30mA)?......thats a standards fail!....even though its safer as it trips out on less current.
............---------------------....................

RCD Type B per Definition does NOT trip at 6mA DC, because it
a) can and shall allow higher DC leakage currents
b) for this reason it is NOT allowed to put a RDC type B behind a RCD type A


Tripping due to imbalance:
A standard RCD is allowed to trip within 50% and 100% of rated currrent, this means a 30mA RCD (AC) is allowed to trip anywhere between 15mA and 30mA. In the factory it is usually produced to get off at something between 20mA and 25mA to allow for some tolerance to both sides.
To answer that question: A EVSE, where the RCD trips outside spec, is considered defective and should be returned to point of purchase. Period.

A RCD Type B has different ranges on DC leakage current: Here we are talking about 15mA to 60mA DC as normative standard.


For all those, that maybe want to dive deeper in RCD technologies: German company Doepke publishes the so called "Allstromfibel", which is also available in english.
https://www.doepke.de/en/services/downloads/

[Simon]

In standardese "Shall" is usually mandatory to comply, "Should" and "May" are quality of implementation things, "Should" generally being stronger then "May", but there will probably be a bit at the start of the document which explains this.

I didn't know the words themselves had um, standardized meanings. But standards are not the law. This is where it's a bit of a mind screw. The beginning of the UK wiring regs standard (BS 7673:2018 - yes I bought a copy, because I felt like it, oh so not expensive at £80 quid) points out that the standard is not the law but laws may impose that these standards are followed or recommend them as best practice. Again, a meek sounding expression - best practice. But in legalese this means it's all shits and giggles about what you care to do until something goes wrong, THEN it is up to you to prove that you followed best practice. This effectively translates into, "just do it this way or else". And this is why installations are supposed to be spot checked at random.

[Faringdon]

For all those, that maybe want to dive deeper in RCD technologies: German company Doepke publishes the so called "Allstromfibel", which is also available in english.
https://www.doepke.de/en/services/downloads/

Thanks, it would be good to know the cost of those, i beleive we have to email them?

[themadhippy]

My house earthing appears weird. It is both connected to the neutral and the conduit from the substation

sounds like they've converted TNS to a TNC-S  at some point,what is unusual is conduit from the substation,plastic ducting yea,some type of armoured cable,yea,but never seen conduit on the suppliers side.

[Faringdon]

but fluxgate sensors for DC measurements like that seem to be an off the shelf Part, Kemet FG-R05-3A for one example among others.

Thanks, that looks a great part, and for $50

BTW, i am still a bit baffled why a RCD trip at 5 or 10mA would be a fail of standards...surely if the standards  threshold is 15mA , and you trip at 5mA, then youre safer than the standards? I cant see how you could be jailed over this,  or even fined.

[Simon]

My house earthing appears weird. It is both connected to the neutral and the conduit from the substation

sounds like they've converted TNS to a TNC-S  at some point,what is unusual is conduit from the substation,plastic ducting yea,some type of armoured cable,yea,but never seen conduit on the suppliers side.

Well it is some sort of armoured cable I suspect as it's external braiding that has be twisted into a wire and bonded to the neutral. Yea it does the electricians head in.

[Simon]

BTW, i am still a bit baffled why a RCD trip at 5 or 10mA would be a fail of standards...surely if the standards  threshold is 15mA , and you trip at 5mA, then youre safer than the standards? I cant see how you could be jailed over this,  or even fined.

Oh FFS get a job stacking shelves in your nearest supermarket!

Who said you go to jail or get fined if it trips lower than the standard? Again unless YOU buy the standard YOU will never know, those of us here not trying our hands at designing car charger access points could not give a flying toss what a standard we do not need to buy says.

If an RCD is too sensitive it will start tripping out at random due to leakage currents!!!!! That will make your customers mighty pleased now wont it? Your company will go bust due to all the returns of goods that are unusable! If the earth and neutral are ultimately connected together don't you think that the capacitance between the live and the earth will cause some permanent leakage? especially where the usual flat cable is used in the UK with the earth wire locked between the live and neutral.

When I connected my parents house that we rented when I lived in italy where dodgy electrics abound to earth I actually got current passing through me, the whole of the leakage current in the house was passing through me as I held the earth wire in one hand and the water pipe that was I was about to bond it to.

[nightfire]

For all those, that maybe want to dive deeper in RCD technologies: German company Doepke publishes the so called "Allstromfibel", which is also available in english.
https://www.doepke.de/en/services/downloads/

Thanks, it would be good to know the cost of those, i beleive we have to email them?

Nope. Free to download.

[Faringdon]

If an RCD is too sensitive it will start tripping out at random due to leakage currents!!!!! That will make your customers mighty pleased now wont it? Your company will go bust due to all the returns of goods that are unusable! If the earth and neutral are ultimately connected together don't you think that the capacitance between the live and the earth will cause some permanent leakage? especially wh

Thnaks, this is interesting, i see your point, now an EV will be the only thing coming off that wallbox, and the leakage from the charger will be just that bit from the y caps....so were taking maybe 3mA max.........this wouldnt spuriously trip an RCD which tripped at 5mA  to 30mA.

[Faringdon]

Hi,
We wish to make a RCD for an EVSE single phase  “wall box”. We will use a simple CT for the 30mA  imbalance detection. This may get  “blinded” by any DC leaking  in either line or neutral.
Therefore, we will also have a second CT the same as the first one, but will constantly  also inject 30mA AC 50Hz   through its aperture with a  wire loop. As soon as this second CT does not detect the 30mA , then that means it has been “blinded” by a DC current flow in either line  or neutral, so therefore we turn off the relay.
(We will inject the 30mA with constantly alternating polarity.)
Would this pass regulations? Why not? It surely guarantees safety, and is far cheaper than offtheshelf RCDs for 30mA/6mA  which are £30 upwards.

[nightfire]

In a single-phase setup DC recognition is not used by many of those ICCB that are use to charge a car from an ordinary outlet.

Reason: During operation, DC currents cannot happen.

Depending on the stuff you want to do, it might not be necessary- here you would have to look it up in the applicable standards.

Other question: If you really need a 6mA DC detection, why is a ready-made RCM module no way? They are regularly used by some big brands in their 3-phase wallboxes, which are sold in germany... In those a standard RCD Type A is used and a dedicated RCM for 6mA DC recognition is coupled via relay to a circuit breaker to switch off the box upon detection.
Makes the setup modular and repairable.

[Faringdon]

Thanks, but those offtheshelf RCDs are upwards of 30 euros each.

With an EVSE, i am not sure if the standards require DC detection, or just an assurance that any DC fault current wont blind the RCD.?

I know the metrel EVSE testers actually do a test for DC leakage current, and if the EVSE doesnt trip at >6mA DC  than thats a fail, (or at least, a fail according to the metrel tester)....whether its a standards fail i dont know.?

[Gyro]

REM  Loop indefinitely (buy standards, understand standards, implement according to standards)!
GOTO 10

[mikeselectricstuff]

Cheap Type B RCD (£20)
https://www.toolstation.com/axiom-incomer-devices/p90500?store=I6&utm_source=googleshopping&utm_medium=feed&utm_campaign=googleshoppingfeed&mkwid=s_dc&pcrid=515847200330&pkw=&pmt=&gclid=Cj0KCQjwvaeJBhCvARIsABgTDM4osmCd5dA_6O6J_ZpILg38mJGsz5GgxGBm0qkjN-WYnVQY5aEzKKoaAqE7EALw_wcB&gclsrc=aw.ds

That is not a type B

Type B's start at around £100

The type is indicated by these symbols. It can be confusing, especially with RCBOs as the same letters are used for RCD type and overcurrent time characteristic, which was a really dumb decision by somebody

[mikeselectricstuff]

With an EVSE, i am not sure if the standards require DC detection, or just an assurance that any DC fault current wont blind the RCD.?

Why are you not sure ? You have either read and understood the standards or you haven't. In the latter case you are just wasting yours and everyone else's time.
Standards are not always the easiest things to understand, but things like this are usually pretty well specified.
I only have access to BS62752, for in-cable level 2 EVSEs but I'd be surprised if the standard for fixed units is any different in this respect.
This states :

8.8.4 Operating characteristics with smooth d.c. residual current
The IC-CPD shall verify that smooth d.c. residual currents do not exceed the value of 6 mA
and ensure a break time according to Table 3 in case that this limit is exceeded.
Compliance is checked by the tests of 9.7.6

What's not to understand?

[Faringdon]

Type B's start at around £100

Thanks, BTW, sorry  i forgot to add, that we dont need the type B RCD to contain the relay...just the signal.

Also, BTW, RCD in EVSE is covered by standard EN62955 (which we cant afford yet) .....as you know, its a "30mA/6mA" RCD,  and these can be made  "home-brew" style very cheaply (a lot less than £100).....what is the cheapest you think a  "home-brew" RCD (Type B)  could be done for? (for use in EVSE)...and not inclduing the relay...just the signal

[Zero999]

Thanks, but those offtheshelf RCDs are upwards of 30 euros each.

With an EVSE, i am not sure if the standards require DC detection, or just an assurance that any DC fault current wont blind the RCD.?

I know the metrel EVSE testers actually do a test for DC leakage current, and if the EVSE doesnt trip at >6mA DC  than thats a fail, (or at least, a fail according to the metrel tester)....whether its a standards fail i dont know.?

For goodness sake. Don't be such a niggard. Spend the money. Failing to design this properly will result in greater costs, later on!

[pardo-bsso]

Tangential to this,

I'm not sure if I understand the problem properly.
I assume (based on this thread and the others...) the current flows from mains, thru the charger and thru the car making a closed loop under normal conditions.

A fault to earth or unsuspecting engineer will divert some of it, no matter if it happens before or after the charger and that difference will be picked up by an rcd upstream.


At least here a simple test of that worked (and the quality of most of our breakers/protection devices is on par with Simons recollection of Italy)

[fcb]

Tangential to this,

I'm not sure if I understand the problem properly.
I assume (based on this thread and the others...) the current flows from mains, thru the charger and thru the car making a closed loop under normal conditions.

A fault to earth or unsuspecting engineer will divert some of it, no matter if it happens before or after the charger and that difference will be picked up by an rcd upstream.


At least here a simple test of that worked (and the quality of most of our breakers/protection devices is on par with Simons recollection of Italy)

Correct. But.... if there is a DC leakage current (EV battery charging electronics?) it can 'blind' a normal RCD (by magnetizing the CT).  The 'Type A' RCD must not be blinded by <6mA DC - technically any more than this can blind/desensitize a 'Type A'.

[Faringdon]

Thanks, is the DC leakage of 6mA (or more, eg 20mA DC)  actually dangerous, or is it just bad because it can blind the RCD?

[fcb]

Thanks, is the DC leakage of 6mA (or more, eg 20mA DC)  actually dangerous, or is it just bad because it can blind the RCD?

IT'S THE FUCKING RULES!
Irrelevant - it's just what the standards say.