Note that this is a draft for public comment. and anyone can download it here by leaving a name and email address.  You do not need to be an IET member, or logged in to do so. The Email can be a burner if you are paranoid.

However the copy so downloaded is only yours, and may not be further reproduced, except fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988

So, as with many similar documents, providing we are reviewing it or criticizing it.but not reproducing great chunks of it, we are fine to discuss the contents, and arguably to do  so will lead to a more complete set of review comments.

Here are my areas of suggested improvement. I do realise this is very much a work in progress, and my comments/observations  are intended to provoke debate to lead to a better final, not offend.

First a general observation.

1) That some sort of standard is being developed for PEN fault detection is an excellent step forward and to be applauded - for a while we have had the slightly unusual situation that a potentially safety critical device is called up in a standard, but there is no ' and it shall perform like this' requirement to refer to.

2) I am unclear who this document is aimed at - in parts it reads as if it is solely for manufacturers of such devices - some clearly are, e.g. requirements  like terminal labeling etc, and yet the testing proposed seems to be a mixture of production sample factory tests, (IK and EMC tests, glow wire tests, are destructive, or can be, and tested units would not normally be allowed into the supply chain) and other things that would normally be verified on each device, perhaps in factory, or for programmable / configurable devices at installation time and others that probably should to be a part of periodic tests ( i.e. repeated during an EICR in "sparks speak" ).

Perhaps (as I suspect it really is a bit of both) it would benefit from a foreword or two explaining which requirements are which.  Then, in a few places the definitions and the language used in parts would do well to better match the  language of their respective audiences. Electronics equipment designers are not usually sparks as well, and while the physics is the same, the rules of the game inside equipment are not the same as outside, where things are a lot less well under the designer's control, and the meta-language to describe what is happening is also different.

3) Definitions (part of the above comment) the device specific stuff is mostly good but I suggest the following to be at odds with normal electronics use of the same words, and liable to cause confusion.

" Pulsating direct current Current of pulsating wave form which assumes, in each period of the rated power frequency, the value 0 or a value not exceeding 0.006 A DC during one single interval of time, expressed in angular measure, of at least 150 deg"

In the pulse power world, a DC pulse is simply one where the direction of current flow never reverses.   (the magnitude may  fall to zero (or for all intents and purposes zero), or rise to any arbitrary value, but the key thing that makes it pulsed DC, is no current reversal...)

What is being described here however is rectified and chopped mains derived waveforms, not a general pulsed DC. And if we mean less than 60mA at the highest point in the cycle, as opposed to the RMS (or do we elsewhere it implies we might ?) then it should say so.

"Current delay angle Time, expressed in angular measure, by which the starting instant of current< conduction is delayed by phase control."

I  assume we mean the no of mains cycle degrees or radians delay, relative to the rising zero crossing of the voltage waveform, or perhaps relative to where the current zero-crossing  would have been had there not been 'phase control',  or as I suspect is mean, conduction angle or duty cycle control...

4) Factory Test or in service re-test limits are not spec limits.

A device picked off the production line will not trip every time a ramping supply reaches 262.200000 RMS . It will trip within some tolerance of this. (say between 260 and 265, for an equal balance of failures on either side false alarms and missed dangerous events, or perhaps between 255 and 262.5 if the latter is intended as a limit that may never be exceeded, so that we err on the side of the false alarm)

It would be better if the voltage spec was written in a way that  acknowledged this reality.

5) A thought about the supplies waveforms of the future. Around line 2012 in the draft it considers power factors for short circuit tests - I presume the low PFCs for high PSSC are assumed to be inductive (i.e. the current waveform is related to the voltage by phase offset but would if sustained, be a sine wave.) I can see how a reduction in R and the predominance of L can be  related to moving up the street main towards a substation transformer, but considering the greater incidence of inverter derived mains, is this a wise assumption for the future ? Note also that in many areas with a lot of solar farms, at certain times of day the mains voltage as supplied  has become quite noticeably trapezoidal, rather than sinewave I suspect this is  a trend that will continue.

Personally I have nothing against "Acme thread" waveforms rather than traditional sines - they are easier on the rectifiers in terms of a lower and  flatter peak current after all, but we probably need to include them in our consideration,  as substations with intelligent inverters to push energy from lightly loaded phases into more heavily loaded ones will likely become more common, as the electronics is already cheaper than a second transformer..

However it affects both power factor (which is a mix of phase shift an non-sinusoidal distortion) and the detection thresholds - as most fast methods are really a peak detector with a degree of integration - true RMS can be done, but is either slow if analogue, or involves waveform digitization and a calculation. More common is a 'near enough' RMS that works for near sine-waves and is progressively more inaccurate as the waveform is distorted.

Anyway this is getting a bit war and peace, and that was not the intention. Comments/flames etc awaited with interest.

Mike.

Note that this is a draft for public comment. and anyone can download it here by leaving a name and email address.  You do not need to be an IET member, or logged in to do so. The Email can be a burner if you are paranoid.

However the copy so downloaded is only yours, and may not be further reproduced, except fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988

So, as with many similar documents, providing we are reviewing it or criticizing it.but not reproducing great chunks of it, we are fine to discuss the contents, and arguably to do  so will lead to a more complete set of review comments.

Here are my areas of suggested improvement. I do realise this is very much a work in progress, and my comments/observations  are intended to provoke debate to lead to a better final, not offend.

First a general observation.

1) That some sort of standard is being developed for PEN fault detection is an excellent step forward and to be applauded - for a while we have had the slightly unusual situation that a potentially safety critical device is called up in a standard, but there is no ' and it shall perform like this' requirement to refer to.

2) I am unclear who this document is aimed at - in parts it reads as if it is solely for manufacturers of such devices - some clearly are, e.g. requirements  like terminal labeling etc, and yet the testing proposed seems to be a mixture of production sample factory tests, (IK and EMC tests, glow wire tests, are destructive, or can be, and tested units would not normally be allowed into the supply chain) and other things that would normally be verified on each device, perhaps in factory, or for programmable / configurable devices at installation time and others that probably should to be a part of periodic tests ( i.e. repeated during an EICR in "sparks speak" ).

Perhaps (as I suspect it really is a bit of both) it would benefit from a foreword or two explaining which requirements are which.  Then, in a few places the definitions and the language used in parts would do well to better match the  language of their respective audiences. Electronics equipment designers are not usually sparks as well, and while the physics is the same, the rules of the game inside equipment are not the same as outside, where things are a lot less well under the designer's control, and the meta-language to describe what is happening is also different.

3) Definitions (part of the above comment) the device specific stuff is mostly good but I suggest the following to be at odds with normal electronics use of the same words, and liable to cause confusion.

" Pulsating direct current Current of pulsating wave form which assumes, in each period of the rated power frequency, the value 0 or a value not exceeding 0.006 A DC during one single interval of time, expressed in angular measure, of at least 150 deg"

In the pulse power world, a DC pulse is simply one where the direction of current flow never reverses.   (the magnitude may  fall to zero (or for all intents and purposes zero), or rise to any arbitrary value, but the key thing that makes it pulsed DC, is no current reversal...)

What is being described here however is rectified and chopped mains derived waveforms, not a general pulsed DC. And if we mean less than 60mA at the highest point in the cycle, as opposed to the RMS (or do we elsewhere it implies we might ?) then it should say so.

"Current delay angle Time, expressed in angular measure, by which the starting instant of current< conduction is delayed by phase control."

I  assume we mean the no of mains cycle degrees or radians delay, relative to the rising zero crossing of the voltage waveform, or perhaps relative to where the current zero-crossing  would have been had there not been 'phase control',  or as I suspect is mean, conduction angle or duty cycle control...

4) Factory Test or in service re-test limits are not spec limits.

A device picked off the production line will not trip every time a ramping supply reaches 262.200000 RMS . It will trip within some tolerance of this. (say between 260 and 265, for an equal balance of failures on either side false alarms and missed dangerous events, or perhaps between 255 and 262.5 if the latter is intended as a limit that may never be exceeded, so that we err on the side of the false alarm)

It would be better if the voltage spec was written in a way that  acknowledged this reality.

5) A thought about the supplies waveforms of the future. Around line 2012 in the draft it considers power factors for short circuit tests - I presume the low PFCs for high PSSC are assumed to be inductive (i.e. the current waveform is related to the voltage by phase offset but would if sustained, be a sine wave.) I can see how a reduction in R and the predominance of L can be  related to moving up the street main towards a substation transformer, but considering the greater incidence of inverter derived mains, is this a wise assumption for the future ? Note also that in many areas with a lot of solar farms, at certain times of day the mains voltage as supplied  has become quite noticeably trapezoidal, rather than sinewave I suspect this is  a trend that will continue.

Personally I have nothing against "Acme thread" waveforms rather than traditional sines - they are easier on the rectifiers in terms of a lower and  flatter peak current after all, but we probably need to include them in our consideration,  as substations with intelligent inverters to push energy from lightly loaded phases into more heavily loaded ones will likely become more common, as the electronics is already cheaper than a second transformer..

However it affects both power factor (which is a mix of phase shift an non-sinusoidal distortion) and the detection thresholds - as most fast methods are really a peak detector with a degree of integration - true RMS can be done, but is either slow if analogue, or involves waveform digitization and a calculation. More common is a 'near enough' RMS that works for near sine-waves and is progressively more inaccurate as the waveform is distorted.

Anyway this is getting a bit war and peace, and that was not the intention. Comments/flames etc awaited with interest.

Mike.

mapj1 said:

1) That some sort of standard is being developed for PEN fault detection is an excellent step forward and to be applauded - for a while we have had the slightly unusual situation that a potentially safety critical device is called up in a standard, but there is no ' and it shall perform like this' requirement to refer to.

I agree, I think that a UK standard for PEN fault detection and a testing system (factory testing and also Electrician onsite testing.)  is well overdue.  I think that the likes of Megger, CA, TIS, Fluke need to get involved as I suspect that a real world test of a PEN fault detector in a Domestic Dwelling or a Commercial/Industrial scenario will have its own set of technical chanllenges.

For the last few years people have been fitting PEN fault devices and also EVSE which claim to have PEN fault detection built in them with no real way of knowing if they will work correctly.

As a side note.   Could a PEN fault detection not be enabled in a Smart meter in some way?

Could .... detection not be enabled in a Smart meter... ?

Oddly we did discuss this very thing a while back but that part of the forum is no more, or I would just link to it.

But the conclusion was "yes" at least for the simple mode of detection that does not need an electrode - i.e. cut off if the L-N voltage goes widely out of spec for more than the usual inrush period of large loads, it certainly would be technically feasible, and as a bonus save the most expensive damage that a lost PEN causes.

Indeed some drawings that look  very like the vector phasor-grams that appear in the draft spec were produced by one of the forum regulars to aid the discussion and to show the corner cases that the technique would not cover.

Sadly, rather like the argument that 'we would like an isolator (or even an RCD) at every meter point as part of the smart meter roll-out' that ship has sailed, and the cheapest solution with no such additions has won the day as the de-facto standard.

And despite promises to do so , even the idea of periodic visits to check the smart meter meter and main fuse for signs of impending failure seems to also be falling by the wayside.

Mike

Could a PEN fault detection not be enabled in a Smart meter in some way?

In principle, yes, we have discussed the possibility here before (and would have other advantages, e.g. informing the DNO of the problem as the same time, so reducing the risk as time to repair would be reduced, not to mention reducing the magnitude of the diverted N current by disconnecting loads). The gotcha in term of EVs though is that disconnecting the PE at the origin is often defeated by main bonding to services shared with other installations - to the EV might still be connected to the dangerous PEN - via the  c.p.c. to MET to main bonding to pipe to next door to their bonding to their MET and to the DNO's PEN. Putting the device next to the EV should eliminate parallel paths when the contacts are open.

   - Andy.

The meter can't even disconnect the PE in the consumer's own installation, as we currently wire things.  We would have to switch to the American system where the N-E link is on the consumer's side of the meter.

I must admit that when I saw the title of the thread, I thought that this was a new proposal for yet another safety device at the origin or in the DB.

Might there be some merit in this option for new builds where mains services arrive in plastic pipes and a new type of meter could be mandatory?

Or just have a triple pole switch in the meter rather than a single pole one?

   - Andy.

Chris Pearson said:

I must admit that when I saw the title of the thread, I thought that this was a new proposal for yet another safety device at the origin or in the DB.

This sort of device wouldn't be effective in installations with extraneous-conductive-parts that are shared with other installations, or other connections to the PME system ... it would only work for equipment outdoors.

A device that reports back to the DNO so they could investigate areas they are getting multiple "pings" from, would be far more effective.

gkenyon said:

A device that reports back to the DNO so they could investigate areas they are getting multiple "pings" from, would be far more effective.

Surely not difficult to put in a smart meter.

I fully appreciate the limitation concerning extraneous CPs, but do modern dwellings have them?

Or indeed if it allowed the DNO to decide to automatically switch off the local substation after some no of fault reports.

None of this is possible at the moment, politically as the functions of DNO and metering company are separated, and physically as substations currently do not have remote controlled LV switching, and the HV areas, that do, cover too large an area for practical fault isolation. But if far sighted good engineering were paramount in the design of the smart grid, it is the sort of thing that would be going in about now to keep it going for the next couple of hundred years.

Mike

None of this is possible at the moment, politically as the functions of DNO and metering company are separated,

Not entirely - SMETS 2 meters don't talk direct to your supplier (unlike the older SMETS 1 ones) - but to one centralized organization (the Data Communications Company (DCC)) - which then passes data on to the appropriate supplier (thus allowing the meter to remain 'smart' when you switch suppliers). So data could be passed onto the appropriate DNO just as easily (and apparently some DNOs are getting power loss data and have been acting on it - e.g. sending someone out in a van to berate whoever had pulled the service fuse without authorization).

   - Andy.