For techies - Potential Flaw in Zs/ADS calculations?

[JohnW2]

Discussion in another thread has made me think of something that has never occurred to me before.

In order to ensure that ADS will happen as intended in response to a ‘negligible impedance’ (i.e. taken as zero for calculations) L-CPC fault, we divide “the voltage” by the measured total loop impedance (Zs) to get the fault current, and then compare this with the characteristics of the OPD which, in the case of a Type B MCB, we assume that, in the ‘worst case’, could require a fault current of 5 x In.

Traditionally, “the voltage” we used for this calculation was the nominal supply voltage (230V), but this would mean that magnetic tripping would not occur if the Zs were right on the limit and “the voltage” were less than 230V. Accordingly, as we know, Amd3 of BS7671:2008 introduced the concept of “Cmin” (with a value of 0.95), with the effect that the calculations were done on the basis that “the voltage” was 218.5V - thereby decreasing the specified ‘maximum Zs’ figures.

However, I’ve never previously stopped to think about what “voltage” is relevant, subconsciously probably ‘assuming’ that it is the voltage supplied to the installation in question. However, I’ve just realised that that is wrong ....

The ‘fault loop’ we are interested in, whose total impedance is Zs, is the entire loop from the final circuit in the installation all the way back to the transformer at the substation. The voltage driving current around that loop is the voltage at the secondary of the transformer, so the fault current will be equal to that voltage (at the transformer) divided by the Zs (total impedance) of the circuit concerned - regardless of what ‘supply voltage’ is being delivered to the consumer installation in question.

The supply voltage to premises is allowed to be anything in the range 216.2V to 253V. However, if/when the voltage of supply seen by a customer is fairly low in that range, it is almost invariably due to voltage drop in the LV distribution network between the transformer and the consumer, not due to a very low voltage at the transformer. I find it very hard to believe that the voltage at the transformer is ever (under normal circumstances) less than 230V, and would suspect that it would rarely even be below 240V.

That being the case, unless I’m missing something or thinking all wrong, could it not be said that the regulations are now being almost ridiculously cautious/ conservative in implicitly requiring that ADS will still ‘work’ when the voltage at the transformer is as low as 218.5V - particularly given that that would almost certainly mean that many/most of the customers being supplied from that transformer would be receiving a supply below the ‘minimum permissible’ 216.2V ???

Kind Regards, John

 

[securespark]

I have seen voltages out of that range (as low as 205V), so it would be reassuring to know that ADS would do its job outside the voltage range.

 

[EFLImpudence]

I am a bit confused.

Are you saying that the voltage at a fault would be different than the voltage read - at the same place - when measuring Zs?

 

[JohnW2]

I have seen voltages out of that range (as low as 205V), so it would be reassuring to know that ADS would do its job outside the voltage range.

Yes, but you're presumably talking about the supply voltage at a consumer's installation - which, as I said, is, in itself, irrelevant to the question of whether ADS will work. I still find it hard to believe that, except under conditions of major faults/problems in the supply system/network, the voltage at the transformer would ever be as low as 230V.

More generally, there is surely a limit to how many (each very rare) simultaneous 'co-incidental' happenings we have to cater for - since multiplying them together results in a 'vanishingly unlikely' scenario (which I/we would call 'vanishingly improbable' if EFLI was not looking!). For there to be a problem if, say, we stuck to pre-Amd3 figures/calculations, we would need, simultaneously:
1... A voltage at transformer secondary <230V
2... A Zs which was right on the limit of a 230V-based calculation
3... A 'negligible impedance' L-CPC fault arising
4... In terms of personal injury, a person unlucky enough to be touching a part which became live because of the fault (and, simultaneously touching something close to true earth potential) during the, say, 10-20 second window following onset of the fault before it got cleared by the thermal part of the MCB.

I know that the phrase 'vanishingly unlikely/improbable' gets thrown around quite a lot, but .... !!

Kind Regards, John

 

[JohnW2]

I am a bit confused. Are you saying that the voltage at a fault would be different than the voltage read - at the same place - when measuring Zs?

I'm not quite sure what you're saying (I don't know what you mean by "voltage at the fault" - the voltage across a zero impedance fault will obviously be zero), but what I am saying is this ...

... the circuit through which fault current flows is incredibly simple - a voltage source (the secondary of the transformer) across which is connected a 'load' which consists of the total impedance of all the wiring between transformer and fault (i.e. Zs) in series with the impedance of the fault (assumed zero). The current which flows (obviously {per Mr Kirchoff} the same anywhere in the circuit, including through the MCB) is simply (and this really is 'Electricity 101', aka Mr Ohm!) the voltage at the transformer divided by Zs - regardless of what 'supply voltage' is seen in the installation in question.

Kind Regards, John

 

[EFLImpudence]

There don't seem to be any other takers, so -

What is the difference between the voltage measured when taking a Zs reading and the supply voltage?

 

[JohnW2]

What is the difference between the voltage measured when taking a Zs reading and the supply voltage?

If I understand your question correctly, the answer relies on an understanding of how "Loop Impedance Meters" actually work, and I am not totally sure about that.

You seem to be suggesting/implying that the measurement of Zs involves a measurement of the supply voltage at the time, but I don't think that's the case.

As far as I can make out from the documentation and other reading, my Fluke 1652 measures loop impedance by introducing a 'fault current' of 12A for 10 milliseconds, and measures the voltage change that results (very briefly!) from that introduced current. If that is correct, then the measurement (of Zs) does not require knowledge/measurement of the absolute 'supply voltage' at the time.

Kind Regards, John

 

[EFLImpudence]

You seem to be suggesting/implying that the measurement of Zs involves a measurement of the supply voltage at the time, but I don't think that's the case.

I thought that was the basis for your original query; that supply voltage is different from the transformer voltage.

As far as I can make out from the documentation and other reading, my Fluke 1652 measures loop impedance by introducing a 'fault current' of 12A for 10 milliseconds,

That is correct.

and measures the voltage change that results (very briefly!) from that introduced current.

Doesn't that depend on what the voltage actually is?

If that is correct, then the measurement (of Zs) does not require knowledge/measurement of the absolute 'supply voltage' at the time.

I do not know for sure but as the machine displays the voltage at the time, I would have presumed that it used that figure for its calculations.


Whichever it is, your original question is becoming even less clear.
That is - we have to use Zs for 218.5V even thought the transformer is always 240V (or whatever).

 

[John D v2.0]

Excellent observation, Well at least we've found where the 15% or so leeway is in the standards! Makes you wonder why they made up the .95 bit.

Although remember it is possible to have a negative impedance, if you have a regulated power supply.I don't that applies to the lv network though!

 

[JohnW2]

I thought that was the basis for your original query; that supply voltage is different from the transformer voltage.

My original point was that (since Zs is the impedance of the loop all the way back to the transformer) the Zs required to produce a current adequate to magnetically trip an MCB is dependent only on the voltage at the transformer - and, for a given transformer voltage, that critical Zs will be exactly the same, no matter what the installation's 'supply voltage' is.

Doesn't that depend on what the voltage actually is?

No. An increase in current of 12A will produce a 'voltage change' of 12 x Z (where Z is the impedance of the circuit) regardless of the absolute 'starting' value of the voltage (relative to some reference). For example, if Z=1.5Ω (so that the 'voltage change' will be 18V) then an initial voltage of, say, 250V will fall to 232V whilst that 12A is flowing, whereas an initial voltage of, say, 200V would fall to 182V.

I do not know for sure but as the machine displays the voltage at the time, I would have presumed that it used that figure for its calculations.

I don't think so, and I don't see how it could use that voltage to calculate a valid Zs (for the entire loop).

I assume that it does (easy enough to check) use that voltage (together with the measured loop impedance) to calculate PFC or PSCC - and, if the argument I've been presenting in correct, then those answers will be 'wrong'. As I see it, one could calculate PFC or PSCC correctly by dividing the supply voltage by R1+R2 (for PFC) or R1+RN (for PSCC), but if one is using the full loop impedance (i.e. Zs = R1+R2+Ze), then it is the transformer voltage (not the local supply voltage) that one needs to divide by that impedance.

Whichever it is, your original question is becoming even less clear. That is - we have to use Zs for 218.5V even thought the transformer is always 240V (or whatever).

I'm not sure I understand why it is less clear. My original question is, indeed, why is that second statement apparently the case.

Kind Regards, John

 

[John D v2.0]

I think John's original question is

• the zs is measured through all the wiring back to the transformer
• The voltage across the transformer is unknown and variable but let's assume they make it 250v minimum
• The voltage at your incomer drops from the 250v depending on the current taken on all the people on the same cable and the PFC at the incomer. The more current and the lower your PFC, the lower the voltage you'll see. That could be 205v but should be over 218 or so.
• The relevant equation is basically If=V/Zs
• John's point is that V should be the voltage at the origin of that whole circuit including the transformer, not the incoming voltage, which would be influenced by the current taken as mentioned previously.

 

[JohnW2]

I think John's original question is ... The voltage at your incomer drops from the 250v depending on the current taken on all the people on the same cable and the PFC at the incomer. The more current and the lower your PFC, the lower the voltage you'll see. That could be 205v but should be over 218 or so.

I'm not quite sure how PFC comes into that - are you perhaps talking about the very brief period whilst a fault is present, before it gets cleared?

The relevant equation is basically If=V/Zs
John's point is that V should be the voltage at the origin of that whole circuit including the transformer, not the incoming voltage, which would be influenced by the current taken as mentioned previously.

That is, indeed, exactly my point.

Kind Regards, John

 

[EFLImpudence]

Ok. I see what you're saying.

Why then, is the voltage always the same - all circuits are from the transformer and back to it - and why is it not "influenced by the current taken as mentioned previously" as there are common cables






I have noticed that the Fluke specifications say:

Measuring range: 100 - 500V AC (50/60Hz)
Maximum test current at 400V 20A for 10ms
Maximum test current at 230V 12A for 10ms

Does that not suggest that the meter uses the actual voltage?

 

[EFLImpudence]

These are all red herrings.

You are saying allowing for 218.5V is unnecessary because the transformer is always 240V (or whatever) but you may have the shower and oven on when a fault occurs.

 

[JohnW2]

Ok. I see what you're saying. Why then, is the voltage always the same - all circuits are from the transformer and back to it - and why is it not "influenced by the current taken as mentioned previously" as there are common cables

The transformer will not provide perfect regulation but, within its limits, the output (secondary) voltage will be relatively unchanged by varying load. What will be "influenced by the current taken a mentioned previously" is the supply voltage as seen at one of the supplied premises - that will vary considerably according to the total current flowing in the parts of the LV network which are common to multiple premises - but that variation is almost entirely due to VD in the supply cables, not to a fall in output voltage of the tranny.

I have noticed that the Fluke specifications say:
Measuring range: 100 - 500V AC (50/60Hz)
Maximum test current at 400V 20A for 10ms
Maximum test current at 230V 12A for 10ms
Does that not suggest that the meter uses the actual voltage?

I don't think so. As I've said, if it used the 'local' supply voltage to installation for its calculations, it will get an incorrect answer for Zs.

I think what those figures are saying is that it uses the same inserted 'fault impedance' regardless of voltage, since 12A x (400/230) = 20.9A (not far from 20A). With a fixed 'fault impedance' applied, the current which flows will be determined by the value of Zs. If Zs was high, the currents would probably not reach anything like those quoted 'maxima'. However, if the Zs was very low, the current could rise to a value which could damage the meter - which is presumably why it limits the current to a 'maximum' value which will not result in damage. Why they are imposing different limits for different voltages, I don't know.

Kind Regards, John