Electrics, Smoke Alarms and Building Control

[JohnW2]

They have to go via the coil, otherwise there could be no magnetic trip action

Whoops, good point - I was thinking of the trip solenoid in an RCD .

However, that inductance is present in the current path of the circuit in question, as well as the connection to other circuits - so would it not restrict the high frequency components in the circuit itself to the same extent as in its connection to other circuits?

Kind Regards, John

 

[endecotp]

[the MCB'S] inductance is present in the current path of the circuit in question, as well as the connection to other circuits - so would it not restrict the high frequency components in the circuit itself to the same extent as in its connection to other circuits?

No.

I've just tried to measure the inductance of a 6A MCB; it was something like 10 uH, though the reading wasn't very stable. For comparison, a 1m long 1 mm diameter straight wire has an inductance of about 1.5 uH.

 

[JohnW2]

No. I've just tried to measure the inductance of a 6A MCB; it was something like 10 uH, though the reading wasn't very stable. For comparison, a 1m long 1 mm diameter straight wire has an inductance of about 1.5 uH.

I don't really understand your "No". I didn't really believed that the reactance of the coil would be enough to achieve much, even in terms of relatively high frequency 'noise' components, but bernard clearly believed otherwise. However, what I was suggesting (to which you appear to have answered "No") was that, whatever (very little) effect the reactance had on transfer of HF components to other final circuits (via the L busbar), one would expect it to have a similar effect on the circuit with the noise itself (well, on reflection, about half, since there would only be one MCB coil in the path, not two). If you are disagreeing with that suggestion, could you please explain?

Kind Regards, John

 

[endecotp]

Two things on the same side of the filter are not separated by it. Things on opposite sides of the filter are separated by it.

As another comparison, typical mains filter chokes are typically of the order of 1-10 mH.

 

[JohnW2]

Two things on the same side of the filter are not separated by it. Things on opposite sides of the filter are separated by it.

Needless to say, I realise that, but I wasn't talking about a 'filter' between the two circuits. I was suggesting that the presence of a significant inductive reactance (anywhere) in the 'primary' circuit (the one with the noise source) will surely reduce high frequency currents flowing in (throughout) that circuit, hence reducing the potential effects of any (HF) noise-producing sources on anything else connected to the same circuit - thereby comparable with the 'filtering' effect of an inductance connecting two circuits. Is that not the case?

Kind Regards, John

 

[endecotp]

No, that's not the case.
High-frequency transients propogate locally due to distributed capacitance and inductance, and may even reflect off distant filtering inductors.
I think a "sound in pipes" analogy is appropriate here.

 

[JohnW2]

No, that's not the case. High-frequency transients propogate locally due to distributed capacitance and inductance, and may even reflect off distant filtering inductors.

Fair enough. However, in addition to those 'local' capacitive and inductive effects, there are presumably also HF components of the current travelling in the conductors?

Kind Regards, John

 

[endecotp]

The wire can be modelled as distributed inductance and capacitance:

https://goo.gl/images/kTBKrb

Current flows through the inductors and in/out of the capacitors.

The current is not constant along the wire.

 

[JohnW2]

The wire can be modelled as distributed inductance and capacitance: https://goo.gl/images/kTBKrb ... Current flows through the inductors and in/out of the capacitors. ... The current is not constant along the wire.

Well, yes, one can regard the cable as a transmission line, but I wonder to what extent that is relevant in relation to the pretty modest frequencies we are presumably talking about in relation to noise generated by electronics (e.g. in lamps and SMPSUs) - in particular, when one is talking about a complete circuit (rather than an unterminated transmission line, as in your link) .

Whatever, I think what I was trying to say still remains. That if one has a complete circuit (as above), if one inserts and additional inductance in the circuit, one will reduce AC currents in that circuit, increasingly so as frequency of the component rises.

In any event, we are letting ourselves get a bit carried away by theory. In relation to bernard's suggestion which started all this, do you believe that there is a significant risk of alarms (which are required to have a certain degree of immunity to EMC) are likely to be affected by HF noise generated by the likes of CFLs and lighting SMPSUs?? If I had been asked to guess, I would probably have said that the very worst I could imagine happening would be inappropriate triggering of the alarm, and that even that would be exceedingly unlikely/rare.

Kind Regards, John

 

[endecotp]

The switching transistor in a switch mode supply is driven by a square wave, and some effort is made to give it sharp edges for efficiency. The frequencies involved - harmonics of the switching frequency - are high. This is exactly what the whole business of EMC testing is about.

Mains circuits (to the substation!) are approximately infinitely long for these purposes.

No I don't believe Bernard's theory, as I said at the time.

 

[JohnW2]

The switching transistor in a switch mode supply is driven by a square wave, and some effort is made to give it sharp edges for efficiency. The frequencies involved - harmonics of the switching frequency - are high. This is exactly what the whole business of EMC testing is about.

Sure, but I would imagine that all but the worst examples make at least some effort to minimise the amount of the higher of those frequencies that gets into the supply.

Mains circuits (to the substation!) are approximately infinitely long for these purposes.

True, but most of that 'infinite' length will be common to the circuit generating the HF noise and the circuit which allegedly might be 'affected'!

No I don't believe Bernard's theory, as I said at the time.

Indeed you did.

In support of that view, as I said I would have thought that if there were any tiny risk, it would probably be of false triggering of an alarm - and, common though it is to experience, or hear of, smoke alarms being activated by cooking etc. activities, I certainly have personally never experienced, and don't think I've even heard of, incidents of inexplicable 'spontaneous' activation of such alarms, such as one would expect to be happening all the time if bernard's theory were correct. I certainly still stick to my personal preference not to have alarms on dedicated circuits

Kind Regards, John

 

[bernardgreen]

Some tests on the impedance of a Wylex B20 MCB using a Wheatstone Bridge circuit

Circuit

If the impedance of the sample is zero ohms then the bridge balances and the voltage at Channel 2 is the same as the voltage at Channel 1

If the sample has a simple impedance ( resistive ) then Channel 1 will be greater than Channel 2 but without phase shift.
If the sample has a complex impedance then there will be phase shifts as well as voltage differences.

The tests were without any load current through the MCB.

If the impedance is zero then the green trace ( Channel 1 minus Channel 2 ) will be a flat line.

At 50 Hz the green trace shows a very small difference which can be explained by the tolerances of the resistor values used in this set up and the resistive impedance of the MCB's coil ( magnetic trip ) and heating element ( thermal trip ).

At 5 kHz sine wave


At 20 kHz sine wave


At 50 kHz sine wave


and square waves got complicated

so I didn't bother to go further.

Enough to suggest that the MCB is an active impedance to high frequency components in the wave form of the mains supply current passing through the MCB

 

[JohnW2]

Some tests on the impedance of a Wylex B20 MCB using a Wheatstone Bridge circuit ... Enough to suggest that the MCB is an active impedance to high frequency components in the wave form of the mains supply current passing through the MCB

Well, yes - I don't think anyone has suggested (or could sensibly suggest) that such is not the case. Qualitatively, it is inevitable, uncertainties being about the magnitude of reactance/impedance of the MCB, information which is not directly available from what you have done.

If you used the Wheatstone Bridge as a Wheatstone Bridge (ideally with inductors {one variable}, rather than resistors - but that would probably be impractical) (or, indeed, simply put the MCB in series with a known inductance and measured the voltage across each when AC of an appropriate frequency was flowing through them both), you could presumably get a quantitative handle on the reactance/inductance of the MCB, and hence the impedance at any particular frequency.

Kind Regards, John

 

[bernardgreen]

I didn't record the results, it was a lash up to get a rough measurement of the effect. It was obvious when applying high frequency sine and impulse waveforms that the MCB had a complex impedance due to the internal construction. An MCB I opened a few weeks ago had two coils in series, one magnetic and one the wound heater. Proximity of components created several distributed capacitive components.

I can recall several cases where PowerLine units ( data over the mains ) would only work when both were on the same MCB. They wopuld not work when on different MCBs. ( they should of course be banned as they radiate RF energy from the mains wiring )

 

[JohnW2]

It was obvious when applying high frequency sine and impulse waveforms that the MCB had a complex impedance due to the internal construction.

Indeed, but I would imagine that is what we all assumed and expected, anyway - how could it possibly be otherwise?

I can recall several cases where PowerLine units ( data over the mains ) would only work when both were on the same MCB.

Yes, I gather that's quite common. However, we are then talking about frequencies which are several orders of magnitude greater than those use in your experiments (and quite probably orders of magnitude greater than significant components of the 'noise' created by electronic devices) - so even tiny bits of inductive reactance presumably could have a major impact.

Kind Regards, John