IT system - how do RCD's ever work

[Simon47]

No TN.. and TT are defined as being earthed directly, whilst IT is not directly connected, the principle being the impedance to earth being substantially high and very reduced fault currents - much less than typical full loads.

OK then, where does the line lie between "IT with a highish impedance to earth" become "TN or TT with a highish impedance to earth" ? There are the extremes - something resembling a "very low" impedance is TN or TT, something resembling a "very high" impedance is IT. When "very low" changes to just "low", or "very high" changes to just "high" - where is the threshold ? Somewhere in between those extremes you will change from one to the other.

No, a dangerous situation can arise due to the single fault, as this basically becomes similar to TT , one outgoing circuit can have a fault - effectively connecting L1 to earth, thus every other circuit (on other phases) has a potential of 400V to earth rather than 230V

We were talking about a two wire 240V IT system. Earthing L1 or L2 will only create 240V between the other line and earth.

insulation failure in contact with a seperate earth (effectively different from the first earth) means that the "earthed" body on circuit 1 has a difference in potential to the "earthed" body on the other circuit.

That's not unique to IT systems - and why we have bonding to avoid having different "earth" potentials within the bonded zone.

 

[timhoward]

OK then, where does the line lie between "IT with a highish impedance to earth" become "TN or TT with a highish impedance to earth" ? There are the extremes - something resembling a "very low" impedance is TN or TT, something resembling a "very high" impedance is IT. When "very low" changes to just "low", or "very high" changes to just "high" - where is the threshold ? Somewhere in between those extremes you will change from one to the other.

Isn't the distinction high impedance by deliberate design (IT) versus high impedance through natural environmental circumstances (TT)?

 

[Simon47]

But still, at some point you have to decide what protective/control measures are needed. Then it doesn't rally matter whether it's deliberate or imposed ...

 

[Julie.]

But still, at some point you have to decide what protective/control measures are needed. Then it doesn't rally matter whether it's deliberate or imposed ...

Well it does, if you have an earth loop impedance of 5k ohm, that tends to be a fault condition on a TT and TN system, which needs to be fixed so the usual protection (std RCD and OCPD) will work, whilst that would likely be somewhat low - a potential fault meaning a first fault on an IT system.

The distinction is clear in the regs, TN & TT systems are directly earthed implying similar fault levels to short circuit faults or at the very least similar fault currents to flc.

IT on the other hand are not directly connected, and therefore have fault currents significantly lower than flc.

If the impedance is between these, it represents an unexpected situation outside of the design of either system.

 

[pc1966]

The distinction is clear in the regs, TN & TT systems are directly earthed implying similar fault levels to short circuit faults or at the very least similar fault currents to flc.

IT on the other hand are not directly connected, and therefore have fault currents significantly lower than flc.

That is not really the case though as even a domestic TT system can have currents to tens of amps on each final circuit, but a 200 ohm rod so PFC below 1A. I guess the defining point would be not the FLC but the operating point of the designed protection system (so in such a TT case the RCD would be tripping typically for faults below 7.6k ohm at most).

Which kind of returns to the original discussion about IT systems and that they are normally designed to continue under single fault conditions, and ideally safely (as far as practical) so you might be looking at earthing impedances of tens of k ohm if trying to keep the 2nd fault touch shock current below, say, 10mA (assuming that is even possible depending on filter caps, etc).

 

[Julie.]

That is not really the case though as even a domestic TT system can have currents to tens of amps on each final circuit, but a 200 ohm rod so PFC below 1A. I guess the defining point would be not the FLC but the operating point of the designed protection system (so in such a TT case the RCD would be tripping typically for faults below 7.6k ohm at most).

Which kind of returns to the original discussion about IT systems and that they are normally designed to continue under single fault conditions, and ideally safely (as far as practical) so you might be looking at earthing impedances of tens of k ohm if trying to keep the 2nd fault touch shock current below, say, 10mA (assuming that is even possible depending on filter caps, etc).

Yeah, I am not sure how to word it though, TT systems still have a decent amount of current in the order of amps (minimum should be above 1A ) whilst IT systems are usually 2 or more orders of magnitude smaller than this - usually 10mA or less.

What I am trying to say, is that an earth fault on a TT OR TN system is a reasonable level to detect - you could use a 100mA rcd and still have fault current in excess of 10x the trip setting, these are easy and almost foolproof systems to be able to detect an earth fault.

This is a vastly different situation to an IT system where the current, on first fault will be in the order of mA or less.

There isn't really a cross over where you might have an IT system having a first fault current of 100mA, or a TT system of also 100mA (200 ohm is the expected minimum).

Detection of first faults is likely to have to be much less than 1mA unless you are forcing a first trip situation by upping this to 10mA or so.

 

[Simon47]

Where are these dangerous touch voltages coming drom

... if trying to keep the 2nd fault touch shock current below, say, 10mA

That implies a situation where at least some items are neither Class II nor earthed/bonded - which is not generally permitted for obvious reasons.
Yes, there will some cases where the IT system is there to directly control touch currents - shaver sockets come to mind. But in the context of the OP's IT supply to a distribution panel, all the normal protective measures (principally bonding via CPCs) are (or should be) there to control touch voltages/currents.

 

[pc1966]

My thoust was as much a "if you need to chose a single-fault threshold" sort of point than seeing big IT supplies as touch-safe.