.75mm flex used from junction box to downlights

[Wilko]

Hi - agree Reg 434.5.2 is important here.
So for fun I apply some values to the equation given, being k=115 (table 43.1), S = 0.75mm (drop flex conductor being discussed) and picking current as 160A, I get t = 0.29 seconds.
From Figure 3A5 we know that a 16A C type MCB will trip in less than 0.1 seconds when passing 160A.
So I think the 0.75 mm conductor is potentially up to the task of handling the fault clearance current without heat damage.

 

[Lucien Nunes]

Agreed. The crux of this is that the load on each 0.75mm² flex is well below its rating, regardless of whether the transformer is wirewound or electronic, and of whether the lamp is LED or filament. If any of the above were to pass a current sufficient to overload the cable then it will self-destruct rapidly or instantly. Therefore, it can be seen as a fixed load in all cases, and only fault protection is needed. According to Wilko's numbers, that is OK on a B6, how about 10A OCPD?

TBH, I would be more concerned to find out whether the flexes have been correctly installed at the junction boxes. I have no issues with flex having been used per se, but the boxes must be suitable for it. If they are not, and the connections may be poor or subject to stress. In any event the cable, flex or twin+earth, should be supported clear of the luminaires.

 

[Shoei]

No said anything about compromising fault protection.
Unless the 0.75 flex is too small regarding the adiabatic equation for the csa of the cpc and is too small for the thermal constraints of the live conductors during fault conditions, then there’s not an issue there.( protection by a class 3 current limiting mcb say 6 amp type B will likely satisfy this )
No one said lump it on a 16 amp mcb that point was brought up with the fact that a lamp holder can have omission of overload protection.

Regarding the use of the cable, my only involvement of this thread was in response to another member about whether all aspects of the 0.75 flex required overload protection or not.

Thanks for the clarification Ian, agree with your points about thermal constraints and that a B6 MCB would be fine when 0.75mm flex is in circuit.

No.
The regulations allow omission of overload protection as detailed in section 433 for certain applications.
After all for example, a pendant lamp holder does not require overload protection and is allowed to be protected by an ocpd of up to 16 amp( see section 559)

I took this to imply that a pendant lampholder can be installed without overload protection. Maybe I read this the wrong way,

The problem with 559.5.1.204, is that the only 16A overcurrent protective devices that would satisfy disconnection times would be MCBs or RCBOs.
As such, I don’t quite understand how it is determined that a 0.75mm2 conductor on a 16A MCB would not comply with 434.5.2?

559.5.1.204 doesn't recommend a 16A MCB, it sets the maximum limit when any of the 5 lampholders listed are used. To comply, the conductors should be sized to withstand any prospective fault current/disconnection time.

If a fault caused (say) 70A to flow in the 0.75, using a start temperature of 70 degrees and table 43.1, the limit temperature is reached in 1.5s. A compliant B16 can allow 70A to flow for over 10s (Fig 3A4) so the design fails 434.5.2, it needs more copper or a smaller/faster MCB/RCBO.

Referring back to the OP's question I think he needs to check for a 6A protective device on his 0.75mm connected circuit, otherwise the design needs to be reviewed.

 

[Wilko]

Thing is, I would say you can't get a fault of 70A constant, that would be a load. A fault should cause current to flow limited by the impedance at that point. Otherwise you could make this argument for all conductors being affected by "excessive currents that are almost faults". Just my thought .

 

[Ian1981]

A lamp holder supplying one lamp is not going to overload period.
If overload protection was required for every single pendant set and manufacturers live and neutral supplies to led drivers and the like then nothing would comply

 

[Shoei]

Thing is, I would say you can't get a fault of 70A constant, that would be a load. A fault should cause current to flow limited by the impedance at that point. Otherwise you could make this argument for all conductors being affected by "excessive currents that are almost faults". Just my thought .

Fair point Wilko, all I can say is that faults don't always follow the script and I've seen this happen. For example, a mains LED lamp contains a switch mode current source. If a transient turns the electronics into a piece of slag it's anyone's guess what current will flow. I would indeed argue that all wiring should be sized so that it can't start a fire at any current level/trip time in the range of its OCPD.

A lamp holder supplying one lamp is not going to overload period.
If overload protection was required for every single pendant set and manufacturers live and neutral supplies to led drivers and the like then nothing would comply

We'll have to agree to disagree Ian, my understanding is that the type of load is irrelevant. The regs are designed to prevent fires by holding temperature rise in wiring to safe limits in any single fault condition. The core principle is that the protective device must disconnect before any downstream conductor overheats, even on lighting circuits.

 

[Lucien Nunes]

If a transient turns the electronics into a piece of slag it's anyone's guess what current will flow

SMPSUs and transformers (unless they are non-compliant junk) include internal protection such as a fuse, thermal fuse or fusible resistor to prevent this scenario, which qualifies them as fixed loads. Even if that protective measure fails, one has to consider what will happen if this arbitrary current does flow.

To cause 0.75mm² to get dangerously hot in less than a minute or two would need maybe 15A or more. At that current, the faulty 'load' is dissipating 15 x 230 = 3.5kW, so the 'load' will be on fire long before the cable. There is no realistic situation where a small electronic device will dissipate enough power for long enough to damage the cable, without being destroyed itself and either going open-circuit and interrupting the current, or short-circuit and tripping the OCPD.

 

[Shoei]

SMPSUs and transformers (unless they are non-compliant junk) include internal protection such as a fuse, thermal fuse or fusible resistor to prevent this scenario, which qualifies them as fixed loads. Even if that protective measure fails, one has to consider what will happen if this arbitrary current does flow.

To cause 0.75mm² to get dangerously hot in less than a minute or two would need maybe 15A or more. At that current, the faulty 'load' is dissipating 15 x 230 = 3.5kW, so the 'load' will be on fire long before the cable. There is no realistic situation where a small electronic device will dissipate enough power for long enough to damage the cable, without being destroyed itself and either going open-circuit and interrupting the current, or short-circuit and tripping the OCPD.

I understand the logic Lucien, but the regs boil it down to straightforward calculation, they reasonably put the onus for safety on the fixed installation because that's what the designer controls. There is an awful lot of non-compliant junk about and common faults which don't involve the load at all.

Your scenario assumes the fault resistance is quite large, the concern is low-impedance faults where most or all of the power is dumped in the fixed wiring. For compliance the wiring must be sized to remain below the specified limit temperature under any fault current/power that the OCPD can deliver before tripping. 434.5.2 is unambiguous about routes to compliance, it doesn't make an exception for lighting circuits and it's easy to work out. I'm a bit surprised that anyone would consider deviating from it without a serious risk analysis that couldn't be worth the trouble to validate in any normal installation.

 

[Lucien Nunes]

In the case of

low-impedance faults where most or all of the power is dumped in the fixed wiring.

the impedance of the defective load must be much lower than that of the circuit, in which case the current will approximate to the PSCC and will trigger the fault protection. If the impedance of the defective load is high enough to significantly reduce the current below this value, it will dissipate enough power to self-destruct.

But I agree; my statement 'only fault protection is needed' should have been qualified with 'for safety' as I'm not claiming it to be explicitly compliant, just practically safe.

Historically, though, there was an exemption for the wiring of light fittings specifically from overload protection, subject to being of a minimum size. There was a standard conductor known as a 'fittings wire' that complied. I can't recall when that was revoked, I should read up.

 

[Shoei]

Thanks for the clarification Lucien, I agree it would be bad luck to start a fire with a B16 MCB on a 0.75 mm wire and mostly it would be fine, but if enough people do it often enough then someone will have a bad experience.