UPS and high earth loop impedance

[Ahr35181]

I would be interested to know how others handle the high earth loop impedance which is generally measured on UPS circuits.

The contractor who tested our office told me that they do not normally do EFLI tests on UPS circuits for fear of damaging said device. I insisted they were done.

In our office disconnection times would not be met on circuits fed by the UPS (15kVA unit). I have had a 300mA RCD installed on the output of the UPS and to date fortunately had no trips.

 

[rbulsara]

Pardon my ignorance, I am in the USA.

Explain high earth loop impedance. What loop?

What is RCD (leakage/residual current detector). Does it use a window type CT surrounding all current carrying condcutors?

It appears you are talking about ungrounded UPS? Is this a single phase UPS? Does it have a bypass circuit?

 

[ScottyUK]

Ahr35181,

This is a common problem with relatively small, high-impedance sources such as your UPS when used with normal distribution MCBs. The UPS will provide a fairly high source impedance when compared to that available from a typical utility connection. This high impedance will prevent a large fault current from flowing, with consequent slow response of the MCB.

In the UK it is required that socket outlets be disconnected in 0.4s in a fault condition, and the fault level available from a small UPS makes this condition difficult to meet. The use of an RCD is a reasonable solution to the problem. You could consider discrimination using 30mA/instantaneous RCDs on final circuits, with a 300mA/time-delayed RCD on the UPS output.

Rbulsara,

I'm second-guessing here, but I expect the UPS neutral is grounded, and the loop in question is the path from UPS 'live' to the point of test and returning via the earth conductor to the UPS.

RCD is an abbreviation for Residual Current Device. We used to call them Earth Leakage Circuit Breakers or ELCBs in the past. RCDs are differential devices, working on the principle that line and neutral currents should have a vector sum of zero; a non-zero sum is treated as a fault and the device trips if the imbalance current exceeds a threshold. A large RCD will probably use a window CT as you describe, although it may well be hidden within the device casing.

 

[rbulsara]

Thanks, scotty.

 

[peebee]

Hmmm. . . .

I'm more familiar with large UPS systems than small ones so this might not apply. But most of the UPS systems I'm familiar with would transfer to static bypass in the event of a fault, they need to in order to supply the fault current.

Now, maybe your UPS systems don't have bypass (likely), or maybe the static bypass system is not fast enough (unlikely).

But if they don't have static bypass, wouldn't they need to shut themselves down to protect themselves against the severe undervoltage and overcurrent?

Scotty -- thanks for the English/English translation