Anyone else done this?

[Bizron]

Hi,

Got distracted doing an EICR as tenant's were chatting and security Alarm on RCD 2 was going off as I had opened all mcbs and RCD2.

I linked out Rcd1 busbar to neutral to do IR test and hadn't opened the main Switch. All breakers were open but RCD was still closed - Lots of sparks and flashing. Lost my jump leads and part of the Busbar. Amazing amount of power involved in such an event. Anyone else done similar.

Here's my question why didn't the RCD actuate, main Switch open or cut out fuse blow?

Cheers

 

[SparkyChick]

The RCD didn't activate because there was no fault it was capable of detecting... the absurdly high currents that were flowing during the fault flowed in the line and the neutral conductors, so it kept supplying power.

An RCD will only trip when there is an imbalance between the current in the live and neutral conductors.

The main switch... if it's 60947-3 (I think that's the correct standard) is not capable of detecting any fault... it is simply a switch. And you test leads can't carry as much current as the main fuse which is why they got vaporised and the main fuse didn't.

 

[SparkyChick]

I have done something similar... I have a client with a double row MK board (older metal casing with plastic front face) and I was just checking tightness on the major connections... not realising it wasn't physically attached to the busbar, I gave the line conductor terminal that fed the second row a tweak.

As I say, I wasn't aware it wasn't physically part of the bus bar and was in fact just clamped in the cage clamp of the main switch... it twisted and shorted out on the neutral bar to it's left... I clearly wasn't the only one to make this mistake as I had already observed arc splash damage on the neutral bar but I'd failed to put 2 and 2 together.

It's never good when that kind of thing happens.

 

[loz2754]

The RCD didn't actuate because it is not designed to detect idiotic stupidity!

(I'm not saying I've never done anything similar)

 

[timhoward]

The Metrel tester uses the blue and brown plugs for IR tests and the plugs are stackable which can be really handy, and also a bit dangerous if you don't pay attention to how you are using them...!

I'd been trying out a different way to do L+N to E IR tests, by hooking up an extra loose lead to the blue clip and stack it on the brown tester lead and brown clip, with the green clip on the blue tester lead. I'd thought this was reasonably fool proof as the clip colours reflected where everything was going. Before I start, I check L to E and N to E both give zero. All fine, so I did the tests.
Then I swapped clips for probes to test at the top of a DP RCBO and did that.
Then there were a couple of distractions, a phone call with a customer arguing about an invoice followed by tea and jam doughnut arriving (happy days!)

Then I picked up tester, didn't engage brain at all, and tried to do an RCD test without remembering the blue and brown probes weren't where the manufacturer intended and were actually directly joined together! As I pressed hard into the terminals to ensure a good connection there were some exciting sparks, the end of one probe nicely deformed, and a blackened terminal screw....I felt a right muppet!
I went back to using R1+R2 clips after that fun experience.

 

[littlespark]

And this is why we shouldn’t be IR testing inside a live board.

 

[pc1966]

I always would suggest linking the output of a small MCB to N for L+N testing due to the smaller bang when it goes wrong.

Yes, we always should do the full proving dead test before touching anything, but there is still the risk that you get distracted and restore power before removing the link...

 

[pc1966]

Also to expand a bit on what @SparkyChick said about RCDs and interrupting the power:

• They don't really care about the magnitude of an L-N fault, as you found out the hard way!
• Even for a L-E fault they don't really have any energy-limiting in the strict sense like a fuse or MCB has, they could take a couple of cycles of 50Hz to open and still be in specification (40ms).
• They often have a limited break capacity, typically in the 1-1.5kA region, which is often below the PFC/PSCC of the supply so for their own safe use they need to be partnered with a MCB or similar that will start to open before they do and thus deal with the majority of the fault energy. Worst-case even a 50A D-curve MCB will have hit the "instant" magnetic trip point by 1kA fault level and so be opening faster than a RCD. For info, MCBs have arc-cutes inside for energy dissipation while HRC fuses have a sand filling to absorb the energy.