Advice needed

[foxc01]

Hi guys,

I've just started working in a commercial premises as a technical advisor.

The building has 2 consumer units - one older MEM board and a newer Proteus board, which has a 61008 RCCB protecting ALL of the circuits on this board

My problem is that I've arrived at work this morning, and the RCD has tripped overnight, causing a nightmare with our computer server and systems.

As nobody is ever in the building and all of the computers are switched off when the RCD trips, I can't see it being an overload problem.

I'm going to carry out an RCD test after work, but my assumption is that the RCD may be faulty. Can anyone advise?

Thanks - Chris

 

[ProDave]

Cleaners plugged a faulty hoover in?

Wiring fault?

Start with EVERYTHING unlugged from that circuit and a L-N and L-E insulation test on that circuit.

Surely the servers are still plugged in and turned on overnight?

 

[apprentice87]

Hi, go here and download the "beama Handbook" :Circuit Protection Catalogue,Brochure,Price List and Technical Paper Downloads

Read through this; Everything you need to help you is in here. The only problem you might have, is the cause of the tripping might well be EXTERNAL to your installation.....

Post back if you have more problems, someone will be able to help you sort this.

An rcd by the way, does not protect against overload, merely earth leakage. You may well have to make special arrangements for computers, as these things tend to have problems with high levels of earth leakage anyway....Could well be the wrong type of rcd for the job.

Is it wise to have ALL the computers fed from the one rcd protected circuit?? Could you not split it up into several circuits all fed from an rcbo??

How are you proposing to test the RCD Have you got the neccessary test equipment??

Hope all this gives you a few ideas!!

Other forum members will be able to help a lot more though!!!

John...

 

[foxc01]

Hi Dave - The servers are indeed turned on overnight.

My initial thought is that perhaps the servers are leaking an earth current which is tripping the RCD, as none of the 60898 MCB's on that board aren't tripping.

The wiring and consumer unit are new within the last 2 years, and I've seen the installation certificate this morning and the insulation resistance results on the certificate are well above 200 mega ohms.

---------- Post Auto-Merged at 11:31 ---------- Previous post was made at 11:22 ----------

John - thanks. I didn't wire the installation, there are a few different power circuits supplying the computers, but they are all protected by one RCD, which is the problem.

I've recommended that going forward all the circuits are changed to RCBO's to irradicate nuisance tripping.

I've just changed from being a apprentice trained electrician fitting domestic bathrooms for 3 years to a technical advisor, so I'm now working for this company that are having the problems!

I'll test the RCD later so that I can see if it's tripping at 0.5x, and then go from there. If it's OK, I'll start insulation tests.

Thanks guys

---------- Post Auto-Merged at 11:31 ---------- Previous post was made at 11:31 ----------

John - thanks. I didn't wire the installation, there are a few different power circuits supplying the computers, but they are all protected by one RCD, which is the problem.

I've recommended that going forward all the circuits are changed to RCBO's to irradicate nuisance tripping.

I've just changed from being a apprentice trained electrician fitting domestic bathrooms for 3 years to a technical advisor, so I'm now working for this company that are having the problems!

I'll test the RCD later so that I can see if it's tripping at 0.5x, and then go from there. If it's OK, I'll start insulation tests.

Thanks guys

 

[S60TEM]

what earthing arrangement is it ie: TT with a 100mA RCD upfront ?

How many servers are running and how many PC's etc etc?

As you've said, I'd be tempted to replace the mcb's with rcbo's for the IT circuits aslong as they don't have more than 6 PC's on each circuit. Are the circuits solely for IT use and do they have High Integrity earthing too?

A lot of things to look into really.

 

[kme]

Computer power supplies are designed to leak a little - get too many of `em, and your quiescent background leakage may be approaching the trip point of the RCD. Not really a good idea to have a front-end RCD in this situation - in order to minimise (rather than eradicate) fault tripping; the RCBos are a good way to go.

Maybe you have seen the EIC - but the RCD wasn`t tripping then! It is now; so you need to do an insul test now. Similarly, you should be doing the full raft of RCD tests (including a ramp test), to determine the state of the RCD.

We`ve said it before, but its worth repeating......If you haven`t found the fault yet, you haven`t done enough testing.

KME

 

[Evans Electric]

1) Need to know why the there is an RCD controlling the Consumer unit first .

2) Test RCD ,

3) As KME says computers leak to earth and (apparently) dump data down the earth .

4) What is the rating of the RCD (not the load, the tripping current )

5) Technical Advisor Eh ! That sounds....... er....well.....technical !!

6) Also Computers need back ups.

 

[VoltzElectrical]

as evans said in point 6, although additional to your first query but a critical observation for you in your role as technical advisor. UPS should definitely be a consideration.

 

[foxc01]

Hi - thanks! UPS already backs up the server, but according to IT technician, the battery backup is very short....

Its a 61008 63A 30mA RCD protecting one of the consumer units that is at fault. My tests concluded that I had 18mA of earth leakage from all the various computers/equipment - Proteus state that their 30mA RCD's can trip from 17mA to 25mA. Also the RCD trips on 0.5x its rating, and at over 400ms for 1x.

I've advised that the RCD is replaced immediately and that an RCBO setup for each circuit is the way forward.

Thanks for everyones help, please understand commercial installations are a bit alien to me at the moment with only having domestic experience.

 

[Sidewinder]

Tell the IT tech they need to install a monitoring system for the battery UPS that instigates a controlled server shut down in the event of a power failure and that is batteries must last long enough to implement the controlled shut down without data loss and that the software he uses should ensure that there is no data loss.

If he can't tell him you know someone who can! ]

 

[Manator]

Tell the IT tech they need to install a monitoring system for the battery UPS that instigates a controlled server shut down in the event of a power failure and that is batteries must last long enough to implement the controlled shut down without data loss and that the software he uses should ensure that there is no data loss.If he can't tell him you know someone who can! ]

Very good point, there can be some very sensitive data kept, and should be looked after properly

 

[batty]

Sounds to me that its a poorly designed system. The installers weren't NICEIC by any chance.

 

[Evans Electric]

As the installation is commercial and under the control of a Technical Advisor why don't you dump the RCD for a main switch . Its a liability in your case. What is the Zs ?

Assuming your supply is not a TT.

 

[S60TEM]

As the installation is commercial and under the control of a Technical Advisor why don't you dump the RCD for a main switch . Its a liability in your case. What is the Zs ?Assuming your supply is not a TT.

exactly my question in post number 5.

If it's a TT system then it SHOULD have a 100mA upfront RCD (good practice) NOT a 30mA!

As said, as it's commercial I would ditch the upfront RCD in favour of a BS EN 60947-3 100A DP isolator and replace the mcb's for rcbo's for the applicable circuits.

 

[foxc01]

Hi Guys - its not a TT system, so no 100mA RCD.

My guess is that as the (NICEIC) installer was installing new circuits, he wanted to comply with the 17th edition; but obviously he's not thought about the installation as a whole.....

 

[PC Electrics]

Very poorly designed system by the sounds of it. Now, that might be the fault of the installers, or it might be the fault of the company who may have set an unrealistically low budget for the job. If the latter, well, they made their bed........etc. The use of a Proteus board suggests to me that someone was more interested in cost than quality.

The simplest solution would be to change over to RCBOs. Are any of the circuits feeding PCs ring finals? Might be worth splitting them into 2x 20A radials to lighten the leakage load on the RCBOs.

If you've got to buy a load of RCBOs, it might be worth changing the board. A Hager (from a decent wholesaler) with RCBOs might well work out cheaper than a load of Proteus RCBOs from Citys anyway.

 

[Manator]

I had to install a Proteus DB today, my usual boards out of stock, nasty little things they are, but I was allowed 1 hour to do the change so just put it in.

 

[Manator]

The screws on their RCBO's are awfull, just to add insult to the injury of installing it.

 

[ProDave]

In defence of Proteus, I was rewiring a garage yesterday with a small proteus CU

The one thing I liked about it was the whole top panel just slid out, making the task of making an extra knock out much easier.

So they may be rubbish in other respects, but I thought that was a nice feature that I've not seen in any good make of CU before.

 

[PC Electrics]

You should try Hager, Dave.