PC SMPS leakage?

Talk Electrician Forum

Help Support Talk Electrician Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

NozSpark

Seaman stains™
Supporting Member
Joined
May 28, 2008
Messages
12,330
Reaction score
141
Location
North Wales Riviera
I know this has probably been answered before, but I'm currently on a ferry coming back from Ireland (visiting inlaws) and the internet connection is really slow,,, I've also not had any internet all weekend:( :( :( :(

I'm going to be wiring my neighbours computer workshop for him in the next week or so...

The house supply is TNCS so I'm running the workshop as TT, he's going to have a server rack, possibly 10 PC's, a freezer, a tumble dryer, lighting and some sort of heating (just thought about that one: O) . I'm going to run a dedicated circuit for the server with other circuits for the rest of stuff.

Now the question is how much do PC SMPS's typically leak (same question for server rack) as I'll need to decide how many circuits he'll need and if high integrity earthing is required as per BRB 543.7????

I'll join back in tonight when I get home as the connection speed is painful at the moment.

 
I always wire servers with a dedicated circuit, the tower servers I have wired up come complete with a bank of socket outlets for all the hardware to plug into. A dedicated earth from the MET is also a good idea. I am not really sure of the leakage levels, I have never measured them on any server racks, but I have never had any problems. The server racks I have seen have an earth connection, and sometimes a couple of strappers for interconnection.

 
Noz,

I would suggest separate circuits for the server & the pc's IF possible with high integrity earthing on RCBO's.

If you are going to TT it put a 100mA S type up front.

Else correctly extend the equipotential zone.

Just IMHO!

 
I'd be tempted to look at making it TNCS as well

How far from the house is it? What is it made of? concrete/metal/wood etc?

Any metallic services?

How likely is it that folk are going to be in contact with class 1 equipment and not very well insulated from physical earth?

For example, I have a concrete workshop, its only about a metre and half away from house (footpath in between them), No water or gas in there. I've done it as exported TNCS, logic being that if instead of a workshop, it was an extra room extension on the house (taking up the footpath as well) it would be TNCS because you would extend the existing circuits in that situation. Workshops have a hard concrete floor and are entered wearing workboots, not a muddy floor entered while wearing sandles as would a greenhouse be (which I would make TT).

As to circuits, I'd go for one for the server, a general one for the freezer and whatever else and two to the computer test bench, as sidewinder says RCBO all socket circuits. But as wiring is probably surface, I wouldn't bother with RCD for the lights. High integrity earthing to server and test bench sockets. for the test bench if you have two radials then its permitted to add a section of eatrh between them so that the earths are effectivly a ring, and then terminate in separate terminals in the sockets as normal and that meets the requirements for HIE rather than pulling in two separate earths back to the board. Or if seeing as its a workshop, you elect to wire in metallic conduit/trunk then don't forget that this can act as your HIE

 
I'm now back home after setting out at 10am, poping out to collect dog from my parents and a few other bits and pieces.... Long day travelling

The workshop/store room/utility room is block builtnon a concrete base and is probably only about a meter away from the house and there are no extraneous services etc. Extending the TNCS equpotential zone would be best as I can then do away with the RCD for the server (dedicated socket outlet;)) I do tend to TT all TNCS sub DBs in out buildings, all power cabling and network cabling is being run in 3 compartment daido. I guess that I'll have to run in a general socket radial and a couple of radials for the PC's with high integrity earthing

 
So, I've done my demand and cable calcs and while 6mm SWA would be big enough for load, Zs and volt drop,,,, I'm thinking that I'll go with 10mm 3 core with one core and armour as earth to make sure that the earth is big enough for TNCS..

What do you think guys?

 
As you've got no control over the use of the sockets I think I would tend to run the radials with HI earthing for all circuits, that way there are no worries?

 
The building is going to have 2 rooms,,, his main workshop and a server/store room which will also house a freezer and tumble dryer, which may or may not stay there - I'm assuming that it will though.

Does my reasoning for going to 10mm on the sub main sound ok?

 
I'd have to check GN1 first.

I think I might be tempted to put a BEMT "in" though.

As you have no extraneous or exposed conductive parts, nor anything bringing an earth potential in then probably OK.

---------- Post Auto-Merged at 16:02 ---------- Previous post was made at 15:13 ----------

OK on 3C 10mm sq IMHO even for table 54.7.

Ref: GN1 table D7, D.8, D.9, D.10a & D10b

I would however, ensure that the connection between the earth paths both ends is excellent!

Tails from banjo's etc. all earths terminating at the same earth bar at both ends, and this being the main earth bar at both ends, i.e. a BEMT.

I would also use non conducting accessories & containment if possible, however, if you need metal clad accessories then perhaps you need to give it a little more thought.

 
Not that I've thought about it too much, but I was going to use "pirranha" nuts at each gland, using plastic/pvc enclosures with the banjo on the outside so it doesn't pull through. Like I said before, he has now got his 3 compartment trunking and all sockets will be in that,,, except for the store sockets which will be surface mounted and probably fed by mini trunking... The only other thing of note is that he has got 3 x 5' flours for his lighting... he originally wanted recessed downlights (with led's).... Most of the lighting cables will be behind plasterboard!

 
OK sounds OK, just remember the migration of polymer enclosures under compressive forces under glands allowing them to come loose.

I like "Pirhana" nuts, just don't skimp on the bonding tails.

You don't really need the banjo if you are using the earthing nuts.

A washer would be oK, also you are then not drilling an additional hole in the enclosure for the brass banjo bolt, possibly reducing the IP in the process.

 
The reason I was going to use the banjo as a washer is because they are that little bit thinner and just using the gland "bolt" shoulder can leave only a very small edge on the PVC box, I wasn't going to necessarily bolt through though;)

I'll probably lug some 10mm earth for the bond

 
I wouldnt get too hung up on worrying about plastic accessories.

This is a workshop which is 1m away from the house. Lets suppose he built it 1m longer so it touched the house? still worried about TNC-S?, what if he knocked through and the door was off the house rather than the garden?. How about now? What if he put a bed in it instead of a workbench and called it a bedroom... is he alright to have brushed stainless sockets in that room to match the decor?

Not being arkward, I'm just trying to point out that electrically there is very little difference between using the TNC-S earth here, and using in a extension to the house..... now if he was running to a green house, or he was running a temp down the garden to a marguee for an event and there would be people possibly barefoot outside and in contact with metal water heating appliances, etc then we'd probably be right to be concerned, or if he was supplying a metal framed caravan/site hut.

I think with this TNC-S / TT thing, you get two main camps of people... those who will use TNCS regardless and don't understand the issues it can create, and those who are often overly cautious and perfer to err on side of safety and will TT nearly everytime. I'm in favor of an approach that asks if the risk from TNCS is any greater than it is in the main installation and comming up with an engineered approach

Like Sidewinder... I also don't like SWA into plastic enclsoures where it can be avoided

 
Top