Connection to earth of computer rack drawers

Assuming the safety of the devices is not in question (each device having its own cpc), we are really only talking about EMC measures (as the parts in question are typically neither extraneous-conductive-parts not exposed-conductive-parts).


The standards in question (referenced from Section 444 of BS 7671) are:

• BS IEC 61000-5-2

• BS EN 50310

First "curly" or "not curly" - usually the recommendation is generally NOT curly, as the inductance is increased by both the length of the lead, and by the action of providing a coil. The object of earthing for EMC would be to conduct higher frequency currents away from the affected part (whether it's shielding radiated emissions or sinking conducted emissions).


Second, is the bonding to these parts actually required or not? Well, that depends. Reasons to do so include:

• Rack contains measurement instruments sensitive to high frequencies and radiated / conducted emissions.

• Rack contains a mixture of equipment with industrial emissions (or Class A for some product standards), and domestic/light industry immunity, and shelves/trays are used as part of the mitigation measures in the EMC control plan (although, there are orders of magnitude difference between the light industry immunity and the Class A emissions, so this is not always necessary)

• Someone is worried about mains cables breaking, and making contact, thus it's considered in that risk assessment as an exposed-conductive-part.

A way of avoiding lots of bonding conductors in a rack is to use internals made of substantially conductive materials that are protected against corrosion, such as zinc-plate colour-passivated, meaning each shelf fixing is electrically connected to the support rails, front panels etc. - in fact, some ranges of 19 " rack are specifically designed with this in mind, such as the Cooper-Bussmann EMC range. Even then, one could argue that the earthing of a pull-out drawer or shelf was only "fortuitous", but so long as it's not for safety (pull-out shelf or drawer is not an exposed-conductive-part), and we're not talking about SIL-level functional safety risks from EMC, then that shouldn't be too much of a problem.


The latest edition of BS EN 50310 goes down to rack earthing and bonding level, and is an interesting read ... but doesn't expressly contain the bullet list above.

+1 for avoiding curly pig-tails. If you did need something then a flat braid is usually a better choice.


That said I'm not convinced of the need for anything. If it's for shock protection (because someone thinks the shelf could be an exposed-conductive-part if a flex gets trapped in the sliding mechanism somewhere) then there are probably better solutions - e.g. cable management arms that concertina down when the shelf is pushed home) to stop flexes getting trapped in the first place.


If it's for EMI I doubt the curly tails will have much effect - there will almost certainly already be good contact between the shelf and the main body of the rack - those slide rails are usually full of steel ball bearings and so have multiple points of contact despite any grease etc. Besides the earthed outer shell of the rack provides the main 'Faraday cage' all the shelf could do at best would be to reduce the interference between one item of equipment and the item below it- and if they were that sensitive (or noisy) I've expect them to have their own earthed enclosures anyway.


  - Andy.

Well,  thank you both for the replies.  In our case, we are first institutionally so risk-averse that you have to wonder how anything ever gets done , so the safety aspect is at the heart of the enquiry.  Nobody wants to be found omitting anything.  But then, on a much more interesting level;   Graham's reason 'for' applies - the rack contains measurement instruments sensitive to high frequencies and radiated / conducted emissions.  Because we are a pulsed power operation, you can imagine what happens for that critical nano second (or micro second for the slow stuff) and it is happening to our measurement devices just at the point when we need the information. There is a noise issue.


I have a confession to make - this is on a par with the day Zs admitted that she liked a range of LED lights and Andy J posted 'who are you and what have you done with Zs?' ...

I obtained a copy of BSIEC 61000-5-2 following your post Graham and have spent quite a bit of time reading it.  What an excellent standard it is.  Clearly written by a knowledgeable enthusiast and such an education.  It's like reading a tutorial not a dictatorial. Thank you for that.  I reckon I should read it in secret and just pretend I knew all that ;).   Complete with pictures of the effect of curly tails and I'd never before have given a thought to the use of round or flat braids - and so on.


Our signals largely get back to the scopes via fibre optics by the way.


Andy, you are right about the ball bearings and you may well be right about not dealing with the EMC.  There are already various layouts of ferrite rings around the place and even a huge plastic barrel, wrapped with about 100m of flex for insertion into the circuits by way of mitigation.  That last one might seem a bit Heath Robinson but it works, after a fashion.


Thank you, I now have to make a decision on what they should do with the racks and offer it forward.  


Zs

 

Pulses that are short compared to the time it takes for the signal to get along the wire are always a problem, and earthing in such a case is largely nugatory. And one way the common mode choke, (for that is the function of the  barrel of cable) works, is by allowing the signal so long to get used to being between the conductors of the cable, that it has forgotten which , if either side, was originally earthed, or off-earth  by the impulse voltage, when it set off.

If it is any help we do not bother with curly pigtails to earth  racks in the HV facility here, but if they arrive with them on, we do not bother to take them out. It may very well be that the confusion is the result of a similar approach.

Earthing for safety would normally be done via the mains leads, unless there are high leakage currents to consider, and then the mains lead develops an earthed screen , and maybe a side-along green and yellow. All of this is then choked to RF with ferrite beads -which are to all intents invisible at 50Hz, but as others have said, can make all the difference in the hundred nano-second domain.

Apologies for the late response, I have been away from the internet, in any case others have it covered.