19" rack xU blanking panels

Timothy Durkin said:

Should blanking panels utilized with 19" racks be earth bonded?

It depends ...

Yes, in the following cases:

• The panels are used to form a metallic enclosure for equipment in the rack, and would become an exposed-conductive-part. This is sometimes the case with some arrangements of distribution equipment racks, or custom-build racks that contain specialist on-off equipment.
  
  
• The panels are used to help form an 'EMC screen'

Not always otherwise.

If the 19" rails and connecting internals are unpainted - say zinc-plate colour passivated - and are themselves earthed/bonded, it may be possible to 'earth' or 'bond' the blanking plates in that way. There are some "EMC rack" products on the market that operate in this way ... sometimes the fixings for the blanking plates and side panels incorporate serrations etc. to ensure they make good contact through paint or oxidation.

See also BS EN 50310 (as referred from BS 7671) for IT and communications racks, and BS EN IEC 61000-5-2 (also as referenced from BS 7671) for general EMC good installation practice.

I agree with Graham, but I'd phrase it differently - what is behind the blanking plate and does it have the potential with any credible fault to make the panel become live ? If not, no real need to bother. Much as you don't bond the door handles at home, unless living in the world that gave us the 15th edition regs.

Wire tail bonds are not actually especially good from an EMC point of view - the panel gaps tend to form slot antennas above a few MHz, and to stop  that you really need paint removal and finger stock that makes a well graunched contact every cm or so, 
or for higher frequencies still,
a wire clad elastomer gasket

- sort of wire knitted sock with a spongy rubber filling.

A 'bond'  wire is a nH or two per mm of length and can under some conditions make an EMC problem worse rather than better by forming an accidental antenna.- 100nH (4 inches)  is ~ 6j ohms at 10MHz and proportionally more as the frequency goes up.  By WiFi and Zigbee frequencies (2400MHz) really it may as well  not be there.

That said I do tend towards to earthing things if in doubt, and would use star-washers where possible. Bonds tend to get left off or broken at servicing.
Mike.

mapj1 said:

what is behind the blanking plate and does it have the potential with any credible fault to make the panel become live ?

This might, however, include being wired (inside the panels) in longer runs of tri-rated rather than insulated-and-sheathed cables.

The bonding requirement for this kind of 19 " rack most likely would be to an assembly or product standard, though, rather than BS 7671.

mapj1 said:

A 'bond'  wire is a nH or two per mm of length and can under some conditions make an EMC problem worse rather than better by forming an accidental antenna.- 100nH (4 inches)  is ~ 6j ohms at 10MHz and proportionally more as the frequency goes up.  By WiFi and Zigbee frequencies (2400MHz) really it may as well  not be there.

Sometimes we are concerned with lower frequencies, though ... it all depends on the particular environment. All covered by BS EN 50310 and BS IEC 61000-5-2 !

From my past experience with 19 inch rack based telecoms equipment. General practice was for the power supplies to be a full self contained removable module with bonded front plate. Ad soon ad the module was removed it was disconnected from the supply.

Power distributec to the rest of the rack was 48v dc or lower and not a risk. Therefore from a safety point of view other than the power supply bonding of the front panel wasn't an issue.

But as emc became an issue and power/heat density increased my industry tended to move more towards 1u an 2u high racks with single front panels and cooling fans at the back. Power supplies also at the back and in their own enclosure within the main rack to prevent access to dangerous voltages and help contain electromagnetic radiation.

My experience of 19" racks has been similar - mains equipment in its own enclosure with nothing more hazardous than sheathed flexes outside the individual boxes. Blanking panels were primarily there for airflow control - enforcing a in-front out-back or similar policy - i.e. stopping warm exhaust air from one server being drawn into the intake side of the server above. In many cases blanking panels would make pretty poor enclosures for hazardous live parts (e.g. live parts covered with only basic insulation) - as there would still be access from the rear or from above/below when other units were slid out for servicing - although I dare say they will have been pressed into service in that way somewhere or other.

   - Andy.

AJJewsbury said:

My experience of 19" racks has been similar - mains equipment in its own enclosure with nothing more hazardous than sheathed flexes outside the individual boxes. Blanking panels were primarily there for airflow control - enforcing a in-front out-back or similar policy - i.e. stopping warm exhaust air from one server being drawn into the intake side of the server above.

This might be true of 19 " racks for information technology, CCTV and similar equipment today

BUT, even in the IT industry, computer racks at one time contained huge custom linear (not switch-mode) power supplies that required blanking panels to prevent access to live parts at the rear.

.... and even today, other industries use 19 " racking solutions differently (and in fact in the way I described).

AJJewsbury said:

. In many cases blanking panels would make pretty poor enclosures for hazardous live parts (e.g. live parts covered with only basic insulation) - as there would still be access from the rear or from above/below when other units were slid out for servicing - although I dare say they will have been pressed into service in that way somewhere or other.

One example of an alternative approach is the temporary power industry ... yes, some distro boxes are made of "self-contained" plug-together units, but others are "wired internals" and need blanking panels where there are no "facia-plate" modules installed.

There are other examples, including laboratory test and measurement equipment, etc., that utilize racking in a similar way, so that the whole "internals" (meant to be accessible only to skilled and/or instructed persons) behind the panels are inaccessible without the use of a tool.

It really isn't that uncommon, and I've worked on/with equipment like this for much of my career.

The original post wasn't specific about the purpose of the racks, so I didn't want to make assumptions ...