Residual current protection – in the data centre
The latest developments in RCD technology stop nuisance tripping, enable regular testing of IT equipment redundancy and offer quantum leaps in safety, writes Brisbane data centre engineer David Morley.
For a long time residual current devices (RCDs) within the data centre were, and commonly still are, viewed as inviting trouble. Nuisance tripping issues, phone calls in the middle of the night and unexplained outages have all been experienced. These are common stories based on the reputation of the old Type II residual current devices.
Sadly, this perception of RCDs still exists within the ICT (information and communications technology) industry and is inconsistent with the latest developments in RCD technology. There have been great improvements in reliability and performance of RCDs, and it should be noted that the latest changes to WHS (work health and safety) legislation and Australian Standards make these devices mandatory in most situations.
Thankfully, technological strides in reliability over the last few years have demonstrated that RCD protection is not only about improving workplace safety, but also enabling reduction in operational risks. RCDs are well-established as a fire-prevention strategy and regular testing ensures that all components of an ICT system or service operate redundantly and this redundancy is regularly tested.
So what is residual current and why does it require protection?
Residual or leakage current is any current flowing to earth, and is generated by all IT power supplies, typically .5 to 1.2 milliAmperes (mA) per connected power supply. Leakage current is normally a by-product of the power supply, but in cases where a human is involved it is likely to be a situation where contact has been made with an active conductor and current is flowing to earth.
Just 30mA of leakage current is enough to stop the heart. Often situations occur where hardware or cabling is being modified or due to an actual fault within the equipment, cabling or rack and typically when personnel are involved. When a person makes contact with a live conductor it may lead to fatal consequences without RCD protection.
RCDs are installed to prevent a maximum of 30mA of current flow within 300mS. Standard commercial Type II RCDs will commonly trip as low as 15mA of current flow well under 30mS. They are great at personal protection, but not so reliable when used in a data centre environment.
Thankfully, there’s a new generation of RCDs now on the market from several manufacturers. They are commonly referred to as SI (Super Immunised) or Ai (High Immunity) and have become a game changer with regards to leakage current protection within IT environments. They handle high-leakage currents with very high reliability, while still offering excellent protection for equipment, personnel and environments. Racks filled with SAN, network equipment and servers, some producing greater than 15mA of leakage current per circuit, will not experience nuisance tripping with the Immunised RCDs.
One of the key benefits realised through regular RCD testing is that not only the RCD is being tested but also the redundancy of the system, and all devices within the rack are being tested. Many failed power supplies, incorrectly connected devices and single supply devices can be detected during each test period.
It should be noted here that, in order to effectively use RCDs with ICT environments, equipment should be dual- corded/dual-power supplied, or be connected to an in-rack automatic transfer switch (ATS)
There are several locations allowable for RCDs under AS3000, and locating the devices adjacent to the socket outlets greatly eases the testing process. It is possible to locate the RCDs into distribution boards; however, this can use excessive and valuable chassis space. Installing the RCDs into small enclosures at the rack level has proven to be the most economical and flexible solution to date; however, some manufactures can now build the RCDs into the power rails. While still being evaluated, this is the simplest way to deploy the devices into the data centre.
The dual or redundant supply nature of the devices is where one of the key considerations with regards to RCD protection and compliance with AS 3000 comes into play. The RCD exemption sections in AS3000 are often read out of context for IT equipment, particularly enterprise-class equipment that will utilise redundant power supplies and is not at risk of failing with the trip of a single RCD.
AS3000 also states that RCD protection is required for all socket outlets not exceeding 20 amps. Commonly dual 32A circuits are provided to racks to avoid installing RCD protection.
Typically, power rails bristling with 10A and 15A socket outlets are then connected to the 32A circuits. The problem is back again with socket outlets less than 20A requiring protection.
There has been a lot of talk among operators in regards to the exemptions listed in the clauses of 220.127.116.11.1 of AS3000. However, there is only one exemption that can be considered within the context of appropriately redundant equipment: “Where the disconnection of a circuit by an RCD could cause a danger greater than earth leakage current.”
In essence, this clause is asking for a risk assessment process to exclude RCD protection.
The RCD is there to protect life, which is a good place from which to start a risk assessment. Evaluation of the services being provided by the equipment to determine if they are life, health and safety in nature could show a good case for not fitting RCDs. Naturally, this needs to be discussed and evaluated in conjunction with the owners/operators of that particular equipment, and documented risk assessments kept.
Typically, a lot of operators will cite the following clause to exempt themselves from installing RCDs: “The connected equipment is required by the owner or operator to perform a function that is essential to the performance of the installation and that function would be adversely affected by a loss of supply caused by the RCD operation.”
This would be valid if the facilities only contained single- corded equipment. With dual-corded equipment, the loss of one supply does not equate to the device not performing its intended function. Almost all modern IT equipment has dual- supply capability or can be fitted to in-rack ATSs to enable the devices with redundant feeds.
But what if multiple RCDs nuisance trip for the same rack, resulting in both circuits being isolated? If multiple RCDs activate for the same rack (two or more circuits), it will be a genuine fault and most certainly has saved your facility from a dangerous event. It may have dropped the equipment, but most likely saved a life or prevented activation of fire suppression systems.
In addition to this, there is the Harmonised WHS (work healthy and safty) legislation, which is being gradually introduced across Australia. Section 4 of this legislation calls up the requirements for Residual Current Protection to be fitted to all applicable circuits within the workplace.
The fines associated with the various WHS and Electrical Safety authorities across Australia do differ between each state and can range to hundreds of thousands of dollars with a serious incident.
The installation of RCD-protected IT equipment and the redundancy proving nature of the devices, while being regularly tested, can deliver significant benefits along with reduction in risk to staff, contractors and visitors to the data centre.
The question has to be asked as to why is its use not more widespread in the industry?