Forthcoming standards may lower humidity floors in data centers, making some wonder if electrostatic discharge will result.
by Patrick McLaughlin
The complex whole that is a data center comprises multiple systems and technologies. Several recent articles in this publication have described the data center as an ecosystem, in which technological and mechanical systems are interdependent and affect one another. In ecosystems-and the data center is no exception-sometimes these systems work in conflict with each other, so that what is good for one is detrimental to another.
That appears to be exactly the case when it comes to the prospect of electrostatic discharge (ESD) accumulating on unshielded twisted-pair copper cables. ESD has been a concern for several years, evidenced by the fact that in 2003 the Telecommunications Industry Association (TIA; wwwtiaonline.org) issued Telecommunications Systems Bulletin TSB-153 Static Discharge Between LAN and Data Terminal Equipment. That document, as well as the TIA’s 942 Telecommunications Infrastructure Standard for Data Centers, addresses ESD mitigation.
Jonathan Jew, principal with J&M Consultants (www.j-and-m.com) and co-chair of the committee that wrote the 942 standard, recalls that when the standard was being developed, he conducted ongoing discussions with representatives of the American Society of Heating, Refrigeration, and Air Conditioning Engineers (ASHRAE; www.ashrae.org). “ASHRAE develops guidelines for mechanical engineering professions,” Jew explains. “I was in contact with them during the development of 942. They recommended a 35 to 55% humidity range for Class 1 spaces, which included data centers. Some of our members had noticed that humidity levels below 40% increased ESD issues. We suggested to ASHRAE that 40% may be a better number.
“ASHRAE prefers a lower value, because if humidity is too low, they need to add humidity-raising energy costs,” Jew explains. “But ESD concern was significant enough that they did not recommend below 40%.”
In sync, for now
Ultimately, the 942 standard and ASHRAE specifications harmonized in that both specified 40 to 55% humidity levels in such spaces. Recently, however, ASHRAE has been reconsidering that humidity-level floor, and may soon specify lower humidity levels.
ASHRAE’s Technical Committee 9.9 Mission Critical Facilities, Technology Spaces, and Electronic Equipment recently was given the go-ahead to conduct a study entitled “The Effect of Humidity on the Reliability of Data Centers.” In what the group expects to be an 18-month study, it will seek “to determine the role of lower humidity limits and humidification on the reliability of data centers, including the impact on ESD,” according to the stated objective in the group’s official proposal. Also from that proposal, specific objectives include:
- Establish a lower humidity limit for protection of data processing environment equipment (servers and data storage devices). Since this limit is currently set at 40%, with the objective of reducing the risk of ESD, it is believed that this research will either significantly reduce or eliminate a lower humidity limit.
- The result of the reduction or elimination of a lower humidity limit will significantly reduce energy usage in data centers through increased use of air-side economizers and reduced occurrence of simultaneous or sequential humidification and dehumidification due to errors in computer room conditioning control units.
Mark Hydeman, PE, a member of ASHRAE’s TC 9.9 and principal of Alameda, CA-based Taylor Engineering (www.taylor-engineering.com), stresses the significance of the energy consumption inherent in deliberately raising humidity levels. Citing a recent report from the Iowa Energy Center (www.energy.iastate.edu), he says the humidity sensors used within computer-room air-conditioning units (CRACUs) generally are low-end devices and do not gauge humidity with a high degree of accuracy. Additionally, relative-humidity sensors’ readings can vary from one CRACU to another because those units may have different return temperatures, depending on air-flow management.
“In many cases you have poor sensors that are poorly located,” Hydeman says. “They are controlling humidification and de-humidification.” Many times, he says, he has entered a data center to find some units humidifying the air while others are dehumidifying it.
The use of air-side economizers, as described in a previous article in this publication (see “Hot-air isolation cools high-density data centers,” September 2007, p. 47), holds the potential to significantly reduce the energy spent on data-center cooling. ASHRAE’s study will further pursue these possibilities.
ASHRAE’s research-project proposal included further detail on the potential benefits of air-side economizers, and why high humidity levels may be an obstacle to those benefits. The proposal states, “Given the constant loads in most data centers, air-side economizers can be extremely effective in reducing mechanical cooling. Unfortunately, there are barriers to air-side economizer use in data centers. Air-side economizers are of little value if tight humidity control is required because the sensible cooling savings of the economizer can be offset by increased humidification of the outside air.” It also states that concerns about dust and contamination are potential barriers to the use of air-side economizers in data centers, adding that Lawrence Berkeley National Laboratory (www.lbl.gov) is currently funded to conduct research on the issues of dust and contamination.
In addition to producing energy inefficiencies, Hydeman asserts there’s no guarantee that higher humidity will eliminate the potential for ESD. Network Equipment Building Systems standards, put forth primarily for equipment used in carriers’ central offices, have wide guidelines for humidity control. Hydeman also points to the most recent edition of the ESD Association’s (www.esda.org) standard “ANSI/ESD S20.20-2007 For the Development of an Electrostatic Discharge Control Program for Protection of Electrical and Electronic Parts, Assemblies and Equipment (Excluding Electrically Initiated Explosive Devices).” That standard, he points out, does not list humidity as a primary means of ESD control. “The reason,” Hydeman says, “is the humidity does reduce the risk of ESD, but does not eliminate it. Raising the humidity level may give people a false sense of security.”
What might provide a justified sense of security? Many say grounding will. The use of such devices as grounding plates, as well as personnel grounding, are their recommendations for ESD control. The foreword to ANSI/ESD S20.29-2007 states: “All conductors in the environment, including personnel, shall be bonded or electrically connected and attached to a known ground or contrived ground.”
Hydeman says that ASHRAE TC 9.9’s current research project will gather data on why servers fail, and whether there is any correlation with ESD events. The project also will correlate personnel grounding and the size of ESD charges over a range of relative humidities.
J&M Consutants’ Jew believes that the concerns among cabling professionals are not necessarily the same as those who wrote the ESDA standard: “Their standard is for environments working with electronics, for example. They recommend procedures such as wearing wrist straps,” for personnel grounding. A difference, though, as he points out, is that, “Where they’re concerned with short-term ESD buildup, we’re concerned with long-term buildup”-over the length of a cable, and over long periods of time. He also states that from a practical standpoint, relying exclusively on personnel-grounding techniques also relies on enforcement of those procedures. In environments in which large numbers of people are working, enforcement will almost assuredly be a challenge.
So, in the ecosystem that is the data center, two concerns are at odds with each other. It is likely to be some time before a determination is made concerning which interest will prevail. But in a world that is increasingly conscious of energy efficiency, it appears those favoring the elimination of minimum humidity levels have a compelling argument.
Patrick McLaughlin is chief editor of Cabling Installation & Maintenance.