The generally accepted rule is that the vapor barrier goes on the warmer side of the wall. There are two reasons for this:
In summer, the building's air conditioning will remove moisture so humidity will be relatively low. In winter, heating will increase air temperature, but the outside air will be crisp and very low in humidity. If you are in the South, you will have high heat and humidity outside in summer, which the air conditioning will reduce. But it won't reduce it as much as in the North. During winter you may still use air conditioning in regions where it never gets really cold, which means lower humidity as well.
In any event, no single analysis holds completely true for every part of the country, but we can still draw a general rule. For an interior data center the "office" or "exterior" side of the wall should still be "warmer" than the data center side. However, this is where things get a little sticky. In a data center, we are not only cooling to levels below the normal office environment, we are also forcing the humidity to abnormal levels relative to the temperatures we are running. Therefore, it is a virtual certainty that the data center will be both colder and more humid than the offices, at least while the building environmental systems are operating in the office areas.
This is not in line with "norms." Even though we would usually think of warmer air as being more humid, this doesn't hold true between offices and the data center. This throws a bit of a monkey wrench into conventional theory -- except for one thing: While we specify both a relatively low temperature and a relatively high humidity range for the data center, we always tack on the words "non-condensing."
Obviously, we don't want a situation where moisture starts condensing on cabinets or equipment. The air coming out of the CRAC units should not go below the dew point when humidity is added. And if it won't condense on our equipment surfaces, it certainly won't condense on or in the interior wall. Therefore, the "outside" wall is still the place where any concern might arise. But let's recognize that this condition is likely to arise only on summer weekends or holidays when building systems are shut down and the office air truly becomes both hot and more humid than that in the data center. Considering this explanation, the vapor barrier isn't really going to do much of anything most of the time.
Contractors love poly sheeting, but it is more easily damaged and compromised during construction than most any of the other materials. If your insulation is for noise-control purposes, foil-faced insulation is probably not a good choice (unless it's rigid board insulation) since standard batt insulations do not have the density to provide much sound attenuation.
But thermal control inside the data center is usually the reason for insulating the wall, not sound attenuation. It is expensive to maintain a good data center environment and thick, loose batts are the right choice for that purpose. If you also want noise control, it would probably be best to use thermal insulation between the studs, put a layer of rigid, noise control fiberglass under the gyp board on the data center side of the wall and use foil-faced gyp board on the office side. But frankly, for the amount of condensation control you are actually dealing with, the specific type of vapor barrier is not very important. Whatever your contractor is most comfortable installing is probably the best choice.
ABOUT THE AUTHOR: Robert McFarlane is a pioneer in the field of building cabling design. He has been asked to speak at countless seminars on building infrastructure for electronic communications, evolving technologies and the requirements of trading floor and data center design. Mr. McFarlane served for twelve years as President of Interport Financial, Inc., a firm specializing in designs for financial trading floors and critical data centers.
This was first published in January 2007
This was first published in January 2007