Sizing Air Conditioners: If Bigger Is Not Better, What Is?

John Proctor is the managing partner of Proctor Engineering Group in San Rafael, California. Peggy Albright is an independent writing consultant for researchers in the electric utility industry.

Equipment sizing is not the only key to comfort-small details in register placement and design can have a great impact on air flow and overall comfort. For example, the louvers of this air-conditioning register are designed to spread air in four directions.

Air Conditioning Contractors of America (ACCA) has published design manuals (Manuals J, S, D, and T) that produce far better results than the rough-and-tumble rules of thumb used by the vast majority of HVAC contractors. A contractor will achieve (and the customer will enjoy) a much higher-quality job if these manuals are followed in the design and installation of central air conditioning systems. A recent investigation of new houses has shown that an air conditioner delivering a capacity equal to Manual J would be adequate even during extraordinarily hot summers (see How Big Is Big Enough?).
 

How Big Is Big Enough? Figure 2. Hourly sensible cooling load versus outdoor temperature monitored for a house in Phoenix, Arizona during an extraordinarily hot summer. Manual J overestimated the sensible cooling load for this house by at least 50%. Even during this hot summer, an air conditioner sized to two-thirds of Manual J would have been more appropriate. An air conditioner sized to ACCA Manuals J and S is big enough. Industry specialists who design and sell air conditioners have long used Manual J as a standard method for determining the amount of cooling needed to deliver thermal comfort to single-family residences. The procedure is used to calculate room-by-room loads for duct design purposes and whole-house loads for equipment selection. It was jointly developed by ACCA and the Air-Conditioning and Refrigeration Institute (ARI), and is based on a number of sources, including the ASHRAE Handbook of Fundamentals. Despite the widespread use of this procedure, many contractors have been reluctant to believe that Manual J can deliver adequate cooling under design conditions. One reason for this reluctance has been the lack of information about how actual cooling loads compare to Manual J estimates. Many who have used Manual J extensively have long suspected that it has an oversizing margin. Until recently, however, no field studies had been performed to verify this anecdotal evidence. New data show that Manual J indeed overestimates the sensible cooling load in hot, dry climates. It is likely that the same holds true in hot, moist climates. Proctor Engineering Group, the Electric Power Research Institute, Nevada Power, and Arizona Public Service monitored air conditioning systems installed in new homes in Phoenix, Arizona, and Las Vegas, Nevada. By testing the actual cooling capacity required to maintain comfort under severe conditions, these tests have yielded the first measurements that confirm and quantify the overestimation by Manual J. The studies showed that even during an extraordinarily hot summer, when almost 200 hours exceeded design conditions (design conditions are exceeded only 73 hours in a typical summer), the actual sensible cooling loads of the houses were less than Manual J estimates. At the most intensively monitored sites in the studies, we recorded air flow, temperature drop, and moisture removed from the conditioned air. The research team calculated the actual capacity delivered by the air conditioner for every air conditioner cycle. The systems were monitored from July 30 through September 25, 1995. Occupants were free to adjust their thermostat settings to any value, but most kept a constant thermostat setting. Most of the systems monitored were typical installations (including leaky ducts, which increased the cooling load that the equipment needed to deliver).  Figure 2 shows the hourly sensible cooling load and the outdoor temperature in one typical house. The duct system had a 12% return leak and a 6% supply leak. Outdoor temperatures at this house ranged as high as 116oF (according to ASHRAE Fundamentals, the mean extreme temperature for Phoenix is 112.8oF). Even though this time period was extraordinarily hot, the sensible load requirements for all but 3 (0.2%) of the 1,316 monitored hours were less than the Manual J estimated cooling load. Manual J overpredicted the design load for this house by almost 50%. There was no need to oversize the air conditioner beyond the Manual J cooling load because Manual J already overestimated that load. The air conditioner installed in this house had a design sensible capacity 24% larger than Manual J-excess capacity that was not useful. The homeowners paid approximately $330 in additional first costs, and they will pay unnecessary additional operating costs every summer month for the life of the system.

We're not sure what manual the installer of this creative air conditioning system was working with, but we don't recommend it!

The main problems typically found in the field are improperly sized air conditioners, improperly designed duct systems, poor grille selection, and poor installation of all three components. These problems are most easily avoided in new construction, but retrofit contractors can and should follow the recommendations in this article whenever feasible.
 
 

As we talked the air conditioner came on and a strong stream of cold air moved by my shoulder. The owner went over to the supply register and closed the damper. He came back to the table explaining that with the register open he was blasted with cold air that made him uncomfortable. The noise coming from the closed register made it hard to have a conversation at the table. He stated that the system was always noisy. When I suggested that we move to another room for our conversation, he said, That wouldn't make any difference, there are only hot places and cold places; no place is right in this house. We are looking for a new house.

The situation we found in this house exists, in various degrees, in millions of homes across the United States. The heating and cooling distribution system was not matched to the cooling loads of the individual rooms or to the needs of the occupants. On top of that, the air conditioner was not matched to the distribution system. Discomfort and expense are the inevitable results.

The comfort zone was found to be acceptable to 90% of test subjects drawn from a range of age groups and genders, with work and life-styles involving varying levels of activity and clothing. An air conditioning system that establishes and maintains indoor conditions within this zone will provide thermal comfort. It will produce a neutral sensation-occupants will feel neither too hot nor too cold.

An air conditioner can easily bring the temperature inside a house into the comfort range. In fact, bigger air conditioners virtually ensure that the temperature at the thermostat can be as cold as we set it. Unfortunately, cold alone is not comfortable. In fact, it is distinctly uncomfortable. To maintain a general level of comfort, the moisture level must also be controlled. This is best achieved by smaller, not larger, air conditioners.
 
 

Figure 1. Smaller air conditioners remove more moisture from a house. In this example, a 5-ton unit running for five minutes removes 1.4 pounds of water. A 2.5-ton air conditioner in the same house, running for ten minutes, removes 1.7 pounds of moisture-an increase in moisture removal of 21%.

An air conditioner's ability to remove moisture increases when the equipment runs for longer periods of time. At the beginning of every cycle in hot moist climates, the air conditioner actually puts moisture into the house as water evaporates off the inside coil. Once it's been running a while, it begins to remove moisture. Since a smaller air conditioner runs longer to keep the house at the temperature setpoint, it removes more moisture than a larger unit would be able to (see Figure 1).

The amount of moisture removed is a function not only of how long the air conditioner runs, but also of its Sensible Heat Ratio (SHR)-the percentage of the total capacity delivered as lower house temperature. A low SHR will result in more moisture removal. For hot, wet climates, air flow across the coil should be reduced slightly to decrease the SHR, and the air conditioner condensing unit and indoor coil combination should be chosen to have a low SHR. Typical matched units from major manufacturers have SHRs in the 68%-80% range when it is 95oF outside and 75oF with 50% relative humidity inside. Note that if you don't use a matching indoor coil and outdoor unit from the same manufacturer, you shouldn't expect to get their published SHR.
 
 

A place in the sun may be a good thing, but too much sun can make a house's cooling load soar. The lack of overhangs and sun protection on this new house will concentrate heat gains in certain rooms, making it difficult to properly balance cooling distribution.

Uneven temperatures have become more common due to the modern practice of severely reducing overhangs above the windows. Without overhangs, rooms with west-facing windows will overheat in the afternoon, since their need for cooling can easily double.

An inefficient method of attempting to get proper distribution and mixing of the air is to use a large air handler fan to circulate air all or most of the time. This is sometimes effective in mixing the air, but at a high price. There is an old rule of thumb that between four and six house volumes of air must pass through the air handler in an hour. At six air changes per hour (ACH), a 1,400 ft2 home would need a continuously running fan that delivers 1,120 cubic feet per minute (CFM)-equivalent to almost 3 tons-regardless of the cooling load of the house. The common practice is to install an air conditioner (inside and outside unit) with the capacity to meet those flow requirements. There are many disadvantages to this practice. They include:
 

A better solution is first to design and install a delivery system that properly distributes the cooling to each room, then to select and place supply grilles that throw the delivered air into the right places in the room to promote mixing. ACCA's Manual D: Duct Design and Manual T: Terminal Design can lead the installing contractor through the process of selecting the proper-size duct and type of register, based on the location of the register, the size of the room, the restriction of the duct run, and the dimensions and heat gain of the room. Unfortunately, only the best contractors and builders ever pay attention to these critical details.

The problems of stagnation and overheating can be reduced by proper implementation of ACCA procedures. These problems can be further reduced by ensuring that the assumptions built into the manuals are not violated. For example, it is assumed that there is no duct leakage in the system. Any longtime reader of Home Energy will immediately note that this assumption is violated in nearly all homes (including new ones). Proper installation of the duct system and leakage testing are essential to obtain comfort.

Another assumption is that the conduction losses are the same percentage of the delivered cooling regardless of the length of the duct run. This would be an insignificant assumption in a heavily insulated system (R-4 is not heavily insulated). Long duct runs through the attic lose over 15% of their cooling capacity before the conditioned air reaches its destination. Long duct runs need additional insulation to deliver the proper amount of cooling to the distant rooms.
 

Water and air are being evacuated from the lines and indoor coils of this air conditioner. This process, which is often overlooked or avoided by installers, can assure that units are properly charged, and also provides an opportunity to check for leaks.

An oversized air conditioner is a major contributor to drafts, because it is almost always married to a duct system that is too small. The ducts are unable to deliver the amount of air necessary for proper air conditioner performance (more on this later). The result is a poor compromise-air flow that is too low for the air conditioner and too high for the duct system. The low air flow across the oversized coil produces colder delivery temperatures, and the high air flow through the ducts and grilles produce high pressures, noise, and high velocities at the grilles. The grilles themselves are often too small and without proper throw or spread (particularly the cheapest ones). When low delivery temperatures are coupled with high-velocity discharge through inappropriate and poorly placed grilles, occupants experience drafts.
 
 

When an air conditioner and duct system are properly sized to meet the cooling load, they can easily distribute the cool air without being noisy. To design a quiet system, keep every supply grille below NC-25 with a face velocity below 700 feet per minute.

Grilles with dampers are invariably noisier than equivalent grilles without dampers. When the dampers are partially closed, the pressures and leaks in the ducts increase and the air flow across the coil is reduced. Occupants generally close dampers to redirect air to another room that needs more delivery. If the system is designed correctly, neither register dampers nor inline balancing dampers should be needed.
 
 

The constriction of this duct has the unfortunate result of reducing the air flow across the cooling coil.

In a recent laboratory test of a high-efficiency air conditioner, Proctor Engineering Group found a 7% drop in efficiency when the air flow was reduced by 30%. In order to ensure that the design air flow is being achieved, the installing contractor must measure and correct the air flow across the inside coil.
 
 

In the summer of 1995, Proctor Engineering Group and Arizona Public Service Company monitored a group of 22 newly constructed homes. Nearly all of those homes had undercharged air conditioners. One of the worst units had 62% of the correct charge (and 79% of proper flow). The homeowner complained to the builder that the air conditioner was not working right. She was told that the wrong amount of insulation had been installed in her attic, and an insulation contractor was called in to apply additional insulation. Shortly thereafter, the true problem showed itself when the air conditioner compressor failed.
 
 

These cardboard frames are specifically designed for insulating over ducts. The walls of the boxes allow weatherizers to build up a deeper layer of insulation around long duct runs.

To ensure a tight duct system, the installing contractor must test duct integrity using specialized tools (see HE Sept/Oct '93 for more information on duct testing).
 

Because of the inefficiencies associated with the start-up of the air conditioner, a smaller unit will produce the same amount of cooling with lower energy consumption, under most conditions.

It is not uncommon for poor cooling performance to be attributed to insufficient equipment size, when in fact there is more than enough cooling capacity. We know designers who determine the system air flow based on floor area (this oversizes the air conditioner in energy-efficient homes), and then try to squeeze down the size of the duct system so that it can be installed in the house. They explain that they can't use a higher insulation level on the ducts because there is no room, and, when faced with poor performance, increase the size of the air conditioner.

Most household air conditioning problems will be eliminated when the capacity of the air conditioner is reduced to ACCA Manual J and Manual S standards; an appropriately designed, insulated, and leakproof distribution system is used; and the system is installed to meet the manufacturer's standards.

Manual J, D, S, and T. Available from Air Conditioning Contractors of America, 1712 New Hampshire Ave., NW, Washington, DC 20009. Tel: (202)483-9370.

• The need for a larger and more expensive duct system to handle the increased flow.
• Increased duct conduction due to constant circulation and the larger surface area of the duct system.
• Reduced moisture removal due to short compressor cycles, caused by the oversized outdoor unit.
• Reduced moisture removal due to the constant air circulation, because water re-evaporates from the coil while the compressor is off, and is distributed back around the house.
• Increased cooling load due to duct leakage and fan energy delivered as heat.

  A draft exists when unwanted air movement causes cooling on one part of a person's body. The colder the air and the faster it is blowing, the more offensive drafts are. Air conditioning drafts are characterized by cold, high-velocity air striking the body. Studies show that these drafts are even more offensive if they are intermittent. We all know how noisy forced-air cooling systems can be. The noises can come from the grilles, the ducts, and the air handler fan. Our perception of noise is affected by both the frequency and the level of the sound. Higher-frequency sounds (such as those generated by high discharge velocities at grilles) are more offensive than low- frequency sounds (such as those generated by the fan). For grilles there is a Noise Criteria (NC) rating that mimics the human perception of sound. The NC for a particular grille increases as more air is forced through it. There is a lot of emphasis on the rated efficiency of air conditioners. Unfortunately, this necessary attention to equipment design has overshadowed efforts to improve the selection and installation of the entire air conditioning system. Builders, contractors, and the buying public all incorrectly assume that if they spend the money on a high-efficiency air conditioner, they have gotten all the efficiency they can. But common problems such as oversizing, improper installation, low air flow, and leaky duct systems mean that customers don't get the efficiency they paid a premium for.
 
  Most air conditioners are designed to have 400 CFM per ton of air flow across the inside coil. When the air conditioner is coupled with a duct system that meets Manual D criteria, the proper flow is achieved. However, since air conditioners are commonly oversized for the heat gain of the home and the duct systems are not designed to Manual D, even new systems are usually deficient in air flow. This situation only gets worse as the inside coil picks up dirt. A new split system air conditioner comes from the factory with the proper amount of factory-installed charge for a standard length of refrigerant lines. When the unit is installed, the contractor needs to evacuate the lines and indoor coil and weigh in any additional charge needed if the installed lines are longer. Evacuation also allows the installer to check for leaks. Most of the time, evacuation is not done. As a result, air and moisture are captured in the line set and coil, the unit ends up undercharged, and leaks are not detected. In many cases the amount of undercharge is severe. The evidence against leaky and underinsulated ducts continues to mount. Leaky ducts are a large contributor to system inefficiency that gets worse when it's hotter outside. The Arizona Public Service Company test found that sealing a 13% supply leak saved 22% of the cooling energy consumption when outdoor temperatures were between 100oF and 105oF. Air conditioners are very inefficient when they first start operation. It is far better for the air conditioner to run long cycles than short ones, because efficiency increases the longer it runs. For example, increasing the run time from five minutes to nine minutes resulted in an energy savings of 10% for the unit described in Bigger Is Not Better (HE May/June '95). F.C. Houghten and C.P. Yaglou: ASHVE Research Report No. 673, Determination of the Comfort Zone, ASHVE Transactions, Vol. 29, 1923, p. 361.

Home Energy can be reached at: contact@homeenergy.org
Home Energy magazine -- Please read our Copyright Notice