In ac circuits, the relationship between volts and amps is not straightforward. The relationship is dependent on the type of load. Resistive, capacitive, and inductive loads (both linear and nonlinear) are possible. For simplicity, this article considers only linear loads.

In resistive loads, the voltage (V) and current (I) are in phase. Therefore, the ac input in volts times amps (apparent power, VI) equals the ac input in watts (real power, W). However, for both capacitive and inductive loads, the current either leads or lags the voltage, and so VI is not equal to W. The factor that relates volts, amps, and watts in resistive-inductive or capacitive ac circuits is called power factor (pf). By definition,

(2)

For real power, W = VI cos(f) (V times the component of the current I in phase with V) and so,

(3)

The phasor diagram in Figure 1 illustrates Equation 3. In theory, pf can have a value between 0 and 1. In the real world, pf is usually between 0.6 and 0.8 (inductive) for most power supplies. A higher pf also signifies a better-designed power supply. With this information, the relationship between, h, pf, output watts, and input VA can be obtained.

Figure 1. A phasor diagram illustrating the power factor for real power.

When multiplying Equation 1 by Equation 2, the result is:

(4)

Hence,

output watts = h x pf x input VI, (5)

or

(6)

The product of h and pf (assuming the above values) would be in the range of 0.36 to 0.64 for most power supplies. The data on a power supply's pf and efficiency should be available from the manufacturer. For example, the total dc output for a power supply is usually available. The examples in this article assume the use of a 250-W power supply. Using the formula in Equation 6 and the range of 0.36 to 0.64, the resulting range for ac input (VI) would be 391 to 694. For input amps at say, 100 V, a range of 3.91 to 6.94 A would be obtained, depending on the power-supply design.

The problem is that most power-supply manufacturers rate their power supplies (on the safety agency label) at higher values (ranging from 7 to 10 A). Although the reason for this is unclear, one possibility is that they are rounding off to the next higher value. However, providing only an estimated rating can sometimes cause problems. It becomes a critical issue when the end product (e.g., a computer or server) is rated at the same value as the power supply. For example, if the unit is rated 100 V, 7 A (the same as the power supply), safety agencies require the unit to be loaded during testing (to 80% of the rating, i.e., 5.6 A). Quite often, this results in excessive loading of the power supply. Excessive loading results in higher temperatures of the critical components, to the point at which the temperature limit is exceeded, resulting in a test failure. Moreover, the measured input amps end up being less than 80% of the rated input amps.

An alternative would be for safety agencies to require units to be loaded based on 80% of output watts. Some engineers from the safety agencies accept this method, but many do not. One reason some do not accept it could be because it is much easier to measure input amps than output watts.

The Solution

The solution is to determine the actual input amps drawn by a fully configured system, and then to select a value that is only slightly higher than what is drawn. For example, if the fully configured system draws 3.5 A, the system could be rated at 4 or 4.5 A.

A fully configured system is one in which all drives are present along with the motherboard and controller. All input/output (I/O) slots are populated (if present) with representative I/O cards (modems, sound cards, small computer system interface controllers, etc.). Some of these loads can be simulated by the use of dummy resistive loads connected to the 5- and 12-V dc outputs.

It is important to note that safety agencies do not require the end product to have the same rating as the power supply. Proper selection of the input rating therefore results in a more realistic rating, which, in turn, results in the power supply operating at lower temperatures during testing.

Remember that power supplies are tested on the bench, and all loads (and the resulting heat) are external to the power supply. When used in the end product, all of the heat generated is within the enclosure. This heat produces unpredictable results on the temperatures of the critical components of the power supply. This unpredictable heat generation is, of course, the precise reason that temperature tests are repeated on end products.

Conclusion

The end-product rating should be determined carefully. Products with the same rating as their power supply can be problematic. Select a power supply with high efficiency and pf values. High values indicate a well-designed power supply.

Safety agencies do not require that the end-product rating be the same as the power supply. A more-realistic rating results in cooler operating temperatures, and, therefore, the chance of failure decreases during safety testing.

End-product customers, especially those purchasing multiple units, also consider the cost of the electricity consumed. This cost is calculated based on the unit's label rating. An improperly rated unit would imply that the unit consumes significantly more electric power than say, a competitor's unit, which may be rated lower. Energy consumption could become a major purchasing criterion in the future if buyers must address power shortages and the resulting higher energy prices.

Finally, the issue of a poor pf may be resolved soon, as more and more products are designed to comply with the requirements for CE marked products for export to Europe. To comply with the harmonics test specified in EN 61000-3-2, power-supply manufacturers are being forced to improve the pf. Assuming that the same power supply is sold to U.S. markets, end-product manufacturers must only find more-efficient power supplies to produce a more competitive product.

Reference