Chasing the Elusive Power Factor

Dear Home Power, I’m writing to defend Carol Montheim (HP92, page 48), who had it exactly right until the powerfactor nitpickers descended upon her with such a clatter that she was forced to recant half of her evidence for the virtue of compact fluorescent lightbulbs (CFLs).

The nitpickers were right to say that power companies supply power in volt-amperes (VA). And they were right to say that most CFLs have a power factor rating of about 0.5, which means that a 25 watt CFL takes 50 VA to run (25 watts divided by 0.5). But it is not true that utilities must therefore burn twice as much coal or cook twice as many atoms in order to supply twice as much energy to run Carol’s CFLs. Power factor does not affect the energy consumption of homes either on or off the grid.

Volt-amperes are a measure of apparent power (as in false and tricky). Watts are a measure of true power (as in honest and upright, like Carol). When AC power hits reactive loads like CFLs or TVs or air conditioners or surge protectors or computers, electrical current and voltage tend to get jostled out of phase with each other, causing voltamperes (sneaky false power) to rise above watts (foursquare honest power).

So our reactive 25 watt CFL still uses an honest 25 watts, even though amps and volts go out of phase and the sneaky VA rise to 50, making energy consumption appear to double. The extra 25 volt-amperes is not lost or consumed, just stored or borrowed. It goes to work somewhere else. It may offset the power factor of another appliance or it may go back out to the utility line where industrial capacitors tweak it back into phase. Utility meters may not read the extra volt-amperes that CFLs suck into the house, but neither do they read the extra voltamperes that blow back into the grid from Carol’s house. Everything gets canceled out, and the only thing actually consumed is the 25 watts.

Inductors (which cause current to lag voltage) and capacitors (which cause current to lead voltage) store or borrow volt-amperes that are not consumed as watts. They return those unused volt-amperes to the circuit later. For example, pure inductance does not consume energy; it stores it as a magnetic field. When that magnetic field collapses, the energy is returned to the circuit.

Utilities like the power factor of loads to be close to 1.0 (unity) because that makes it cheapest to distribute energy. These companies constantly adjust loads to bring current and voltage back into phase without using significant extra power. Commercial fluorescent lights usually have power factors of 0.9 or higher. Utilities lobbied to get residential CFLs to have similar power factors, but bulb manufacturers revolted because high power factor bulbs are more expensive to make. But the issue was power management, not energy consumption.

To prove this point, I set up an experiment with CFLs vs. incandescent lights (resistive loads with a power factor of 1.0), and measured watts and volt-amperes delivered by my inverter. Then I measured the amps from my power source, a bank of L-16 batteries, which are DC sources wise to the tricks of reactive power.

A 75 watt incandescent bulb measured 70 watts and 70 volt-amperes out of the inverter, which is what you would expect with a power factor of 1.0. The TriMetric meter measuring true power from the batteries measured 71.8 watts (volts times amps). The slightly higher wattage was due to inefficiency in converting 12 volt DC to 120 volt AC.

Then I lit a 25 watt CFL. It measured 25 watts and 52 volt-amperes out of the inverter, which indicated a power factor of about 0.5. The TriMetric, however, showed the true power needed by the batteries was 26.5 watts. The low power factor did not require the batteries (or the utility) to produce any extra energy.

This is not to say that volt-amperes are not important. Household wiring must be sized for the highest volt-ampere load it will carry, not for the highest watt load. Inverters, too, must be able to handle the highest volt-ampere load thrown at them, not the highest watt load.

Nonetheless, CFLs save as much energy as Carol said in her first story. With our nation’s leadership in full attack mode on the environment, let’s not make Carol’s job any harder. My hat’s off to you, Carol.

Josey Paul, Joyce, Washington

Hello Josey, Thanks for your great letter. The figures in Carol’s spreadsheet, Sylvania’s power factor info sheet, and my response to Carol regarding power factor were misleading and way too high, as you point out. Unfortunately, the affect of power factor on energy consumption isn’t quite as clear-cut as you describe.

One point that definitely shouldn’t be overlooked regarding power factor is that all current causes losses in electrical distribution systems. This is true whether the system is the electrical wiring in a given building, or major utility transmission lines. The degree of loss is what’s hard to quantify.

Here’s an example to illustrate the effects of line loss in relation to power factor. Say a given pump has a power factor of 0.60. To do 1,000 W of useful work requires 1,667 VA of apparent electrical power. At 240 VAC, the electric current demand is 6.9 A. Replacing the pump with one that has a power factor of 0.95, would reduce the amount of apparent power to about 1,052 W, and reduce the circuit’s current to 4.4 amps. This amounts to a 36 percent reduction in the level of current. In almost all cases, this will mean less power loss. How much depends on the length of the wire runs and the gauge or size of the wire in the circuit.

A second way that power factor affects an RE system’s consumption is inverter efficiency. Inverters operate less efficiently when low power factor loads are being powered. Sandia National Labs runs reactive load tests on inverters. Their data shows that Trace SW5548 inverters powering loads with a 1.0 power factor run 4 percent more efficiently than a 1,000 watt load with a 0.5 power factor. That means more amps out of the battery. At 3,000 watts, the efficiency difference is 7 percent. It’s important to note that different combinations of loads and inverter technologies may result in higher or lower efficiency figures. But the bottom line is that low power factor loads decrease inverter conversion efficiency.

We have some comprehensive experiments planned to quantify the affect of low power factor loads on overall energy consumption. I’d be very interested to hear from RE or utility folks regarding this, since it’s a complex and often misunderstood electrical concept.

Joe Schwartz • Home Power