Silence of the Lamps:

Reducing radio frequency interference from small fluorescent lamps

Tubular fluorescent lamps with electronic ballasts are very popular because of their instant-on characteristics, low audible hum, and compact size. However, many such lamps produce unacceptable amounts of radio frequency interference (RFI). This RFI appears at 30-40 kHz intervals up to about 500 kHz. I purchased one of these lamps, and after being disappointed in the amount of noise it produced, I did some tests using a F13T5 bulb with different ballasts and wiring strategies in order to find the best way to reduce its radio frequency interference. Below are some tips that I discovered for reducing RFI from these lamps.

If the lamp is wired and grounded properly, the amount of RFI is only slightly higher than that produced by a regular compact fluorescent (CF) lamp. A properly wired lamp with fluorescent ballast should not produce any detectable buzzing more than 10 feet away on any frequency (at least above 150 kHz, which was the lowest frequency I tested). On the other hand, if wired improperly, buzzing can be heard on a wide range of frequencies, including the MW region as high as 1000 kHz, and will easily interfere with radio reception 50 feet or farther away. If you use one of these lamps, aliens from outer space will visit you and complain about interference in their flying saucers.

Fluorescent desk lamp
This is the lamp used in the tests. The original fixture was all metal resembling stainless steel, except for the base. The original base was a hollow, cross-shaped piece of cheap plastic with a fake metal coating. Inside the plastic base was an unencased electronic ballast. This design resulted in large amounts of RFI. The lightweight plastic base also made the lamp very unstable. These lamps, which were sold in hardware stores, were only on the market for a short time. I replaced the plastic base with a 8×8×3/8" steel plate and mounted a new ballast, enclosed in a steel case, on the back. The wires travel under the steel plate (which is on rubber feet) and up through the tubing. This design provides a much sturdier support than the original base and greatly reduces the RFI.

Findings

  1. Electronic ballasts are very unstable. They may produce interference at different frequencies depending on their surroundings. They also burn out easily, making a distinctive high-pitched "plink" sound.
  2. Enclosing the ballast in a small steel case is essential to preventing RFI. Inside the case, a cardboard insulator is typically used to prevent short-circuits. Cheap ballasts consisting of a bare circuit board should be avoided. Better ballasts are enclosed in a steel case and are sometimes potted with black polymer.
  3. Select a ballast that has low total harmonic distortion (THD). Low THD means higher efficiency and fewer RF emissions. Currently, good ballasts are quoted as having "THD <10%". Keystone KTEB-213 ballasts are highly recommended.
  4. Some ballasts have only two wires (white and black, but no green). On these ballasts a ground point is present on the circuit board, but it's not used. These ballasts can be used with two-conductor plugs. Attaching a ground to the board will result in RFI being conducted to the house wiring. For these ballasts, only the metal case and the fixture itself need to be grounded.
  5. If the lamp fixture is metal, attach the ballast to the lamp. Permanently mounted metal lamp fixtures must be grounded to meet code. If it's a portable lamp, the ballast can be up to 1 foot away from the lamp. However, if it's not attached to the lamp, it's also necessary to worry about heat dissipation in the ballast. Contrary to expectation, very little RF energy was actually radiated from the wires leading from the ballast to the fluorescent tube unless the wires were excessively long (i.e., over six inches or so).
  6. Ferrites on the output lines or on the power cable were of little or no benefit. Shielding the output lines (braided shielded wire, copper tape, etc.) also had no benefit. The reason for this seems to be that most of the RF radiation comes from the ballast and the light bulb itself, and relatively little from the wires. I also used a shielded three-wire power cord.
  7. If the wires travel through the lamp fixture, and the lamp has two sides (like the desk lamp shown above), run all four wires through one side. The wires on both sides of the fluorescent tube should be the same length.
  8. If you have a choice, avoid three-wire ballasts. (These may have, for example, two red wires that connect to one side of the fluorescent tube and one yellow wire which connects to the other side.) Four-wire ballasts, which have four wires, (e.g., two red wires and two blue wires), are often superior. Although three-wire ballasts are not necessarily noisier, more care is needed when installing them. Under certain conditions they can induce a potential on the metal parts of the lamp up to 600 volts. (That is to say, I observed this using one three-wire ballast, which did not appear to be otherwise defective). A potential of 600 volts is easily detected if the lamp is touched by the hand (note: not the recommended way of testing it). Using three-wire ballasts would require scrupulous grounding of the lamp for safety reasons. This is probably a good place to mention that electronic ballasts put out over 600 volts and around 20 milliamps of current; be careful out there. Ballasts intended for two bulbs may have five or more wires (in addition to the white and black wires for the line connection), and are provided with a wiring diagram. Of course, you also need to make sure the ballast, whichever type it is, is designed for the tubes you're using.
  9. For wiring the ballast to a single 13 watt fluorescent tube, the wire must be insulated to withstand 1000 volts.
  10. Keep all power connections and switches as far away as possible from the output connections. This will reduce RF coupling to the house wiring.
  11. It was necessary to perform all testing of the lamp for radio frequency interference during the daytime, since my neighbors do not care about the interference from their own lamps. The interference from neighbors (which blankets the entire spectrum up to 600 kHz from about 4PM to midnight) will usually drown out any RFI from the lamp you are testing, making it impossible to determine whether you are reducing the noise from your lamp.

Disclaimer (i.e., the don't sue me section)

If you rewire a lamp, be sure to use fixture wire, which has insulation that can handle the higher temperatures, and not regular hook-up wire. Always follow the electrical code when wiring lamps. Check for induced potentials and shorts with a voltmeter before using the lamp. Make sure the ballast is rated for the specific bulb you're using. People have been electrocuted by using the wrong fluorescent bulbs (the ballast can overheat and melt the insulation).

RF Spectra of fluorescent lamps

From the figure below, it is evident that different types of electronic ballasts have markedly different RF spectra. These lamps were measured with a low-gain antenna one foot away from the ballast, and the computer monitor turned off. There is a certain amount of ambient noise that I didn't bother to try to eliminate.

Lutron dimming fluorescent ballasts emit their energy in a sharp band with harmonics extending above 500 kHz into the AM radio band. The Tu-Wire was the worst, with peaks up to -70 dB in my setup. When the lamp is dimmed, a small amount of broadband noise is also evident between the major peaks. Non-dimming ballasts, such as the ballast in the GE F21T5 fixture shown here, tend to have broader bands, with markedly reduced intensity above the third harmonic. The sixth harmonic (around 320 kHz) is already down to -120 decibels. This indicates that the ballast in this lamp has a relatively low THD.

Fluorescents with magnetic ballasts also produce small amounts of radio frequency interference over a very broad spectrum, but the levels are very low compared to electronic ballasts (the y axis in the figure is in decibels, which is a logarithmic scale). Light dimmers also produce broadband interference. The light dimmer in this case was farther away, about ten feet from the antenna.
RF spectra of fluorescent lamps

Update

I recently came across another one of these lamps. They are back! They are now using a different ballast that produces much less radio interference than before. The inside of the plastic base also has a conductive metallic coating, presumably to reduce RFI, but the lamp itself is not grounded. Unfortunately, on the lamp I tested, when the lamp was turned on, all the metal parts were about 90 volts AC with respect to the ground. It's easy to feel this voltage with your hand when adjusting the lamp. Using a multimeter on its 20 amp setting, I also found a measurable amount of current. There were no shorts in the lamp. Most likely it is only induced RF from the new ballast. But there's an awful lot of it, and I would be hesitant to use this lamp in an area where there are grounded objects, like water pipes or radiators. If you use this lamp, I recommend replacing the plug with a three-conductor plug to drain it off.
Back