Reducing RFI at Home From Lighting Fixtures?

amper asks: "I'm in the process of building a new home recording studio. When I originally moved into my new (very old) house, I decided that in the interests of conserving energy, I would replace most of the incandescent lighting fixtures or lamps in my home with fluorescent fixtures or compact fluorescent replacement lamps in those fixtures which could not easily be replaced. Unfortunately, these fixtures are creating a massive amount of radio frequency interference in my home. The worst culprits seem to be the dimmable fluorescent fixtures in my living room. Barring replacing all my fixtures and lamps with conventional incandescents, can anyone point me in the direction of alternatives? Is it possible that the decreasing quality of most home goods has led to a decreasing quality in fluorescent ballast systems that are much more noisy from an RFI standpoint? Some of these fluo's are so noisy, they even emit audible sound! It's gotten so bad that I can't even play an electric guitar without turning off all the non-incandescent lighting in my house, which pretty much limits me to playing and recording during daylight hours (when I'm supposed to be out making money)."

Comments

Focus on ground loops, not on RFI sources.

[hankwang]

I come from a physical laboratory where we have to measure small signals with an acceptably low noise despite the presence of laser flash lamps, or even high-voltage discharges in more or less open air. The best way to avoid RF interference is not to remove all RF sources, but to design the electronic circuits correctly.

The main thing that makes electronic circuits sensitive to noise is ground loops. Often, signals travel through cables that have two wires or a central wire with a shield surrounding it. Normally, equipment (whether it is an oscilloscope or consumer-grade audio equipment) has a common ground, which means that the neutral wire of each and every input and output is connected. If you have more than two pieces of equipment interconnected, it is likely that there are loops in the ground wire, for example the cable from mixer to some effect generator, and the wire back. All these loops acts as antennas that can pick up noise. Having shielded cables doesn't help because it is the shield, that acts as ground wire, that causes the problems.

The first and simple step is to have all wires bundled as close together as possible, such that the area inside the loops is as small as possible. The next step is to upgrade your equipment to stuff that has balanced inputs, with those big XLR connectors. Here the shields are really shields against RF interference, while the signal is carried by two wires inside the shielded cable. Balanced signals means roughly that the equipment measures the signals on the two signal wires completely independent from the ground.

Re:Focus on ground loops, not on RFI sources.

[rhild]

To expand on that, any RF noise that gets into the cable will appear in phase on both signal conductors, which will result in it being cancelled out in the input circuit of the device the cable is feeding, if the device has balanced inputs.

CMRR (Common Mode Rejection Ratio) is a measure of how well an input circuit cancels out this kind of noise. When looking at the specs for gear this is one of the important but often overlooked numbers to check out.

Re:Focus on ground loops, not on RFI sources.

[unitron]

The average electric guitar has a high-impedance unbalanced (one "hot" one ground) output and the average guitar amp the same kind of input, and the cord in between has a center conductor for the "hot" side and the shield around that is the ground.

If the OP would care to email me at coastalnet.com I'll explain to him how his guitar can be fairly easily modified to greatly reduce the interference which it picks up, which may or may not be due to his lighting.

Light dimmers, incandescent or flourescent, are notorious sources of RF unless you get very special expensive ones.

RFI Solutions

[Ed Almos]

As a radio ham I have similar problems, try the following.

1) You can get ferrite rings of various diameters. Try winding a couple of turns of the cord leading to the light through one of these rings. The ring should be as close as possible to the fixture.

2) Wire RF chokes in series with the fittings. These work in a similar way to the ferrite rings but are more effective.

3) It's possible to obtain capacitors that are rated for 110/220 volt operation. Have a look inside a PC power supply and you may see one, they are normally encased in yellow transparent epoxy. Wire one of these across the fitting.

And finally.........

4) Ask for help from the right people. Try and find either a local radio ham or (even better) a local ham club. These guys will be experts at fighting RFI and will help you out.

Ed Almos
Budapest, Hungary

Ditch the dimmers.

[david.given]

Trying to dim fluorescent lamps is a really problematic process; see here for a discussion. You can't dim them by reducing the voltage, and have to pulse the power supply instead. This will, as you have discovered, interact oddly with the ballast and the tube itself and is probably what's making the bulk of your radio noise.

(The article mentions a number of problems, including the fact that you might be significantly reducing the lifetime of your tubes by dimming them. YMMV.)

What I have in my living room are some long-live fluorescents to provide the bulk of the light, and some incandescents to change the colour and make the illumination more interesting. You might be able to do something similar without having to hook the tubes themselves up to the dimmer.

How about Faraday Cages?

[VernonNemitz]

Get some wire mesh and surround each light fixture with it, and "ground" the mesh. The size of the mesh-holes will depend on the frequency of the interference (higher freq -> smaller mesh). Chicken wire is fine for some things (60-cycle AC hum) and window-screen mesh might be needed for others (but associated with lessened lighting, alas). If you can find mesh with holes as large as those in a microwave-oven door, that will probably be more than small-enough. (Of course, if this is really a good idea, then I expect some comment from some expert who KNOWS exactly what size of mesh is sufficient. :)

Re:How about Faraday Cages?

[DeComposer]

As I recall, flourescent light ballasts double the line frequency (presumably using a tuned resonant filter--I'd guess an LC tank) to 120 Hz (or perhaps some other multiple of 60 Hz). BTW, Electromagnetic interference in this frequency range is not in the radio frequency (RF) range (ELF transmissions notwithstanding); its correct designation is "EMI".</nit>

Now for the bad news:
a surprisingly vast amount of professional audio gear uses 1/4" unbalanced connectors. Of the synths, samplers, and effects processors in my studio, maybe a quarter of them have balanced outputs. (The mixers, of course, have balanced line and balanced mic inputs and outputs). This means I have to spend considerable effort routing cables to minimize 60 Hz inductance from power cables. The inverse-square law applies in this case: the farther audio lines are from power lines, the better 60 Hz line noise is rejected. In situations where proximity between the two is unavoidable, I try to cross them at 90-degree angles to minimize inductance.

Getting back to your question...
For myself, a music studio is primarily an auditory experience; I don't want a lot of bright light. Reflections on device displays or monitors are distracting and detract from my enjoyment of working in the studio. Plus, when I have other musicians or clients in the studio, I want them to focus on the music, not their visual surroundings. My solution is to use low power (25-40W) incandescent fixtures, usually aimed at walls or ceilings to provide soft, indirect lighting. (If I need to work from a score or other paper material, I'll us a task light.) This solution provides more than adequate light to find my way around in the studio without running into things and creates a warm working environment, all without sucking down huge gobs of electricity.

Incidentally, I have sometimes wondered if tight twisted-pair cables would help with EMI rejection for audio the way it does for data transmission (i.e. ethernet over CAT5). I welcome the thoughts of any EEs out there with some experience in this area.

Re:Flourescent lighing

[DAldredge]

he following introduction is from a publication by the National Electrical Manufacturer's Association (NEMA) entitled "Fluorescent Lamps and the Environment". This publication and a lot more valuable information can be found at the NEMA website devoted to proper lamp disposal called www.lamprecycle.org. An Adobe Acrobat PDF version of the publication can be downloaded from www.nema.org/lamprecycle/nemafluorfinal.pdf

If you need more information about properly disposing of lighting waste click on the link at the left for "lamp disposal rules".

Why Do Lamps Need Mercury?

"Mercury is an essential ingredient for most energy efficient lamps. Fluorescent lamps and high intensity discharge (HID) lamps are the two most common types of lamps that utilize mercury. Fluorescent lamps provide lighting for most schools, office buildings, and stores. HID lamps, which include mercury-vapor, metal halide, and high- pressure sodium lamps, are used for street lights, floodlights, and industrial lighting. A typical fluorescent lamp is composed of a phosphor coated glass tube with electrodes located at either end. The tube contains mercury, of which only a very small amount is in vapor form. When a voltage is applied, the electrodes energize the mercury vapor, causing it to emit ultraviolet (UV) energy. The phosphor coating absorbs the UV energy, causing the phosphor to fluoresce and emit visible light. Without the mercury vapor to produce UV energy, there would be no light. A four-foot fluorescent lamp has an average rated life of at least 20,000 hours. To achieve this long life, lamps must contain a specific quantity of mercury. The amount of mercury required is very small, typically measured in milligrams, and varies by lamp type, date of manufacture, manufacturing plant, and manufacturer."

"If lamp life is shortened, more lamps must be purchased to achieve the same length of service, and the number of lamps that generators must dispose will increase. How Much Mercury Do Lamps Contain? Based on a 1999 NEMA survey, the average four-foot fluorescent lamp contains about 11.6 milligrams (mg) of mercury. This number has been steadily declining as lamp manufacturers work to reduce mercury content to the minimum amount technically feasible without reducing lamp life. The average four-foot lamp today contains over 75% less mercury than the same lamp would have contained in 1985. According to the U.S. EPA, total global natural and manmade emissions to the environment are 5,500 tons. 1 Manmade sources in the U.S. released 158 tons of mercury in 1995. For comparison, all of the lamps sold in the United States in 1999 containonly an estimated 13 tons of mercury, of which only a fraction will be released as an air emission. Proper lamp disposal or recycling will result in keeping almost all of this mercury out of the environment."

I guess you've never toted them up.

[Engineer-Poet]

I use a bunch of circle-tube CF's installed in table lamps. They are rated at roughly the same light output as a 150 W incandescent but they use only 34 watts on full (they're 3-way). They will run at least a year between tube replacements and these units cost about $20 new. Figuring 2000 hours of operation @ 34 watts and 8 cents/KWH and $20 up-front, the total cost for those 2000 hours is $25.44. That's assuming that you have to throw away the whole lamp after 2000 hours instead of just replacing the circle tube; if you replace the tube every 2000 hours at $5 each the cost for each incremental 2000 hours goes down to $10.44.

Lifespan of a 150-watt incandescent is what, 200 hours or less? Figuring 10 bulbs at $1 each plus 300 KWH of electricity @ 8 cents, the same 2000 hours of light would cost $34. Looks like your approach is penny-wise, pound-foolish.

Re:I guess you've never toted them up.

[Engineer-Poet]

I just replaced a tube over the weekend (2-3 years on that tube) and checked the new GE 3-ways at the hardware store. They claim 10,000 hours on a tube for about $25 (not a big-box store, prices are higher). The 150-watt 3-way incandescents across the aisle were $2.29 and claimed 2000 hours. Figuring 33 watts for the GE going flat-out, over 10,000 hours it will use 330 KWH at a cost of $26.40 for a total cost of $51.40 sans taxes. The 5 incandescents would cost $11.50 in toto and use 1500 KWH at a cost of $120 for a total of $131.50.

It looks like the CF saves about $80 on the first tube and, at a replacement cost of $9.00 per tube, $96 on each subsequent tube. Not bad, eh? Oh, these are 3-way ballasts, a simple one-way on/off would be about $10 cheaper these days.

Buy an online UPS

[ponos]

I don't remember how much these things cost but an online UPS (one that converts AC -> DC feeds the battery and then feeds the PC/equipment from the battery) can reduce line noise significantly. It isn't the most elegant solution but feeding pc/sound card/sound equipment from a noiseless source can improve things considerably and it is very easy to install. APC has a specific solution for high-end audio/video systems that appears perfect for you (its called a power conditioner and is, essentially, an online UPS with very little battery time). Caution: the words "high performance AV system" usually mean $$$$$$. Check the link http://www.apcc.com/products/family/index.cfm?id=3 10&ISOCountryCode=US

Alternatively you may try balanced XLR->XLR cables that are used by professionals. I use these for my microphones (which provide a very low signal) and I am quite happy. Note that 4.5mm jack can also be balanced but you have to explicitly request it. Also bear in mind that a strong signal is much less sensitive. Long lengths of microvoltage-level signal are a not a good idea but line-level (-5/+5) signal is quite resistant. You may also buy ferrite cores for your cables (even power cables that have ferrite cores pre-installed!) and see if it makes a difference. Shielded cable may also be useful. Don't go spending a lot before trying.

Finally, always remember: a very good power supply (as in expensive audio equipment) can handle noise quite effectively.

P.

MANY ideas for you!

[Anonymous Coward]

If you use one single-wound pickup, of you will have hum from any RFI/EMI in your vicinity.

But there are alternatives... "Humbuckers" for example are like two pickups right next to each other wired in such a way as to cancel the hum.

(Personally I don't play music that sounds good on humbuckers... but YMMV.) There are other options too.

If you use a G&L "Z" pickup, these are like two half-pickups for three strings each, and they're wired in a humbucking configuration. These are an excellent design, IMHO.

Another great option is to use two single pickups and have them wired in a humbucking configuration. My G&L ASAT 2000 (basically a Telecaster) has a pickup switch with three positions. I can select Front, Rear, or Both pickups wired together in a humbucking configuration... which is the configuration I usually play in. It gives the right amount of treble from the rear pickup with all of the rich tones from the forward pickup, without any RF hum, and with a far richer tone than could be generated with your typical humbuckers.

If your current guitar doesn't provide this option, you could talk to a repair shop or if you're up to it, maybe get out your soldering iron... (You always wanted a custom guitar, right?)

In addition to the above, you should try putting your noisy lamps on power strips that have RF chokes in them! Some power strips advertise EMI/RFI filtering.... those are the kind to get.

If that doesn't help, try winding the cord of each of your lamps into loops of say, 3 or 6 inches in diameter. Clamp a ferrite bead around one part of the ring.

Okay now for the serious voodoo... and don't laugh until you've tried this, because it actually can work. I have used the following trick many times:

Get a long extension cord and plug the lamp or lamps into the cord. Wind the extension cord into a circle on the floor maybe 2 or 3 feet in diameter. Experiment with increasing or decreasing the number of turns in the coil, and also with the direction of the winding (flip the coil over.) You should find that this has an effect on the amount of hum which reaches your guitar's pickup. Through trial and error, you may be able to reduce it substantially by this method. Basically what you're doing is creating a variable RF choke with the extension cord. It works on microphone cables too.

By combining the above methods, you chould be able to win the battle against EMI/RFI without having to construct a Faraday cage to play in. ;-)

Best of luck!!
-The Ruggy