Which ferrite cores are the best ones to stop GSM buzz in my speakers?

August 11, 2008 7:06 PM

Electricity and Magnetism geeks - help me! I would like to stop the GSM interference from my iPhone to my speakers with ferrite chokes. I imagine a bit of information would help me get the best ones...

So, it has become somewhat common knowledge at this point that gsm noise is irritating lots and lots of people. While the hack of cutting these ferrite beads out of USB cables is cute, and all, I figured I'd order just the parts I need from digikey, where I've bought some electronics stuff in the past.

The issue is -- these little buggers come in a variety of sizes, with a variety of impedance levels and frequencies (frequencies? what?). I would link to a DigiKey search results page, but they seem to use POST in their forms, so I can't get you a URL with search results....

First of all, I'm pretty sure GSM is 850mhz with AT&T in the united states. All of the ferrite core snaps I've found on DigiKey have a "Frequency" of either 100mhz or 300mhz ... how can a little ferrite bead have a "frequency"?

Their impedance levels vary wildly -- from 35 ohms to as high as 572 ohms.

So, my question: If we know what frequency AT&T's GSM is that's causing my interference issues with my stereo speakers -- how do I choose the ferrite cores that will be most effective in stopping it? Does it even matter which ones I get?

They're all reasonably priced, ranging between about $0.95 and $2.50... and since I'm only buying maybe 6 or 8 of the things, I'm more than happy to shell out the extra bucks if it means less noise getting through.

Thanks!

posted by twiggy to technology (12 answers total)

Many of the ferrite beads I've looked at actually have a range of frequency choking. They don't generally indicate the inductance. I think you might be looking at the wrong type of component.

http://www.rpelectronics.com/Default.asp?Main=/English/OnlineCat.asp?Menu=/English/Content/Categories/CatM_66.asp%26Detail=/English/Content/Divisions/Div_66_240.asp
posted by robofunk at 7:58 PM on August 11, 2008

I could be wrong, but I don't think this approach will work. I think that what a ferrite bead does is to attenuate ("block" to a certain degree) the transmission of signals higher than a given frequency through the wire around which it is wrapped. The frequency at and above which the bead begins to significantly attenuate signals is probably what is meant by its rated frequency. The higher the bead's impedance, the more it will attenuate the signal.

Cell phones in the US operate in frequency bands around 800 (sometimes called 850 for political reasons) and 1900 MHz. These are obviously higher than the frequencies being sent from your sound card to your speakers, which, like the human hearing ability, range from approximately 16 Hz to 20 KHz.

This would be all well and good, except for the fact that it is not the 800 or 1900 MHz carrier from the phone that you are hearing in your speakers. You couldn't hear anything at these frequencies anyway. What you are hearing are pulses caused by the intermittent transmission inherent to the operation of GSM equipment.

In the GSM protocol, each radio channel is divided into frames, each frame having a duration of approximately 4.615 ms. Each frame is then divided into eight "burst periods," better known as timeslices. During regular voice (and single-timeslice data) operation, a given phone only transmits during one of the eight timeslices in each frame. Other phones on the same channel in the area transmit during the other timeslices, thereby allowing eight phones to share a single radio channel.

The upshot of this is that a GSM phone, when in use, starts and stops transmitting once every 4.615 ms, or every 0.004615 seconds. 1/0.004615 is equal to approximately 217, giving a frequency of 217 Hz. I've never attempted to measure it, but I've heard a lot of GSM interference over the years, and I'd say ~200 Hz sounds about right for the base tone of what one hears in speakers being interfered with. (The cadence of the interference tone varies depending, among other things, upon whether the phone is transmitting during its entire assigned timeslice or only a portion thereof.)

I'm certainly no electrical engineer, but I can't think of any easy and reliable way to remove this type of interference from a cable, and especially not any way using a simple LRC type circuit. However, if the interference is being received by the cable itself (as opposed to by components in your sound card or speakers), I wonder if wrapping the speaker cables in a conductive metal foil might help. "Real" aluminum foil (as opposed to metallized film) might work for this, although I suppose it might be somewhat ugly to have foil-wrapped speaker cables running about.
posted by Juffo-Wup at 8:21 PM on August 11, 2008

Honestly, twiggy, I don't know anything about GSM noise specifics. Generally, however, if you are going to solve an electromagnetic interference (EMI) problem, it is awfully helpful to have the noise source characterized with some sort of visualization hardware. In this case, a spectrum analyzer would be best, followed by a really good scope if you can't get one of those. Either is a helluva lot more costly than a bag full of crap from Digikey. (My spectrum analyzer cost $75,000 new.)

The noise that you hear from RF can be an artifact of the actual interference but you can effectively treat it by addressing the RF frequencies or the audible noise frequencies.

If you are dealing with this 'blind', hacking is as good a method as any. If you can get a spectrum analyzer that will show you the details of the noise you are seeing (either in the RF or the audio range), then you can start being more scientific. Until then, most anything you do will be trial and error. (That's OK, by the way.)

EMI is a bitch. I've seen RFI treated by putting a rock in a mixer console at a radio station, so help me Spaghetti Monster. (ramen)
posted by FauxScot at 8:22 PM on August 11, 2008

Wow.. amazingly interesting answers so far.. I feel like I've learned a lot even though I haven't gotten the answer of which ones to use (understandably so, given what I've just learned!)...

It's clear from posts all over the "it must be true" interwebs that people are really helping dampen the obnoxious noise coming from their GSM-based cell phones that's causing their speakers to go nuts. I do understand that what you're hearing isn't 800mhz or 1900mhz -- but it is phones that operate on GSM frequencies that are causing speakers to "chirp and buzz" obnoxiously when they're talking to the cell towers.

For example, when I get an incoming call, my speakers go absolutely nuts a couple of seconds before my phone actually rings. This did not happen with my Sprint phone.

What you are hearing are pulses caused by the intermittent transmission inherent to the operation of GSM equipment.

That much I understand. This makes sense.

However, if the interference is being received by the cable itself (as opposed to by components in your sound card or speakers), I wonder if wrapping the speaker cables in a conductive metal foil might help. "Real" aluminum foil (as opposed to metallized film) might work for this, although I suppose it might be somewhat ugly to have foil-wrapped speaker cables running about.

That's basically exactly what the ferrite cores are for. If you check out the MacLife article linked in my post, that's what people are using them for - they're cutting the little ferrite cores out of unused USB cables and putting them around their speaker wire to stop whatever interference is being picked up by that wire from making it to the speakers...

Any idea if it even matters which ferrite core would be best for this? Or how to "calculate" that answer?
posted by twiggy at 8:39 PM on August 11, 2008

Right, well, like I say, I'm not an EE, so don't let me discourage you from experimenting. Even if I'm "right," you never know what you might come up with when it comes to RF interference via more-or-less random experimentation like FauxScot said. However:

I do understand that what you're hearing isn't 800mhz or 1900mhz -- but it is phones that operate on GSM frequencies that are causing speakers to "chirp and buzz" obnoxiously when they're talking to the cell towers.

Yes, but GSM and CDMA (technically, "CDMA2000," in this sense) use the exact same frequencies in the United States. AT&T and Verizon use both 800 and 1900 MHz, and Sprint and T-Mobile use 1900 MHz only. But, you won't ever hear this type of an interference from a Sprint or Verizon phone for the reason discussed next:

For example, when I get an incoming call, my speakers go absolutely nuts a couple of seconds before my phone actually rings. This did not happen with my Sprint phone.

That's because the network and your phone are communicating with each other to set up the call, and it takes a couple seconds for this to happen before your phone starts ringing. Sprint phones do this too, but you won't hear any interference because the CDMA protocol used by Sprint and Verizon does not cycle the phone's transmitter on and off while the phone is in use. A CDMA phone will emit a continuous 800 or 1900 MHz carrier the entire time it is in (or setting up) a call instead of turning its transmitter on and off a couple hundred times a second as a GSM phone does. Because of this, CDMA phones don't generate interference in the ~200 Hz range that can be picked up by speaker wire & etc.

Any idea if it even matters which ferrite core would be best for this? Or how to "calculate" that answer?

Yeah, you would want to use a ferrite core that blocks (attenuates) signals around the frequency that you want blocked, which in this case seems to be approximately 200 Hz. The two problems with this are that I'm not sure it's possible to make a ferrite core that effectively attenuates 200 Hz signals and even if you could, your music would really suck if all the ~200 Hz tones were missing from it.

It's clear from posts all over the "it must be true" interwebs that people are really helping dampen the obnoxious noise coming from their GSM-based cell phones that's causing their speakers to go nuts.

I don't disagree with this at all. I chalk it up to what FauxScot said about trial and error. I'm amazed at what a huge change the tiniest difference in the position of either the phone or the receiving wires can make in the amount of interference heard in the speakers. I've been in situations where I was purposely trying to obtain the largest amount of audible interference possible from a GSM phone, and just a couple inches' difference in the position of the phone or the shape of the wire can literally make the difference between loud interference and none at all. Of course that doesn't give you a very clear roadmap of what exactly to start trying, but you might very well be able to mitigate the problem by moving things around -- whether the phone, the wires, or other objects in the vicinity of the phone and the wires.
posted by Juffo-Wup at 9:11 PM on August 11, 2008

My personal experience with this, as a long time user of TDMA/GSM cell phones, is that it's more likely interference in the amplifier itself, rather than the wires. It is impossible for me to cause interference in my speakers when the amplifier is off. There's just not enough energy to move the cones significantly.

So shield/move/reorient your amplifier, not your wires, especially not your speaker wires.
posted by wierdo at 10:34 PM on August 11, 2008

Oh, or buy an iPhone 3G, if you're in an area where at&t has 3G service. ;)
posted by wierdo at 10:36 PM on August 11, 2008

Wierdo: My iPhone is an iPhone 3g, actually. Although in my condo, reception is weak because the building is pretty thick concrete, and I don't get 3G signal in here.
posted by twiggy at 10:43 PM on August 11, 2008

I just wanted to point out that most of what Juffo-Wup says is true, but the part about interfering with your music was incorrect. Wrapping the wires in anything to reduce the electromagnetic interference will help, however the music is flowing through the wire, and any emi protection around the wire will not affect it. Although if you put a circuit on the wire to filter out stuff around 200 Hz, then you would have crappy music as well.
Also, as others have pointed out, it's the amplifier more than the wires.

Another technique you can try to mitigate the interference is to keep your phone next to a window in your condo, you may have to experiment which location works best for you. This does two things, it provides distance between your phone and the stereo equipment (lessening the interference), and hopefully the phone itself will have improved coverage so it can transmit at a lower power, also reducing the interference.
posted by forforf at 8:05 AM on August 12, 2008

Thanks for all the info everybody... Perhaps I'll spare some extra time and money and buy a variety of the little guys from DigiKey -- For Science(tm), although I'm leaning toward just figuring out what the common ones are in USB cables and buying about 8 of those for personal use to save on the time aspect of buying a variety of them, testing, etc, then ordering a bunch of the best ones since only one or two of them won't be useful to me (I've got 5 speakers in my room to shield!)...
posted by twiggy at 10:08 AM on August 12, 2008

This is purely an academic question, since AT&T and T-Mobile USA use UMTS-FDD, but I wonder if the transitions between Tx and Rx modes in UMTS-TDD cause any type of audible interference... :)
posted by Juffo-Wup at 1:01 PM on August 12, 2008

It is my understanding that if you have a phone with any carrier that is utilizing GSM, you are going to get &!@(*#&!@(* interference from them. I could not believe that the FCC (if that's who's responsible) would let this go on during my week long attempt with trying just about every cell phone carrier there is. Apparently people do not give enough of a crap about it to bother. Isn't GSM what all of Europe uses?

I once heard somebody say "hey, if it wasn't for that preemptive noise, my band would never know we were getting an incoming cell phone call."

Uhh......

I refuse to own a GSM phone as long as I work in audio. Guess I could just keep it turned off the whole time. Stupid.
posted by bitterkitten at 3:30 PM on August 12, 2008