[Ham-Mac] Computer RFI

[Chuck Counselman]

On Feb 23, 2011, at 12:17 PM, Todd Jonz wrote:

> CT2IWW, Paulo, writes:
> 
>> ...can anyone provide a report on RFI caused by the new 11" MacBook Air?


I have no knowledge of the MacBook Air 11", but my experience with the aluminum-body PowerBook G4, 1 GHz, 15", may be relevant.

I found that almost all of the RF noise detectable at a distance greater than a few feet from my aluminum PowerBook was eliminated by putting a series of good common-mode chokes on the AC power cable between the PowerBook's external power supply (that white plastic rectangular block having dimensions 2 7/8 x 2 7/8 x 1 1/8 inches) and the wall outlet.  By "good," I mean that the real part of the impedance (the equivalent series resistance) of a choke, in common-mode and at all relevant radio frequencies, should exceed 1000 ohms.  How to make a good, yet inexpensive choke is described below.  By "almost all," I mean that the noise power spectral density in the lowest 25 kHz of the 3.5-MHz band was reduced by more than 30 dB.  Adding a series of good common-mode chokes to the DC power cable between the PowerBook's external power supply and the PowerBook itself reduced the noise further.

I demonstrated what the preceding paragraph describes in front of a meeting of the YCCC, using my aluminum PowerBook and an HF receiver whose antenna could be moved, coupled to the AC power line, and so on.  The receiver's audio output was connected to a loudspeaker and the gain was turned up.  The sound of the PowerBook RFI was intolerably loud and unspeakably annoying without the chokes.  With the chokes, no trace of RFI could be heard.  I mean RFI from my PowerBook and its power supply.  To show that I had not covertly disabled the receiver, I plugged a Wintel pee cee laptop and power-supply into the same wall outlet.  Hell's fury was unleashed.

Any other cable (telephone, Ethernet, USB, FireWire, external display, audio in, audio out, ...) connected to the computer should also be choked.  By "any other cable," I mean that _every_ other cable should be choked.  Very little RFI is radiated by a well-shielded computer; and an Apple PowerBook or MacBook with an all-metal body is well shielded.  Some other computers are not.  A cable is a special case worth discussing.  A peripheral or network cable, even if perfectly shielded, will radiate if either end is connected to a device that is imperfectly shielded, because displacement current flows from the device onto the outside of the cable shield and continues as conduction current.  The latter current radiates efficiently if the cable's length is sufficient.  Quite a short cable can radiate efficiently at HF if the cable is end-loaded by the capacitance of a big "box" such as a printer or a display.

Instead of choking the original line cord of my PowerBook's power supply, I made a short extension cord for it from a (male) plug, a (female) socket, and some electrically suitable but thinner and more flexible cable; and I threaded this cable through six pairs of ferrite beads, to form a series of six binocular-core chokes separated by intervals of a few inches along the extension cord.  The cable made three complete turns through each binocular core.  Each bead was one Fair-Rite Products part number 2631102002.  These beads are available from Newark for $1.65 apiece in quantity from 1 to 99, and from Mouser for $1.69 apiece for 10, $1.66 for 50, $1.64 for 100, ....  Google for 2631102002, and you'll find many other vendors.

A binocular-core choke comprising three turns of insulated, 0.2"-diameter cable on two p/n 2631102002 beads has common-mode resistance of 1000 ohms from 4 MHz through 29 MHz, by my own and others' properly calibrated measurements with AIM4170 and HP instruments.

Don't wind more than three turns on one binocular core, because the parallel capacitance of the winding would be excessive; it would bypass the choke and spoil its performance for the higher HF bands.  The reason for using a _series_ of several chokes is that one choke provides insufficient isolation/attenuation.  The series of chokes must be stretched out in a more-or-less straight line, not bunched or heaped together, in order to minimize coupling between the cables or other devices at the opposite ends of the choked section of cable.  Remember, you're dealing with RF!


73 de Chuck W1HIS