Tweakers' Asylum

Posted by Jon Risch (B) on May 09, 2005 at 17:56:47

In Reply to: RC network(zobel) Radio Frequency speaker filter question?(rather long) posted by Keith Lockwood on May 9, 2005 at 12:23:41:

The resistor value is not critical, as it is there to prevent a dead short at the very high frequencies, which most power amps would not like to see, even if they do not have any nominal response up that high.

A value that is around the load impedance is probably best, as it will continue the nominal impedance up into the very high frequencies. Thus a range of 8-10 ohms is the usual recommendation. This needs to be of a non-inductiove type for RF issues, a typical wirewound will not do very much good. Look for some of the flame-proof metal oxide types, which come in sizes up to 1 or 2 W. These usually have a decently low inductance.

The cap value would normally want to be the largest one you could use that did not intrude into the audio band. Realize that if you used a honkin big cap, that the voltage divider action of the RC would create an approx. 6 dB attenuation above the range where the cap kicked in, so even what would lok like a "big" cap, is not going to cause that much intrusion into the audio band.

However, the downside of a larger value cap is the lower range of operation. It is generally considered that a cap is no longer effective above it's self-resonant frequency. This can be very high for physically smaller caps, which are also usually not a large capacitance value, but it can get kind of low for the 'larger' capacitance values.

With this in mind, it is possible to some extent to 'tune' the capacitance value to be more effective for a given range of RF, within certain limits. Use the smallest value (and physically) for the higher RF regions, and a medium value for AM radio, etc.

For speaker cable stability, I normally recommend a 10 ohm ressitor and a 0.1 uF cap, but this is intended to stabilize a power amp output, not help terminate for RF pick-up. A smaller value cap would be more efective for RF situations, and even a cascaded set of different sizes, say a 0.001 uF and a 0.001 uF paralleled. Remember to keep the leads as short as possible, and use the smaller value cap as the "main" leads, and piggyback the larger cap onto the smaller cap, so it's leads are the longer of the two.

For 4 ohm s[peakers, halve the R values, and double the C values.

If one set at the speaker does not do the job, double up with another set, similar or different, at the amp.

One further tweak, you could try a compound RC pair, using the example above, use a 10-15 ohm resistor for the 0.01uF cap, and a 3.3 to 4.7 ohm resistor for the 0.001 uF cap. The smaller cap value is not kicking in until a very very high frequency, and so, the slightly lower R value is not as much of an issue, and can more effectively damp or reduce RF.

Note that in persistent or severe RFI situations, experimentation is the order of the day, and R values that are higher or lower may work best, and even trying a very small cap, say a 1,000 to 470 pF ceramic disc, might be effective without any R in series, but this ONLY AT THE SPEAKER.

For some related info, see:
Cable RF Termination and Cable Loading
http://www.audioasylum.com/audio/cables/messages/80110.html

Jon Risch