RFI ingress from the speaker cable is an area that deserves a lot more attention and work. I would feel tempted to stick a small ferrite in series with the feedback network, on the amplifier output side of the network, to try to solve this. Another approach might be to split the f/back resistor and use a good quality cap to gound to try to limit the HF.
Any other thoughts on this?
I had a Marantz 240 that used to pick up SSB transmissions. After a while it blew up (output devices went short). I always suspected it was the RF pickup that caused it.
Any other thoughts on this?
I had a Marantz 240 that used to pick up SSB transmissions. After a while it blew up (output devices went short). I always suspected it was the RF pickup that caused it.
john curl said:
Hi John,
IIM is definitely real. We have no disagreement on that. Just read my paper on IIM on my web site at www.cordellaudio.com.
Once again, Otala is to be commended for coming up with a measurement technique to measure the IIM phenomena that he researched. I duplicated his measurement approach and made measurements to see for myself how IIM was influenced by the application of negative feedback. In agreement with the theory that I showed in the AES paper (also in a WW article), the measurements demonstrated that high amounts of NFB did not exacerbate IIM.
It was again a case of Otala drawing the wrong conclusions from his results.
Bear in mind here that I am referring to IIM as Otala defined it, wherein he drove an amplifier under test with 6 kHz and then back-drove that amplifier through a load resistor connected to another amplifier that supplied a 60 Hz large signal. This was like a SMPTE IM test where the 6 kHz is propogated forward through the amplifier under test, and the 60 Hz is back-driven into the amplifier.
The IM on the 6 kHz carrier at the output of the amplifier under test is then measured.
This test is, by the way, a good way to see how well the amplifier performs with fairly high output currents.
The Otals test and measurement method does not necessarily apply to the process of Interface Intermodulation distortion when it takes place with RF signals. That is a whole 'nother can of worms.
Cheers,
Bob
Perhaps I've missed something important in the ferrite discussion, but isn't there usually an air core inductor in series with the output in 99.99% of all amplifiers?
_-_-bear
PS. the added capacitance of RG-8 would not be a good thing hanging off most amplifiers, imho... now if you could defeat the capacitance and keep the RF shielding effect via some dynamic way of driving the shield... hmmm... bah too complex...
_-_-bear
PS. the added capacitance of RG-8 would not be a good thing hanging off most amplifiers, imho... now if you could defeat the capacitance and keep the RF shielding effect via some dynamic way of driving the shield... hmmm... bah too complex...
Rick,
Some times you can reduce the effect of RFI on an amplifier (line, power, whatever) by terminating the interconnect or speaker cable with something close to the characteristic impedance of the transmission line at both ends. That tends to minimize the voltage and current peaks because the VSWR is reduced on that line at RF.
Obviously, you may not want to place a 15 Ohm resistor across your speaker terminals if your estimated speaker cable impedance is 15 Ohms, so you use a "Zobel" like network instead. I know Rick knows what I mean, but for everyone else, that's a somewhat small cap in series with the 15 Ohm resistor so that the resistor is effectively *mostly* out of the circuit at audio, but not at RF. Many amplifiers already have a Zobel at the amplifier, so you might not need it there.
For everyone who reads this: This is not some magic termination for the speaker cable or interconnect so that the transmission line is terminated at audio frequencies. It's for RF where the cable length can be very significant in terms of the wavelength of the interfering signal. I suspect I'll get skewered anyway, but this might reduce that by 17% or so. thank you.
Some times you can reduce the effect of RFI on an amplifier (line, power, whatever) by terminating the interconnect or speaker cable with something close to the characteristic impedance of the transmission line at both ends. That tends to minimize the voltage and current peaks because the VSWR is reduced on that line at RF.
Obviously, you may not want to place a 15 Ohm resistor across your speaker terminals if your estimated speaker cable impedance is 15 Ohms, so you use a "Zobel" like network instead. I know Rick knows what I mean, but for everyone else, that's a somewhat small cap in series with the 15 Ohm resistor so that the resistor is effectively *mostly* out of the circuit at audio, but not at RF. Many amplifiers already have a Zobel at the amplifier, so you might not need it there.
For everyone who reads this: This is not some magic termination for the speaker cable or interconnect so that the transmission line is terminated at audio frequencies. It's for RF where the cable length can be very significant in terms of the wavelength of the interfering signal. I suspect I'll get skewered anyway, but this might reduce that by 17% or so. thank you.
Air -core inductors don't add much in the way of harmonics or distortion , that is why they are used in RF applications. After that inductor at the amps output there is the typical large woofer LP inductor , with its large ferrite slug , prone to saturation.
(Better speakers use a HUGE air core for this LP)
As far as a small ferrite bead at NFB , I've seen beads on mosfet gates and air core coils in parallel with basestoppers ,so bonsai's "tradeoffs"are being used in commercial designs , just implemented a little differently .
OS
How does one put a "termination" (characteristic impedance) at the end of a speaker cable? There is a speaker sitting there, with a crossover (usually) and a non resistive impedance from the driver?
Then there is the question of the characteristic impedance of a speaker cable... John Dunlavy had some patents on speaker cables with a nominal "8 ohm" nominal impedance, but in reality they were somewhat capacitive...
RG-8 is 50 ohm fwiw.
_-_-bear
Then there is the question of the characteristic impedance of a speaker cable... John Dunlavy had some patents on speaker cables with a nominal "8 ohm" nominal impedance, but in reality they were somewhat capacitive...
RG-8 is 50 ohm fwiw.
_-_-bear
- Status
- Not open for further replies.