The worst RFI I ever experianced was when building up a rig to test and grade around 50 EL34 valves at a time. Grid stoppers had been left out of the circuit diagram (it was not my design but we all make mistakes). Temperature controlled soldering irons 100 yards away were going mental and chattering on and off and the radio a similar distance away lost all reception.
It is my experiance that RF usually can and will find it's way into almost anything by routes and methods that can seem most peculier to those not used to RF work. Decoupling capacitors can be resonent tuned circuits at certain frequencies as can even PCB tracks (they are often used as just that in microwave circuitry), amplifiers and power supplies that have V. low impedance up to say 100Khz can be anything but at 100Mhz. A friend of mine demonstrated to me just a few days ago that Volts of RF at one second intervals from a military missile early warning system were being picked up from just a few inches of wire!
Hifi Choice magazine used to show RF intermodulation results from Paul Miller Associates for all amplifier reviews (don't know if they still do), going up to the Ghz range, which caused an amp we had sent in for review to get a slating....we should have tried it in a microwave oven first maybe
It is my experiance that RF usually can and will find it's way into almost anything by routes and methods that can seem most peculier to those not used to RF work. Decoupling capacitors can be resonent tuned circuits at certain frequencies as can even PCB tracks (they are often used as just that in microwave circuitry), amplifiers and power supplies that have V. low impedance up to say 100Khz can be anything but at 100Mhz. A friend of mine demonstrated to me just a few days ago that Volts of RF at one second intervals from a military missile early warning system were being picked up from just a few inches of wire!
Hifi Choice magazine used to show RF intermodulation results from Paul Miller Associates for all amplifier reviews (don't know if they still do), going up to the Ghz range, which caused an amp we had sent in for review to get a slating....we should have tried it in a microwave oven first maybe
EMI ingress
This thread has taken an interesting turn with the discussion of possible power amplifier vulnerability to sonic degradation as a result of EMI ingress via the speaker cable antenna effect.
Let me first re-emphasize that although I think it is a concern and take measures to guard against it in my designs, I have never attemped to measure it. My bad. Although measurements are not the be-all and end-all of sonics, I think many of us agree that measurements are a good starting place. So maybe I and some of the rest of us should do some measurements on EMI ingress under various home audio conditions.
A starting point on measurements would be to just measure broadband noise appearing at the speaker terminals of an amplifier with a speaker cable and loudspeaker connected in some different home environments. It would be nice also to feed that into a spectrum analyzer.
Some of the posts here have argued that the numbers are really small and likely inconsequential, and they may be right. I don't know.
The other real point of curiosity that this raises has to do with interconnects, speaker cables and power cords. In many of the ads for those, we see claims that they improve the sound by keeping RFI out of your components. We all know that the RF spectrum out there is pretty polluted, but it would be instructive to do some of the sorts of measurements mentioned above at the inputs to, and the power supplies of, power amplifiers and other equipment. Again, this must be done with an open mind that measurements are not the be-all and end-all of sonics, but are a good starting point. Maybe some of you out there have done some of these kinds of measurements.
It is very plausible that some interconnects, speaker cables and power cords may really do a better job of keeping this junk out of the components, but it would be nice to know how big the problem is that we are trying to solve in the first place.
Along the same lines, even if rejection of RFI by high-end cables, interconnects and power cords is a big value-added by them, it would seem that the relative degree of this value-added would be very highly dependent on the design of the particular component. In other words, it should be clear by the discussions we have been having that some components are much more (or much less) vulnerable to EMI ingress at the inputs, outputs and power ports, either as a result of grounding practices, shielding, or circuit design.
By no means is this meant to be an anti-high-end cable rant, and please keep in mind that many interconnect, cable and power cord manufacturers claim other reasons than just EMI for the sonic improvement rendered by their cables.
Cheers,
Bob Cordell
This thread has taken an interesting turn with the discussion of possible power amplifier vulnerability to sonic degradation as a result of EMI ingress via the speaker cable antenna effect.
Let me first re-emphasize that although I think it is a concern and take measures to guard against it in my designs, I have never attemped to measure it. My bad. Although measurements are not the be-all and end-all of sonics, I think many of us agree that measurements are a good starting place. So maybe I and some of the rest of us should do some measurements on EMI ingress under various home audio conditions.
A starting point on measurements would be to just measure broadband noise appearing at the speaker terminals of an amplifier with a speaker cable and loudspeaker connected in some different home environments. It would be nice also to feed that into a spectrum analyzer.
Some of the posts here have argued that the numbers are really small and likely inconsequential, and they may be right. I don't know.
The other real point of curiosity that this raises has to do with interconnects, speaker cables and power cords. In many of the ads for those, we see claims that they improve the sound by keeping RFI out of your components. We all know that the RF spectrum out there is pretty polluted, but it would be instructive to do some of the sorts of measurements mentioned above at the inputs to, and the power supplies of, power amplifiers and other equipment. Again, this must be done with an open mind that measurements are not the be-all and end-all of sonics, but are a good starting point. Maybe some of you out there have done some of these kinds of measurements.
It is very plausible that some interconnects, speaker cables and power cords may really do a better job of keeping this junk out of the components, but it would be nice to know how big the problem is that we are trying to solve in the first place.
Along the same lines, even if rejection of RFI by high-end cables, interconnects and power cords is a big value-added by them, it would seem that the relative degree of this value-added would be very highly dependent on the design of the particular component. In other words, it should be clear by the discussions we have been having that some components are much more (or much less) vulnerable to EMI ingress at the inputs, outputs and power ports, either as a result of grounding practices, shielding, or circuit design.
By no means is this meant to be an anti-high-end cable rant, and please keep in mind that many interconnect, cable and power cord manufacturers claim other reasons than just EMI for the sonic improvement rendered by their cables.
Cheers,
Bob Cordell
Hi, Bob Cordell,
Testing with 1khz single tone will be boring
Is there any kind of measuring equipment nowdays (computized?) that can know the distortion artifacts when the test signal is real music, not a single tone? I predict the artifact will be interesting
PMA has measure up to Mhz region. His result is interesting
http://www.diyaudio.com/forums/showthread.php?postid=1012943#post1012943
Testing with 1khz single tone will be boring
Is there any kind of measuring equipment nowdays (computized?) that can know the distortion artifacts when the test signal is real music, not a single tone? I predict the artifact will be interesting
PMA has measure up to Mhz region. His result is interesting
http://www.diyaudio.com/forums/showthread.php?postid=1012943#post1012943
lumanauw said:
This is the effect I'm thinking of and when it's present it creates a fine sandy quality on voices and a background sort of glare. When it's not there is a bloom to the presentation. I find the input gain stages and the supply to be a more sensitive entry point than the speaker wires.
lumanauw said:
I still don't think the conditions are right for much noise to be present on the output, but the concept of creating a noise source (from Wavebourns post) and trying to measure it on the output sounds intrequeing. The trick would be in controlling the scope grounding so that you don't pickup the noise on it's ground and be fooled. It might work.
Bonsai said:
This is a good example of a problem to solve and a good test source for noise. Take the opportunity to investigate what knobs to turn to minimize the noise pickup. It is also a good way to play around with radiated interference, it's strength, range. This example is primarilly magnetic rather than an electric field.
I'd have to think about this since I'm still not convinced that this is an issue, but it would be an easy test to perform in my system. I can check the impedance of my cables on my HP 4263B and add the resistor and listen. Is this what you're suggesting?darkfenriz said:
MikeBettinger said:
That noise source was generating kilowatts of SSB signal on well tuned antenna. When they started contests there was the real playground for DIYers (nearly all of us were DIYers, it was an University of Automated Control Systems and Raio - Electronics). Nobody won. Neither DIYers, nor modifiers of industrial production. Everything was filtered-grounded-regraunded-shielded-reshielded. Nothing helped but jumming of their bands (once, when I was in a bad mood).
As I told before, looking at schematics may be enough to see how vulnerable the thing is. If output has followers, if that followers have naked base-emitter junctions open to RFI from speaker cables, like on the picture from Bob's article, and if a feedback loop feeds input of the amp by rectified on output emitters signal, it is the real kaput.
However, a lot depends also on a PCB layout and wire connections, but if points of vulnerability are visible well no measurements can help...
mikeks said:
I mean without low resistive component from bases to emitters (look at the picture from Bob's article, where you don't see A class in the driver). Pornographic? Probably, but I am innocent, it was not my work.
Ohhh, sorry: "Naked" is a jargonism of sabre fencers. "Your back is naked!"
"Personally, the worst I've seen recently is the hash comming from CFL's. I had one in my bench lamp which was pulled low over the amp I was working on and I had some strange ringing on the output waveform - I turned the lamp off and the problem dissapeared. I'd be interested to here other peoples experiences here. "
I was working in a lab where some military missile launch detection equipment (radar signature) was being tested once, and it kept indicating the detection of of an incoming missile. The source was eventually traced to a fluorescant lamp in the room. When the security people got wind of this they came down and confiscated the bulb, since we weren't in a "secure" lab at the time where such covert coded signals could be used. Made me wonder if WWIII could be set off by a fluorescent lamp someday. Have moved to the stix since then, not a concern anymore.
Don
I was working in a lab where some military missile launch detection equipment (radar signature) was being tested once, and it kept indicating the detection of of an incoming missile. The source was eventually traced to a fluorescant lamp in the room. When the security people got wind of this they came down and confiscated the bulb, since we weren't in a "secure" lab at the time where such covert coded signals could be used. Made me wonder if WWIII could be set off by a fluorescent lamp someday. Have moved to the stix since then, not a concern anymore.
Don
Re: EMI ingress
I do this on regular basis. This is not only about RFI, but also about D/A residuals arriving through signal path. I succeeded in suppressing the HF content more than 10x by design/construction modifications (shielding, wiring, filtering, box construction).
Bob Cordell said:
I do this on regular basis. This is not only about RFI, but also about D/A residuals arriving through signal path. I succeeded in suppressing the HF content more than 10x by design/construction modifications (shielding, wiring, filtering, box construction).
Re: EMI ingress
This is the discussion I’d like to be having, real world effects. Although I don’t personally buy the EMI into the output of the amp scenario, I have proven to myself that EMI finding various paths into the signal path is a real issue.
Measuring it, as you speculated is another challenge partly because, and here I’ll speculate from my experience, I believe a large portion of the noise circulates on the returns. When we hook up a scope or a spectrum analyzer we use ground as a reference point. Does the noise appear on the display? Maybe a differential probe could be used. I haven’t tried this. I know I can measure the signal developed across 4” of the speaker return in my amps this way. I’ll put it on my to-do list…
Along these same lines I’ll lead into your next point concerning cabling. In general terms whenever you interconnect two separately powered components there will be a difference in potentials between them even if they are powered from the same mains outlet. This was pointed out to me years ago and a quick measurement between the unconnected chassis with a voltmeter will verify this. (This is a topic unto itself). When the interconnects are connected there is a noise current that circulates. What path this takes (and whatever other noise present) determines whether or not it ends up as part of the signal path. With most of the grounding schemes I’ve seen over the years the path is not well thought out in this respect.
So yes, I would agree that the cabling might just be bypassing the noise around the signal path, another possibility was that different dielectrics might affect the amount of noise that gets radiated between the conductors. Another thought is that the effectiveness of a components internal ground structure would cause wide swings in the cables impedance; possibly the effectiveness of it's shielding.
So I am in agreement with you and your thoughts as to a cables sonic effect being dependent on the performance of a components electrical environment. I get excellent results from basic twisted pair construction using a quality conductor material and matched connector metals.
There’s a lot to learn here and the surface has only been scratched.
Regards, Mike.
Hi Bob,Bob Cordell said:
This is the discussion I’d like to be having, real world effects. Although I don’t personally buy the EMI into the output of the amp scenario, I have proven to myself that EMI finding various paths into the signal path is a real issue.
Measuring it, as you speculated is another challenge partly because, and here I’ll speculate from my experience, I believe a large portion of the noise circulates on the returns. When we hook up a scope or a spectrum analyzer we use ground as a reference point. Does the noise appear on the display? Maybe a differential probe could be used. I haven’t tried this. I know I can measure the signal developed across 4” of the speaker return in my amps this way. I’ll put it on my to-do list…
Along these same lines I’ll lead into your next point concerning cabling. In general terms whenever you interconnect two separately powered components there will be a difference in potentials between them even if they are powered from the same mains outlet. This was pointed out to me years ago and a quick measurement between the unconnected chassis with a voltmeter will verify this. (This is a topic unto itself). When the interconnects are connected there is a noise current that circulates. What path this takes (and whatever other noise present) determines whether or not it ends up as part of the signal path. With most of the grounding schemes I’ve seen over the years the path is not well thought out in this respect.
So yes, I would agree that the cabling might just be bypassing the noise around the signal path, another possibility was that different dielectrics might affect the amount of noise that gets radiated between the conductors. Another thought is that the effectiveness of a components internal ground structure would cause wide swings in the cables impedance; possibly the effectiveness of it's shielding.
So I am in agreement with you and your thoughts as to a cables sonic effect being dependent on the performance of a components electrical environment. I get excellent results from basic twisted pair construction using a quality conductor material and matched connector metals.
There’s a lot to learn here and the surface has only been scratched.
Regards, Mike.