Hi JN, I think you mentioned coax send/return currents recently.
Mid-fi gear universally utilises pairs of RCA sockets where the neutral connections are formed from one piece of metal, ie perfectly commoned.
Consequently L and R interconnect plug outers are perfectly commoned at both ends with essentially zero inductance 'separating' them.
Thought experiment.
If the two coaxial interconnects were fitted in reversed directions and the outers/shields were to have any degree of directional characteristic (caused by outer sheath polymer ?....) then the center wire send and the braid return currents would not match causing both the coaxes to be 'unbalanced', and the sign of the unbalance according to signal polarity....ie each shield/braid carrying more or less current than its own central wire.
What would be the effects of such core/shield current imbalance according to frequency and source/load impedances for 75R coax ?.
Dan.
Mid-fi gear universally utilises pairs of RCA sockets where the neutral connections are formed from one piece of metal, ie perfectly commoned.
Consequently L and R interconnect plug outers are perfectly commoned at both ends with essentially zero inductance 'separating' them.
Thought experiment.
If the two coaxial interconnects were fitted in reversed directions and the outers/shields were to have any degree of directional characteristic (caused by outer sheath polymer ?....) then the center wire send and the braid return currents would not match causing both the coaxes to be 'unbalanced', and the sign of the unbalance according to signal polarity....ie each shield/braid carrying more or less current than its own central wire.
What would be the effects of such core/shield current imbalance according to frequency and source/load impedances for 75R coax ?.
Dan.
Last edited:
The first obvious problem is that there is no material which is capable of making the shield or sheath directional at audio frequencies.
Assuming batteries...
If you put the two IC's through a very sensitive current clamp, you will see that it will always read zero. There is always a net zero current when all 4 conductors are within the current clamp, no matter what the signals in the wires do.
If you clamp over one coax, you will read 1/2 of the core current going down either coax. If for example, you pass one ampere DC down the core of the coax you've clamped to, it will read the 1 ampere, minus the 1/2 ampere that is returning through the braid. The other half ampere is going back via the other shield.
The inductance of the loop formed by the two shields may be in the 10 to 20 uH range. If you push a 100 MHz signal in one coax, all of that current will return via it's own coax. This is because the reactance of the other shield as a return path is too high because of the loop inductance, the current will see the coax impedance as the less reactive return path.
Now, add a line cord. It's ground conductor is massive, into the wall, back to the pre. If your IC shield with rca connections is wimpy, and you put 1 ampere DC into one coax core, all the current will return by the safety grounding conductor in the line cord. As you raise the frequency, the ground loop of the line cords will decouple and carry less and less current, the two IC braids will start to carry more and more. Eventually, at high enough frequency, the two braid loop will decouple, and the driving core will be completely covered by the return current of it's signal on it's shield.
Coax by design, only works perfectly when it's core signal is exactly equal to and opposite in direction to it's shield current. If they are not identical, they will create an external magnetic field, and they will be susceptible to same.
To sum up, two RCA's between two line corded three prong plug devices will absolutely NOT have the IC's function as coaxial cables; the frequencies where this changes is dependent on the wire gauges, the cord layout, the contact resistances of the rca sockets (easily 15 milliohms per connection, where a 12 gauge grounding wire is 1.7 milliohms per foot), and will do lots of this in the audio range.
Whenever the currents do NOT return via their own shield, it both creates external magfields as well as becoming more susceptible to external magfields.
Since the loop resistivities are so dominated by even connector contact resistance, how does one consider that when playing cable swap?
Since the loop coupling frequencies are dominated by the resistances in addition to the loop inductances, how do you consider how the wires are layed out?
Before even considering cable swap testing, you have to consider the impact these entities have on the system response. Otherwise, you're wasting time and drawing bad conclusions.
jn
Assuming batteries...
If you put the two IC's through a very sensitive current clamp, you will see that it will always read zero. There is always a net zero current when all 4 conductors are within the current clamp, no matter what the signals in the wires do.
If you clamp over one coax, you will read 1/2 of the core current going down either coax. If for example, you pass one ampere DC down the core of the coax you've clamped to, it will read the 1 ampere, minus the 1/2 ampere that is returning through the braid. The other half ampere is going back via the other shield.
The inductance of the loop formed by the two shields may be in the 10 to 20 uH range. If you push a 100 MHz signal in one coax, all of that current will return via it's own coax. This is because the reactance of the other shield as a return path is too high because of the loop inductance, the current will see the coax impedance as the less reactive return path.
Now, add a line cord. It's ground conductor is massive, into the wall, back to the pre. If your IC shield with rca connections is wimpy, and you put 1 ampere DC into one coax core, all the current will return by the safety grounding conductor in the line cord. As you raise the frequency, the ground loop of the line cords will decouple and carry less and less current, the two IC braids will start to carry more and more. Eventually, at high enough frequency, the two braid loop will decouple, and the driving core will be completely covered by the return current of it's signal on it's shield.
Coax by design, only works perfectly when it's core signal is exactly equal to and opposite in direction to it's shield current. If they are not identical, they will create an external magnetic field, and they will be susceptible to same.
To sum up, two RCA's between two line corded three prong plug devices will absolutely NOT have the IC's function as coaxial cables; the frequencies where this changes is dependent on the wire gauges, the cord layout, the contact resistances of the rca sockets (easily 15 milliohms per connection, where a 12 gauge grounding wire is 1.7 milliohms per foot), and will do lots of this in the audio range.
Whenever the currents do NOT return via their own shield, it both creates external magfields as well as becoming more susceptible to external magfields.
Since the loop resistivities are so dominated by even connector contact resistance, how does one consider that when playing cable swap?
Since the loop coupling frequencies are dominated by the resistances in addition to the loop inductances, how do you consider how the wires are layed out?
Before even considering cable swap testing, you have to consider the impact these entities have on the system response. Otherwise, you're wasting time and drawing bad conclusions.
jn
jn's explanation, the conclusion of which is quoted above, is excellent because it is clearly stated. My understanding of the variables involved has been greatly increased. 🙂
Yes. Have you had a look at this thread? Particularly the posts of ilimzn
The dozens schemes to wire an amp...
No problem, i´m sure i have mentioned the several path issue, but it can´t be emphasized to much......
Hearing the difference between two different conductors may be lost on some people but here are two instances when it was obvious to people besides myself.
I had a pair of thick braided copper "Monster" speaker cables inside clear plastic insulation in my car audio system. My car got flooded by mineral rich brown water that had them marinating for a few days. I hung them out and forgot about them for a long time. I found them again and did an experiment. I compared them to zip cord speaker cables and the sound degradation was very noticeable using the now green looking cables. My friend, who could not tell any difference when they were new now heard a difference.
When I was Chief Engineer of radio stations I would replace the coaxial cables that carried the composite FM stereo multiplex signal to the transmitter with a silver plated solid copper center RG223/U. The first time I did that the original cable had been copper plated solid steel center conductor. The sound difference was pretty significant so much so that two colleagues separately commented about the improved brightness of the station's audio and they had no knowledge of my change.
It's been said that the skin effect comes into play at audio frequencies and that introduces some distortion as the signal propagates differently at varying depths of the conductor relative to the outer surface. Oxides of copper are poor conductors, oxide of silver are good conductors. Plastic insulation is not generally gas tight so oxides will build up over time compared to new cables.
I had a pair of thick braided copper "Monster" speaker cables inside clear plastic insulation in my car audio system. My car got flooded by mineral rich brown water that had them marinating for a few days. I hung them out and forgot about them for a long time. I found them again and did an experiment. I compared them to zip cord speaker cables and the sound degradation was very noticeable using the now green looking cables. My friend, who could not tell any difference when they were new now heard a difference.
When I was Chief Engineer of radio stations I would replace the coaxial cables that carried the composite FM stereo multiplex signal to the transmitter with a silver plated solid copper center RG223/U. The first time I did that the original cable had been copper plated solid steel center conductor. The sound difference was pretty significant so much so that two colleagues separately commented about the improved brightness of the station's audio and they had no knowledge of my change.
It's been said that the skin effect comes into play at audio frequencies and that introduces some distortion as the signal propagates differently at varying depths of the conductor relative to the outer surface. Oxides of copper are poor conductors, oxide of silver are good conductors. Plastic insulation is not generally gas tight so oxides will build up over time compared to new cables.
What was the original RF cable and frequency range of the multiplex? Without knowing that the anecdote doesn't have anything to work on.
That's pretty much like comparing car gas mileage with and without 4 flat tires. I don't think anybody would argue against your observation there.
The left and right signals run from 23Khz to 53 KHz, I would expect it to need high quality coax and connectors, so I wouldn't be surprised by a change.
Unfortunately, that is an audio myth started by a rather badly written article, using rather terrible assumptions and understandings, and included a rather badly designed test setup. I expect it will always surface again and again.
jn
The multiplex signal is mono (L+R) to 15 KHz and the stereo sub channel is a left right difference (L-R) that is amplitude modulated on 38KHz and spanning 23KHz to 53KHz.
No RF involved at this stage.
No RF involved at this stage.
Actually I think plenty of people are happy to argue against it. That is why I offered such an example. They can even try it for themselves.
Ah ok, baseband. Over here multiplex generally infers the full set of stations being sent across so covering MHz. If I were a guessing man then I would suggest the double screen of RG223 make it superior in that application to std single screen (depending on what it replaced.
Ref the corroded cable there is at least one story of someone preferring it that way. if the corrosion goes far enough you end up with almost litz wire. I didn't but that other than an example that the internet can always find a contrary viewpoint. 🙂
Ref the corroded cable there is at least one story of someone preferring it that way. if the corrosion goes far enough you end up with almost litz wire. I didn't but that other than an example that the internet can always find a contrary viewpoint. 🙂
I have an old MC zip cord that has grown a nice green rot for 40ys. I am skeptical that cleaning and dipping the ends in a solder pot wouldn't just fix any issues unless of course the corrosion has gone all the way through the conductors. Were you curious enough to even check the DC resistance against that typical of the wire gauge?
BTW when my favorite college station replaced their all valve signal chain with SOTA SS the sound went down the tubes IMHO of course.
I agree with jn on the last point, and would add that if a skin of a poorer conductor forms the signal does not want to go there in the first place. Hawksford's articles ae full of problems
Wire skin effect table. I know of two audio companies that use 20ga, as they feel that ga is the best choice for signal and low power internal wiring.
American Wire Gauge (AWG) Cable Conductor Size Chart / Table
American Wire Gauge (AWG) Cable Conductor Size Chart / Table
That has to be the ultimate in confirmation bias. You hear a change because you want to, not because it's there. This is true in >>99% of these "magical" changes in audio systems - you know; the ones you can't measure.
There are shysters making enormous amounts of money from gullible audiophools. I was in a factory in Shenzhen, China last month, where they were adding "direction markings" to ordinary , cheap phono cables. No doubt these will turn up "branded" in the near future and be priced at $$$.
The company there sell the unmarked versions of the cables, with gold plated phono plugs on each end, for between $0.04 and $0.09 according to length, if you buy >1000 pieces. It's going to be funny to see the prices the "marked" versions of these sell for to the gullible!
So can I and my hearing is not what it was in the 60's. In fairness however, the difference IMO is very small and easily could be missed in a double blind test. I think this subject was well covered in the first 3 posts of this thread. For those that do hear a difference in cables, it is the structure that rules over the type of metal.
I suppose it is also verboten to mention that tubes from different manufacturers also sound different. No, not suggesting high $ is better.
I also hear a difference in transformers based on the material and winding. Oh no will it never end??
Don't make me mention resistors. It will really light up this thread. Yes I do wind my own from manganin wire for critical parts when an equal resistance is available for a penny. Surely we can agree that there are measurable differences.
And?
The effect is real, as is proximity effect.
The reason for that table is to determine the IR heating caused by the current at frequency.
It has nothing to do with sound quality nor distortion, just heating. Ampacity is the determination of how much current a conductor can carry. It depends on the insulation type as well as bundling in a common jacket or in a conduit.
On occasion, it is necessary to consider 180 hz ampacity, as many simple power supplies pull lots of harmonic current.
Anyone who says it can be heard is pulling your chain. Read the Essex echo article, study it well, understand it. Then you will know what is entirely incorrect.
Funny, that table doesn't even cover the wires I use. Where is the 535 to 900 kcmil? 4 ought is not that large. And, 40 AWG? That's not that small, I use 49.
Jn
It's complete BS. 🙄
The carefully limited and filtered (20Hz to 15kHz) audio feeding the stereo coder is pre-emphasised at 75µs in the Americas and 50µs elsewhere. The "stereo multiplex" feeding the transmitter consists of three components:
A "Sum" or "M" signal, which is what a mono receiver "hears", so they get to hear all the audio from both channels added together.
A "Difference" or "S" signal which is actually Left - Right, transmitted as DSSC (double Sideband, Suppressed Carrier) centred on 38kHz.
A 19kHz sinewave pilot tone, used to indicate to the receiver that it's a stereo transmission and to supply a signal that's doubled in frequency in the receiver stereo decoder to provide the missing 38kHz carrier to allow resolution of the DSSC "S" component.
The stereo receiver's decoder does a little bit of algebra, and adds the M and S components together:
(L+R) + (L-R) = 2L
It also subtracts the S component from the M part
(L+R) - (L-R) = 2R
The separated signals are then de-emphasised, and we recover the 20Hz to 15kHz L and R components as they were fed into the stereo coder at the transmitter end.
Any "magical" cable change is going to affect all the components of the stereo multiplex signal. Increasing the amplitude of any of the multiplex components will just lead to the transmitter over-deviating (very much frowned upon by the authorities).
The audio filtering is an industry standard, and "brickwall" filtering is used at 15kHz, so that there is no audio content whatsoever by the time you get to 19kHz, to prevent interference with the pilot tone. My stereo processor for broadcast uses a multiple gyrator approach (emulating inductors), and there is no measurable content by 19kHz.
The abuse that most radio stations apply to their programme material means that it really can't ever be called "high quality". It's compressed, limited, has "Aphex" added (deliberate second harmonic distortion is added to the "speech band" to make it sound "louder"), there's phase rotation trickery, and then the whole mess is clipped (more distortion) and filtered.
My approach eliminates the "Aphex" and the clipping - I use "feed-forward" limiting that "sees the peaks coming" giving a limiter with "zero attack time" and no overshoot, thanks to the use of a delay line. My stations sound a little quieter than the competition, but I get more listener hours, because they're not so fatiguing to listen to....
No cable change is going to make a significant change to the "sound" of a stereo radio station, unless the one you're replacing is defective!!
Different valves ("tubes" - US) from various manufacturers will add differing distortions as the methods of manufacture may vary, leading to slightly different geometry inside the valve. Just remember - what you're listening to is distorted by the valves, by the audio transformers and by the loudspeakers.
In 40 years of broadcast engineering, I haven't seen any valved amplifier that could anywhere nearly equal the performance of even the cheapest solid state model.
Valves have no place in modern, high quality amplification. The "sound" you like is just even harmonic distortion. The only places they should be used are in guitar amplifiers and high powered TV transmitters!
- Status
- Not open for further replies.