[NagysAudio]

To those who are interested.

I have been getting a lot of questions from people asking me why they need a high bandwidth audio cable, when all they can hear is in the 20-20,000Hz range. Answering this question involves a lot of explaining, so I will touch upon the more important aspects.

First of all, the two most important factors in signal transfer are phase and time. The smaller the phase shift and time distortion, the closer it will sound to live music. This holds true not only for cables, but for all audio equipment, amplifiers, speakers, etc.

To reproduce 20,000Hz with zero phase and time distortion, the implied frequency is around 2MGz. To reproduce 100,000Hz (SACD, DVD Audio) the implied frequency is somewhere around 10GHz! However this is all with just passing one tone. Music is a mixture of many different frequencies, the harmonics can reach into the GHz range even for normal audio CD's.

If the cable is not capable of passing such high requencies, they begin to reflect. This reflection causes an avalanche effect down the frequency range, ruining the correct phase and time in the audible range of 20Hz-20,000Hz.

To reproduce transients and dynamics in music with lifelike quality, the signal must have zero time and phase distortion.

Thanks for reading,
Norbert
   

[Wayner]

How do you tell the copper to be more wide band? It's a mystery to me.

[Bob Reynolds]

To reproduce 20,000Hz with zero phase and time distortion, the implied frequency is around 2MGz. To reproduce 100,000Hz (SACD, DVD Audio) the implied frequency is somewhere around 10GHz! However this is all with just passing one tone. Music is a mixture of many different frequencies, the harmonics can reach into the GHz range even for normal audio CD's.

If the cable is not capable of passing such high requencies, they begin to reflect. This reflection causes an avalanche effect down the frequency range, ruining the correct phase and time in the audible range of 20Hz-20,000Hz.

Can you explain the computation used to compute "implied frequency"?

How do you address the impedance mismatch between the cable and the RCA connector?

Thanks,
Bob

[NagysAudio]

Hi Wayner, it's not the copper that makes the cable have wide bandwidth, it's in how the surface is manufactured. At really high frequencies, over 10GHz, having a smooth surface is critical. This means it must not have any imperfections, or parts of the signal will start to reflect. The very best of coax cables in the world, use a copper center conductor plated with silver, then laser polished to a perfect mirror. These cables are made in multi million dollar facilities and sold for aerospace and defense applications.

[NagysAudio]

Bob Reynolds, everything can be measured about an instrumentation cable with decent euipment. All one needs is a fairly high quality oscilloscope, for example to measure the phase shift at certain frequencies of a specific cable. This data can then be compared to other cables.

Selecting and attaching connectors can make or break a high quality cable. It is by far the most difficult task, especially for audio. Unfortunately, consumers are stuck with RCA plugs, while even BNC connectors are infinitely superior.

I selected an RCA connector that measures really close to the 50ohm military coax cable, at appropriate frequencies. After that, it's all in how it is attached, the quality of the solder joint, quality of solder, how the braids are oriented, how the pressure fit is implemented. This is done while measuring the cable, to find optimum geometry position. The connectors are then epoxy sealed.

[Occam]

I'm not sure I'm following. Now at line level, folks might well run op amps [opamps not specifically monolythic integrated ciruits] configured for a power bandwidth of say 200kHz, and often far narrower. And certainly, in a feedback circuit you have to deal with compensation issues (ala JLH) to avoid slewing and transient intermodulation distortion. (do folks still talk about this?) And I vaguely follow 'skin effect' based arguments, but specifically, I'm not following why a bandwidth into the GHz region is needed and what specific kind of distortion you're trying to avoid.

TIA,
Paul

[Dan Banquer]

Where do I start here, let me count the ways.
Microphones commonly used in audio roll typically roll off sharply after 20 kHz. Then there's the input filter from microphone before the A/D converter, typically somewhere around -3 db at 40 kHz, and rolling off sharply from there. Then there's the Digital Filter after the A/D converter. Typical 8x oversampling at 44.1kHz converters will have a flat frequency response to 20 kHz between -6db and -12db down at 22 kHz and between -80db to 120 down at 24kHz. This will stay down that low until about 325 kHz, where it will have 0 db attenuation.
Microphone cable will get you to an easy 100 kHz.
 Are we having fun yet?
               d.b.

[NagysAudio]

You're absolutely right! However, the medium that the 20kHz signal will propagate in, cable or an amplifier circuit, must let much higher frequencies go through, if one is to hear them with zero phase and time distortion. One can take a simple Denon receiver and measure  20Hz-20kHz sinewaves through it. Sure, it will pass the signal, but the phase at higher and lower frequencies will slightly be off and so will time. Certain frequencies will pass through the unit faster than others. This is all on the micro level, BUT to a human ear it's all it takes to tell that it's a recording and not a live event.

Everything in the chain counts and poor electronics like a mic or dacs, etc. will provide a poor end result. It doesnt matter that cd's are limited in bandwidth, but they should reproduce the 20-20,000Hz band accurately.

[Occam]

....
Everything in the chain counts and poor electronics like a mic or dacs, etc. will provide a poor end result. It doesnt matter that cd's are limited in bandwidth, but they should reproduce the 20-20,000Hz band accurately.

Thats nice, but like saying (in the States) we like Mom and Apple Pie. Nice sentiment but a bit vauge. I'm asking again, what specific kind of distortion are we trying to avoid in this cable, and why does it require a bandwidth into the GHz?

[NagysAudio]

Time and Phase distortion

[Occam]

Read your PM prior to posting again, please.
And who specifically am I exchanging posts with?

Thanks,
Paul, Lab Circle Moderator

[woodsyi]

Norbert or Arnold,

I think you need to send out a few pairs of your uber cables on an audition tour.    I would be glad to test them but I need 2 1meter pairs and 2 2 meter pairs for my actively biamped system.  That would be $3400 list price.  BTW, does it matter that I have tubes in the chain?  Would these ancient valves choke the boadband signal your cable passes?

[Daryl]

Nagys Audio,

All of your claims are invalid.

1. Group delay would have to be considerable before it could even be noticed.

Any decent cables group delay will be far far beyond the ability of human hearing to resolve.

2. Your cable might have a bandwidth to 10 Terahertz when connected to a piece of equipment with a 50 ohm input impedance.

However once you connect it to piece of audio gear with an input impedance of 20 kohm the mismatch would limit your bandwidth probably to under 1mhz due to capacitance.

If you tried to locate a 50 ohm load inside the connector to properly terminate your cable it would cause distortion in the output stage of the connected equipment due to being overloaded (also bass response would sag on most equipment due to the output coupling capacitor becoming a highpass filter when loaded down 200 times more than expected).

Also at 10Ghz the insertion loss of any RCA connector would be sizeable.

At these frequencies the connector must be designed so that the geometry of the transmission line does not change as it hands off to the connector.

Even if both have the same characteristic impedance a sudden geometry change will cause a reflection.

3. If you want to sell audio cables I would recomend offering something that offers the audiophile high performance.

I would recomend a twisted pair or quad with extreme symmetry in both geometry and the resistance of each conductor with minimal capacitance conductor to conductor and conductors to shield which should be 100% and connected electricaly at the insertion end only.

Daryl

[NagysAudio]

Paul, no offense, but I have no idea what you're talking about.

Woodsyi, I am working on demo sets of cables for that purpose, I have not finalized what I want to do, I might offer a 30 day trial period, with money back if not satisfied. I will post back here or on my site. It should be an excellent match with tube amplifiers. I have heard many  properly designed tube amps that sound great.

Daryl, point 2 is correct. It's unfortunate that we as consumers have to deal with RCA connectors, they are the worst of the worst, but all one can do is make the best of it and use the highest quality available. Best sounding amplifiers are not capacitor coupled, and yes, the bass will suffer in a capacitor coupled amp.

Point 1 I disagree, sometimes it only takes minute differences for the human ear to interpret the sound as being closer to live recording.

Point 3 is completely false, the best cables for signal integrity is most definately a coax cable! It doesn't even warrant discussing this issue.

Norbert