Cabling and Grounding for Audio, Video, Speakers and Power. Part 2

A Selection of Various Speaker Cables

Part 1 provided an insight into my views regarding cables and how to provide power and a grounding system that will prevent hum loops and sink electrical interference.

In this post I discuss the various types of audio and video cables, connectors and other related issues. Wrapping up with a few comments regarding Esoteric Cables.

Analog Audio Line Level Cables

Unbalanced

A typical RCA/Phono connector used for unbalanced audio connections.

Most low to medium cost audio systems only support unbalanced line level signal audio connections. These are created using RCA or phono connectors and use unbalanced coaxial cables. For most users these cables are short and rarely over 6 feet except maybe to any active speakers like subs. My recommendation here is always to use plain copper cables with the highest strand count that you can get for the braid and center connector, with the lowest capacitance per foot, typically 20 picofarads. These simple cables have extremely low inductance, typically 0.1 micro Henry’s per foot, that can also be ignored at audio frequencies. In my case I use balanced cabling to create a pseudo-balanced connection by using the two internal cables to carry the signal and just grounding the shield at the source end. You don’t need Litz cabling, silver plated cables or carbon impregnated center shields unless you are connecting up a phono head amp or cartridge that only outputs a few millivolts. See my turntable rebuild here.

Balanced

A typical 3 pin XLR connector used for balanced analog audio connections.

Balanced audio cables were originally designed for very long runs where the source and destination impedances had to be matched in order to prevent a significant loss of signal. These interconnects were based upon a 600-ohm source and destination balanced system. The 600-ohm standard is not used in most balanced equipment anymore as the runs are relatively short and introduces no significant level losses. So most balanced professional or home equipment now use balanced outputs of typically 100ohms or less, with balanced inputs of 10,000 ohms or more. This difference in impedance effectively removes the 600-ohm losses whilst maintaining the balanced cables ability to provide very high interference rejection (this rejection is also a function of the balanced I/O electronics).

Higher priced more exotic amplifiers and receivers often provide balanced connectivity including connections to phono cartridges. All balanced connections use 3 pin XLR connectors for audio, and should at at least be gold plated to inhibit oxidization. The rule is the same, plain copper and high strand count unless it is for phono connectivity – as above. In my case I built my own balanced Litz wire cables to support my moving coil cartridge and head amp connections.

Star Quad

These cables are often used to support hand held microphones were high noise rejection is required and the cable is often being moved and could become microphonic. While having up to a 30dB improvement in magnetic field interference rejection over standard balanced cables, it does have the disadvantage of its capacitance per unit foot being about 50% more than typical balanced cable. This can impact high frequency roll off over long distances. However, in residential applications the distances would not be significant. This cabling configuration has four conductors at its core with the usual outside braided screen. See Belden and Wikipedia for further technical details. My own system uses Belden 1192A quad cabling for all balanced connections, just to be safe, as it has a very high airborne noise rejection ability, is plain copper and multi-stranded.

Generally speaking all the above signal cables are fed by, and drive, electronics that have fairly constant impedances. Typically, 100 ohms or less for the source and 10K ohms or more for the destination. These low output impedances will generally swamp cable effects unless the cables have high capacitance, see above, or inductance values, which professional cables do not generally have.

Be aware that many device input and output circuits contain filter components that are added in order to prevent or reduce RF interference from entering the device after being picked up by the cable. If well designed these RF suppression circuits SHOULD NOT impact the HF attenuation or phase response of the signal within the audio pass band or at least an octave above, up to 40KHz, making their effects inaudible.

Speaker Cables

A Neutrik Speakon connector

Typical Low Impedance Speaker binding posts

Low Impedance 4 to 16 ohms

Unlike line level signal cables, speaker cables carry high alternating currents and feed speakers whose input impedance is not remotely constant. These dynamic load changes will interact with the cables impedance causing the overall frequency response of the system to vary. Furthermore, speakers are influenced by the damping factor provided by the amplifier. Damping is the ratio of source plus speaker cable impedance, to speaker impedance. The higher this ratio the better, as it contributes significantly to the speakers control at low frequencies. Well designed solid state power amplifiers can have output impedances below 0.1 ohms where as valve based power amplifiers can be as high as 0.5 ohms. Providing damping factors for a typical 8 ohm speaker of about 80 and 16 respectively. You therefore need the cables DC resistance and impedance to be as low as possible. Changes to the cables impedance in conjunction with the speakers changing impedance, will effect how the speaker performs and sounds.

A Typical Speaker Impedance Plot

These complex speaker impedance (Z) changes may also impact the power amplifiers performance and therefore what signal the speaker finally gets delivered to its terminals. Also there are several power amplifier signal parameters, depending upon their design, that can change depending upon the load presented to the amplifier. Yet again, another variable in the cable equation!

It is therefore reasonable that cables with high capacitance and inductance, with any significant DC resistance, will impact what you hear. So, nothing new there. You again need cables with low DC resistance, multi stranded plain copper for low skin effect, together with low inductance and capacitance per unit foot. I have used a range of 12 & 14SWG multi-stranded plain copper cables from various manufacturers including; Monster Audio, RCA High Performance Speaker Cable and Marshall Sound Runner OFC, and never heard any significant differences between them. In most cases my cable tests were performed using my bi-amped system.

High Impedance 70 volt systems

70 volt speaker distribution systems are generally only used to ‘pipe’ background music around large homes and spaces such as clubs and stadiums etc. It really has no place in the home. It is designed to support many speakers covering large distances and areas. However, the cable gauge, strand count and use of plain copper still applies to get optimum connectivity. For such systems the audiophile quality is secondary to the installation requirements, coverage and intelligibility of the sound.

Line level

Balanced or unbalanced analog audio line level feeds that drive active speakers are subject to the same issues described earlier in analog audio line level cables. In some of the more high end or professional speaker systems, the signal to active speakers is supplied as a digital AES3 signal over either balanced or unbalanced cabling, see digital audio cabling below.

Digital Audio Cabling – Copper or Optical

Typical 3 pin XLR connector for 110ohm AES digital audio

Typical RJ45 1Gbps network connector

A typical BNC 75 ohm video or digital audio connector

TosLink Optical Connector

A typical LC fiber optic connector used for digital video or audio data

A categorical statement for ALL audio cables that carry digital signals, be they copper or optical. If the source signal gets to its destination above the threshold required for the re-clocking and re-equalization circuits to clean up the signal, the cable type DOES NOT MATTER. They DO NOT impact jitter or any other audible parameter of the signal.

Copper digital audio cabling maybe balanced 110 ohms or unbalanced 75 ohms. There are no audible advantages of either connection type. Very often it is the personnel desires of the client or the available connectivity on the hardware that drives the use of one or other of these connection types. That being said, balanced connections always provide better rejection of airborne interfering signals.

Video Cabling – Analog or Digital, including HDMI

A typical HDMI connector copper or fiber.

A typical 75 or 50 ohm BNC video connector.

A typical non professional RCA video connector – not 75 ohm.

ALL video cabling should have an impedance matched source and destination with an interconnect cable and connectors that have the same characteristic impedance, typically 75 or 50 ohms. Why? It stops reflections and standing waves that degrade the ultimate distance the signal can travel, NOT what it looks like when it is displayed or sounds like. As with digital audio signals, provided the signal equalizing and re-clocking electronics can recover the signal, the cable makes no difference, other than how far the signal can travel. It does not effect any visible parameter of the video signal. With analog video signal levels, low signals at the display will result in increased picture noise and possible loss of chrominance (color) and HF detail. This is should never be an issue in a typical residential system.

Typical SD analog or digital copper video cabling can cover distances of up to 1000 feet with no technical issues. HD signals can cover distances of up to 500 feet while 4K signals can cover distances of up to 250 feet depending upon just how low loss the cabling is. Optical connections have no practical distance limits in residential setups, being able to exceed 10Km or 30,000 feet.

Copper and optical HDMI cables are subject to similar technical restrictions, however their distances are very much more limited as they carry signals up to 18Gb/s, where as professional digital video cables do not generally exceed 3Gb/s. Most professional 4K connectivity being over CAT6(A), optical or quad SDI link. See here for more information on 4K HDMI cables. (There is currently a move to provide professional 4K video signals (12Gb/s) over a single unbalanced coaxial cable. Only time will tell how popular this connectivity becomes in the light of optical and CAT6 connectivity.)

Fiber lengths used in a home are relatively short, and any losses other than those created by bad surface mating due to dust in the connector, can be completely ignored, and cable length has absolutely no effect on the audio or video that the connection carries.

It has become common practice in many home receivers and video processors to use RCA connectors for the video connection. Why? They are inexpensive compared to a BNC connector as are the mating video cables. So in spite of the fact that these connectors are NOT 75 ohms, their impact on signal loss is generally minimal as these cables are rarely more than 10-20 feet.

Audio Connectors and Connections

Male RCA or Phono Connector

3Pin Female XLR

TosLink Optical Connectors

USB Connectors

There are two types of connectors that can generally be found on equipment, copper and optical. I will say nothing regarding optical connections other than provided that the plug and socket is kept free of dust there are few practical restrictions on this type of connection within a home or professional environment.

We are all presented with the need for metallic connectors, and all surfaces must act as if one with no metallurgic effects on the contact surface areas, see oxidization below.

I am fortunate to be able to make all my own audio and video cabling (not HDMI). Based upon some very early spectrum analyzer tests of crimped, wrapped and simply lap soldered joints, the latter two provided better RF conduction results due to the inductance of the wrapped joint. So all joints should be crimped, or soldered with just a conductor overlap. I know we are dealing with audio not RF, but I am just very cautious. Note that virtually all professional video connections are CRIMPED as are many audio connections. Wrapping soldered joints is a thing of the past, so don’t do it, especially for video.

For those of you who can make your own cabling there is the option to use silver solder. I have only ever used this for the cabling to my moving coil cartridge where the signal levels are extremely low. Whether there are any audible benefits I cannot really comment, but its conductivity is higher than that of lead/tin solders. As to whether it would provide any audible improvement to line level connections using RCA or XLR connections is highly unlikely. I have never seen any audiophile hardware printed circuit boards assembled using silver solder, on which there are many hundreds of analog and digital soldered devices and connections inside a typical pre-amp or receiver. So one or two silver soldered connections at line level, external to the device for a cable, aren’t going to audibly help the signal conductivity.

Other Cable Issues

Cable molecular structure, lattice issues

There is a camp that believes that when copper cables are pulled through the dies to create them, that their lattices structures become damaged and the valency electrons become dislodged or unbalanced. These effects result, apparently, in the need for the cable to absorb additional electrons during a “breaking in” period for the lattice and valence electrons to return to their natural neutrally uncharged state. Not being a metallurgist, I cannot comment on this view, other to say that it seems highly improbable and even if the condition existed it would be neutralized by the application of a current in seconds, not weeks or months. Furthermore, it implies that EVERY time you move or deform a cable it would need to be “broken in”, sorry but I don’t buy that, nor have I ever heard the effect of a cable sounding better with age. (unlike bass/sub speakers that do change with time.)

I have also seen the term Triboelectric effect (static electricity) associated with cables. Again, I fail to see how this contact electrification effect, which can take place between two materials that touch each other, or is caused by materials that rub against each other, applies to cables in a fixed setting! Are we now saying that there is an electron exchange between the copper and its surrounding insulator as the power density in the copper varies? If so, it would be minute and unmeasurable as electrons migrate to and from its surrounding insulator as the current and magnetic field vary down a cable. This effect would amount to leakage currents of unmeasurable levels and I fail to see how a speaker could respond to such infinitesimal losses, even if they existed. More snake oil!

Oxidization

Unfortunately, we have to interconnect all our hardware. This generally means lots of copper connectors. The connector is supposed to provide nothing more than an EXTENSION to the devices internal cabling to the next device. These connectors, plugs and sockets, can have the surface of the metal oxidize, or rust. This layer of oxidization not only effects the impedance of the connection but can act as a diode impacting low level signals producing distortion. My preferred material being gold plating, as it is not easily susceptible to surface corrosion, does not react with oxygen, bonds very well at a molecular level to copper and is one of the best conductors. All properties that are required to create a purely passive connection.

Silver plating can also be used but while this provides slightly better conductivity to gold, tarnishes more easily.

Be careful with certain types of plain copper cabling. There are types of insulation material that actually allow the passage of oxygen to migrate through them and cause the copper to oxidize. I have owned such speaker cables which were purchased from a well known, and respected, existing esoteric cable vendor!

Oxygen Free Copper Cabling

There are three main types of commonly used copper specifications:

C10100 – Known as Oxygen-Free Electronic (OFE). This is a 99.99% pure copper with 0.0005% oxygen content.

C10200 – Known as Oxygen-Free (OF). It has a 0.001% oxygen content, 99.95% purity but its conductivity is no better than ETP below.

C11000 – Known as Electrolytic-Tough-Pitch (ETP), is the most common copper. It is universal for electrical applications. It is 99.9% pure and typically has 0.02% to 0.04% oxygen content.

Only C10100 has an improved conductivity of approximately 1% over the other two standards and it is not considered significant to either the resistance or impedance of any audio cable. It is also extremely expensive. However, any improvement in conductivity of cabling carrying high currents, such as speaker cables, may be audible, and even companies like Belden now sell OFC speaker cabling for residential use. I have never seen it used in any professional environment. Remember, that there is a lot more going on in cabling than its conductivity. Other parameters such as its capacitance and inductance and even connector/termination type, can easily swamp any improvements to the DC resistance of a cable. I have compared OFC speaker cabling on my pair of Yamaha NS10’s, a staple in many recording studios, and heard absolutely no difference. Yes, I know, they are hardly the most revealing of speakers.

Jitter/ Re-clocking issues

This effect is limited to the digital processors master clock stability and other design parameters within the receiver or amplifier, rarely the cables. This is true for all digital audio and video signals. The ONLY proviso for this is that the optical or copper interconnect losses provide a level of signal with sufficiently clean rise times such that the equalization, re-clocking and error correction circuits of the receiving device can accurately re-constitute the digital data with no increase in jitter.

Digital Data Loss

A quick note on encrypted or compressed digital signals and signal interconnects. Some of the algorithms and hardware used to process these types of signals have the ability to repair and replace lost, corrupted or missing digital data. The missing data that is created is not a GUESS at what is missing or some form of interpolation; it is an exact replacement of those bits that were lost. This process can only recover a limited loss of data. Generally, above this limit, the video or audio is frozen or muted respectively. So data losses in these types of system interconnects DO NOT, as before, affect the final video or audio signal in some magical way.

Esoteric Cables – Finally, the Roundup!

If you follow both of these posts dogma (guidance) you should NEVER have a need to use esoteric cables, or if you decide to do so you should not hear their effect – or will you?

Esoteric speaker and line level cables, some of which have radically varying coupling between the conductors and their surrounding screen, and significant differences in their distributed capacitance and inductance, will affect the passage of the signal and what you hear. Why? Because they are filters with the inductance and capacitance affecting their response, just as any low or high pass filter would. There is no magic here, and yes, you may prefer the sound that these cables create. However, they are potentially modifying what was a perfect signal connection in the first place, unless you were using some very poor interconnects. e.g. bell flex for speakers.

Don’t forget that complex impedances, like those described above, who’s impedance and voltage/current phase angle change with frequency can have a significant effect on the performance of the source power, or line amplifier, as they are driving a varying load with potentially out of phase current demands. Such effects can have significant impacts on the final sound quality as they can impact both voltage amplifiers (line level) and power amplifiers technical specifications. Nothing new here!

Never think that any audio cables can in any way be likened to RF transmission lines or video cables, where the wavelengths are very short compared to audio. These cables require impedance matching between their source and destination in order to prevent the signal being reflected back from its destination and causing a number of technical issues in particular standing waves. There is no evidence either antidotally, practically or mathematically that supports such an assertion by anybody. (I was originally trained as an RF engineer and worked for years with antennas, transmission lines and amplifiers). Transmission line properties of cables do not start to exhibit themselves until the cables length is equal to AT LEAST 1/10th of the highest frequency wavelength that it carries. So, if you assume the highest audio frequency is 100KHz, having an electrical wavelength of 3000 meters, I am being really generous here, the cable would need to be approximately 300 meters long! I would be far more concerned about other cable parameters at this length!

Impedance matching, as it applies to audio, only applies to 600 ohm balanced systems and is primarily deployed in professional equipment, and then mainly for AES digital audio. This technique was originally required for the transmission of audio over large distances. Today, balanced audio is still used in order to reduce air born interference on long cable runs but provides no other audible benefits. On the assumption that the balanced and unbalanced line level amplifiers are equally well designed. It does however carry the benefit of handling signals levels double or 6dB higher than unbalanced connections.

Modern day line level signal connectivity is based upon using a source impedance that is below 100 ohms with a destination input impedance of typically 10,000 ohms or more. With source impedances approaching zero ohms the effects of cable capacitance and inductance are drastically minimized, but not completely mitigated. This allows for esoteric cables with complex constructions to sound different to a standard well design balanced or unbalanced cable made from Belden, Klotz, Mogami or Gepco. Note here that these established, and highly respected cable manufacturers, produce a whole range analog and digital audio, video and speaker cabling based upon the application. Its interesting that these very specialized and extremely knowledgeable companies with large R&D divisions, don’t manufacture any esoteric type cables; that should tell you something!

Skin Effect. Well it’s real for sure. It is the process whereby the passage of the electrical current down the cable produces a magnetic field that causes the mobile electrons (current) to migrate to the outer few microns of the cables surface. This effect is generally only relevant at, once again, high RF frequencies. This effect implies that in order to keep the resistance of a cable as low as possible you should use a multi-stranded cable in order to increase the available conductor surface area. If you are a cautious person, like myself, use multi stranded cables, the more the better, with plain copper conductors. I NEVER use tin plated anything as dissimilar materials can cause electrical issues at their junction so all my cables are plain copper. The braid doesn’t matter if it’s only a screen and doesn’t carry an active signal. I apply this thought process to all connectors, being gold plated. I know its plated, but there is a limit!

In my early years I built ALL my pre-amps and power amplifiers, described below. Designing my own toroidal analog power supplies that used custom made toroids with Mu-metal screens. Even back then I only used multi-stranded plain copper interconnects, resorting to solid silver stranded cabling for phono pre-amps. Even today I still make all my own cables using Belden Star Quad 1192A (a plain copper multi-stranded cable) for all my balanced and unbalanced cables. Unbalanced cables being pseudo balanced by using the internal conductors to carry the signal while the screen is only connected at the source end.

This brings me nicely to oxygen free cables (OFC). Have I used/tried them yes. Did I hear any difference? Absolutely not when compared to my 12SWG Monster plain copper stranded cables. I sure heard a difference when I substituted 18SWG bell flex and 16SWG Litz wire for the comparison, the results not favoring the bell flex or Litz wire. My original bi-amped system that I used for these comparisons used a Radford ZD22 to drive a Quad 405 power amp and a pair of Leak TL50+ valve power amps, driving a pair of Maudant Short MS600’s. Tests of this nature can be indeterminate, as passive LF speakers react differently depending upon the amount of damping applied to them, so unless the DC resistance and impedance of the cables is the same you would expect to hear a difference between the cables, again, not surprising!

So, to warp up, do esoteric signal cables make your system sound different, most probably yes, but there are good technical reasons for that especially if your using sub-par cabling, power connectivity and poor grounding. Also how do you know which is the correct sound you probably weren’t in the studio when the recording was produced. I on the other hand was often in the studio and on occasion would play the studio 1/4″ master tapes or DAT’s at home! So yes, you may prefer one sound over the other but which is correct? As for esoteric power cables sorry, but if they make a difference to your sound you have bigger issues that need to be resolved and you should re-evaluate your entire high voltage supply chain and grounding system.

Finally, any vendor that tells you that their digital cable will make your signal sound or look better should be locked away for fraud. Digital signals are not susceptible to the same cable effects as analog signals. Most copper digital audio cables are 110 ohm balanced or 75 ohm unbalanced ensuring a source to destination impedance match, exactly the same as the 75 or 50 ohm impedance match for video signals, thereby preventing unwanted termination reflection problems. So, provided the cables losses do not exceed the destinations hardware ability to equalize and re-clock the signal, the cables performance is totally transparent. This ALSO applies to HDMI cables, be they optical or copper. If the signal gets to the destination such that it can be fully re-clocked, any losses in the interconnect are irrelevant, and cannot be heard or seen. Losses in either type of cable that cannot be corrected will always show up as sparkles, loss of signal, image freezes, pops or muting. Not some magical change to the chrominance or luminance content of the video or the sound quality.

Shop closed.

See Part 1 of this post here.

From Vinyl To Plastic

One man's quest for audio perfection

One man's quest for audio perfection

Cabling and Grounding for Audio, Video, Speakers and Power. Part 2