Barry
25-05-2015, 14:25
Multi-party audition of audio interconnects using inexpensive cable.
Introduction
There has been much discussion here of commercially available VFM interconnects, using cable as supplied to professional recording studios and broadcast institutions. These cables are available ‘off the drum’ and are inexpensive to buy. Some of these cables have been designed for use with hand-held microphones, so are very flexible and often have a semi-conducting layer between the insulating dielectric and the outer screen, to discharge any static charge caused by cable flexing (so-called ‘handling noise’).
This report concerns itself with the audition of audio interconnects, again using VFM cable, but this time most of the cables are those that would be used in the aerospace, computer and other industries.
As is well appreciated, the whole discussion of audio cables polarises the audio community, often causing much dissention and heated debate. It might be useful at this stage to briefly outline the philosophy and thinking of the two ‘camps’. They can be conveniently divided into “objectivists” and “subjectivists”.
Those belonging to the objectivist camp would argue that cables are entirely passive constructs and are electrically so simple (relative to the complexity of the items between which they connect) they cannot have much influence on the overall sound quality. Cables, according to objectivists are qualified by three simple values: the loop resistance per unit length (which depends on the metal used for the conductors), the capacitance per unit length (determined by the geometry of the cable and the dielectric insulation used) and the inductance per unit length (which again, depends on the cable geometry) [1]. Of these three values, the most important is the capacitance per unit length. When cables are used as interconnects, their capacitance will form a filter network with the impedance of the items between which they connect. [2]
Subjectivists on the other hand would rightly argue that any audio system is a concatenation of components, including the interconnecting cables. Thus every part of the chain is as important as every other part, regardless of the electrical complexity of the item. It therefore follows that cables contribute to the overall sound quality of the system just as much as the more complex items of electronics used. Subjectivists will claim they can readily hear the changes made by exchanging interconnects to the overall sound and these changes can be quite large; large enough to justify spending a considerable amount on interconnects if necessary.
At this stage I need to state my position. As someone who has had a scientific training and having worked in an electronics research laboratory for thirty years, I’m inclined to put myself in the objectivist camp and am somewhat sceptical of so-called cable sound. Yet there have been times (much to my surprise) where I have heard audible differences made through changing cables. The sound differences have always been small, but distinct and most importantly, repeatable. So I find myself being unable to explain why I can occasionally hear these differences, but willing to accept they can occur; and more importantly, that other listeners can perhaps more readily hear them.
The purpose of this exercise was thus to try and find out whether cables have an intrinsic sound of their own, or whether the changes in sound quality heard is system dependent.
Methodology
To this end I made up six sets of cables, using conventional coaxial cable as supplied to and used in the aerospace, computer and hydrology industry. These cables were then sent to four other listeners, having four quite separate audio systems and listening requirements. Each listener spent time with the cables (a few weeks) and recorded their observations. In three cases (as well as my own), the cables were used between a CD player or DAC and the preamp. One listener preferred to only use vinyl as the source material, so the cables were fitted between his phono stage and preamp.
I don’t claim the exercise was in any way scientifically rigorous – it was not blind (the cables were not identified, they were simply given a code, but on some the identity of the cable is printed on the outer jacket); the cables were not of identical length (most were 1 metre long), and the RCA phono plugs used were not common to all. None of the cables made used expensive cable and most were fitted with inexpensive RCA plugs of reasonable quality.
The cables are listed in turn along with their technical specification, typical cost per metre and a brief description of the RCA plugs fitted. Listener’s comments then follow; and for all the cables, the listener’s comments appear in the same order.
The cables
Cable 1 Belden 8219 Coax – RG-58A/U
http://i780.photobucket.com/albums/yy88/barrydhunt/008_zpsemdknvx4.jpg
This is an RG-58A/U type cable; consisting of a 20 AWG (19x32) stranded tinned copper inner conductor (0.94mm diameter), foam polyethylene dielectric and a tinned copper braid shield having 95% optical coverage. The outer insulation is PVC.
The characteristic impedance is 50Ω (though this is totally irrelevant at audio frequencies).
Nominal loop resistance: 42.3mΩ/m.
Capacitance: 87pF/m.
Nominal delay: 4.5ns/m.
Nominal surface transfer impedance: ~ 13.5mΩ/m [1]
Typical cost: < £1.50/m
Plugs: Neutrik ‘REAN’
Listener’s comments:
Listener A
Unquestionably my favourite cables of the bunch. Very detailed and open sounding, with no hint of harshness or artifice. Lets the music through relatively unhindered. I could live with them, although swapping back to my reference Klotz revealed a slight lack of spaciousness, ‘musical flow’, and ultimate resolution.
Listener B
Way too much bass. One note bass. Everything ill defined - hard to separate the different elements of the performance. Very little air or ambiance where it should be.
Listener C
This is one aggressive, forward sounding cable with plenty of slam. I would imagine that an extended listening session involving a lot of bright sounding CDs would see your ears sliced off! The aggressive bright sound gave the impression that there was a lot of detail retrieval but at what cost? Can't fault the bass though - powerful, detailed and with some real slam. Great soundstage too. Mids were a bit screechy and thin sounding - okay, but the more natural and fuller tones of (the Maplin microphone cable) were better. Clearly a cable that needs careful system matching but in the right set up could be a bit of a star. I suspect 'What Hifi' testers would love it.
was something of an enigma (can a cable be described this way...?) and potentially a real star. You certainly get a lot of detail retrieval and a fast punchy sound but at times I could barely stand it!!
Cable 2 Boston Insulated Wire (UK) - RG-58 hydrophone cable
http://i780.photobucket.com/albums/yy88/barrydhunt/005_zpsxnu9lazv.jpg?t=1432233356
This cable is designed for use with hydrophones. Essentially it consists of an RG58 coaxial cable, clad with a tough silicone outer sheath. Hydrophones use the piezoelectric effect, so it is important the hydrophone signal is not distorted by any change in capacitance of the connecting cable due to dimensional deformation caused by pressure of the water. Hence these cable need a thick (7.8mm OD), tough and waterproof cladding.
Being an RG-58 type cable, the construction and performance details are very similar to those for the Belden 8219 listed above.
Unfortunately the thick and tough, silicone plastic cladding makes this cable extremely difficult to work with as well as making the cables ‘springy’ and not particularly flexible.
I cannot find any information about Boston Insulated Wire (UK), Esher, Surrey, UK, so I assume they are no longer in business. Cost (if still available): probably ~ £2.50/m.
Plugs: Pro Signal (8mm cable diameter)
I like these plugs because the screen and outer insulation are gripped by a collet, preserving a true coaxial geometry. http://www.farnell.com/datasheets/320074.pdf.
Listener A
Very nice indeed! Sound remarkably similar to my current Klotz MC 5000 reference: Big, wide-open sound with naural musical flow, offers bags of insight and detail into recordings, and instruments portrayed with realistic weight and authority. They reproduce ‘Explosive sounding’ transients.
I suspect that the Boston are probably the best cables I’ve heard that you’ve made. There’s certainly little between those and my favourite Klotz MC5000s. I would like to hear them made into shorter lengths. as I only need a 0.5m pair to go from my DAC to preamp.
Listener B
[This] we didn't like at all - the bass was very overpowering, and it was difficult to follow the different strands that the various players were contributing to a piece.
Listener C
With pop and rock music there was good bass but the mid and top were a little suppressed. However on classical music the cable was found to be neutral.
Listener D
Playing Joan Baez 5 (Villa Lobos), the Boston was a little more edgy than the Chord ‘Signature’ Tuned ARAY (~£765/stereo metre!), violin softer with good ambience. Bass definition was not quite as good as the Chords.
Cable 3 Belden 9907 Coax
http://i780.photobucket.com/albums/yy88/barrydhunt/012_zpsyqkjtwrz.jpg
This cable was designed as a thin Ethernet (“Thin-net”) interconnection. Essentially it is the same as RG-58A/U, but the screen/shield consists of a bonded aluminium foil-polyester tape-aluminium foil sandwich providing 100% coverage, with a tinned copper braid having 93% coverage. The electrical properties are very similar to those of Belden 8219, thus:
Characteristic impedance: 50 ± 2Ω.
Nominal loop resistance: 50mΩ/m.
Capacitance: 86pF/m.
Nominal delay: 4.2ns/m.
Surface transfer impedance: ~ 19mΩ/m. [1]
The presence of the aluminium foil sandwich in the shield makes these cables a little stiff - any cable bend tends to remain ‘set’, so these cables are probably best described as ‘semi-flexible’. As such, it is important not to bend these cables too tightly otherwise a ‘kink’ will form.
Cable cost ~ £2/m (when bought in bulk)
Plugs: Neutrik ‘Profi’
Listener A
These have a more musically cohesive presentation, with a less ‘spotlit’ midrange and lack of ‘grain’ [than the Trompeter cables (vide infra)]. Again, as with the ‘Bostons’, voices and instruments have realistic weight and ‘scale’, although transients weren’t quite reproduced with the same sense of dynamic alacrity and ‘attack’ available with the ‘Bostons’.
Listener B
[We] didn't think that any [of the cables] were an improvement to the cheap Belkin cables that the system is almost completely wired up with normally, but of course, they are familiar to us. Having said that, the one that came closest was the grey Belden one with the Neutrik Profi plugs.
Listener C
They made a noticeable difference to the sound of my CD player, especially:
a) greater 'attack' and a more lively presentation (but NOT increased brightness) probably because the notes start and stop with greater precision.
b) greater clarity of bass drums and cymbals (essential for me!).
c) a bit more space around the instruments.
d) wider sound stage.
I am using them to connect my CD player to the amplifier. Superb accuracy, treble, depth, pace, masses of detail coming off the CD, and a little wider soundstage was a little wider too. Lovely!!!!
Listener D
[When listening to] Joan Baez singing ‘Brasilieras Baccianas’ (Villa Lobos) there was good focus - everything flows. Good integration with good depth and ambience. Voilins were smooth, treble forte control best of all.
Cable 4 Trompeter - TC-75-2 (75Ω triaxial cable)
http://i780.photobucket.com/albums/yy88/barrydhunt/010_zpspjhtavsa.jpg
This cable is used for Ethernet connection. It is a triaxial cable, that is, there is a central coaxial cable with an additional screen braid. The two screens are electrically insulated from one another. The inner central conductor consists of 7 strands of AGW34 tinned copper wire (overall diameter 0.47mm) with a solid clear polyethylene dielectric (diameter 2.96mm), followed by the first single braid screen of tinned copper wire, having a coverage of 93%. The inner jacket (separating the two screens) is clear polyethylene. A second single braid screen of tinned copper wire also has an optical coverage of 93%. The outer jacket is yellow PVC. The electrical properties of this cable are as follows:
Characteristic impedance of the inner coaxial cable: 75 ± 2Ω
Inner conductor DC resistance: 145mΩ/m
Capacitance of inner coaxial cable: 67.3pF/m
Nominal delay: 5ns/m
Trompeter have not published the DC resistance of either screen, so it is not possible to quote the loop resistance, or to estimate the surface transfer impedance.
To simplify matters the cables were constructed with both screens connected in parallel, that is, both screens were connected together at the plugs at each end. These cable were shorter than the others, being only 0.675m long.
Cable cost ~ £5/m
Plugs: Maplin (code JU06G) http://www.maplin.co.uk/p/maplin-7mm-metal-barrel-gold-plated-phono-plug-red-ju06g
Listener A
Ok, but nothing special. Rather midrange prominent presentation, with a slightly ‘grainy’ sounding top-end, and overall lacklustre way of music-making, certainly in comparison with the musically entertaining and addictive ‘Bostons’ above.
Listener B
This came second after the Belden 9907 cable and was preferred over the Boston RG58 in the opinion of Listener B.
Listener D
When listening to Joan Baez ‘5’ (the Villa Lobos track) there was moderate focus and good articulation. Violins were smooth with good ambience, but a bit more edgy on treble forte.
Listening to Elkie Brooks’s ‘Shooting Star’ the (Trompeter) cables had a firmer, smoother and full-bodied bass, compared to the Van Damme Ultra Black cables. Overall the Trompeter cables were very slightly preferred to the Van Damme.
Cable 5 BICC - Uniradio 70
http://i780.photobucket.com/albums/yy88/barrydhunt/007_zps3ixf00xa.jpg
This is identical to URM70 cable, a 75 Ω characteristic impedance coaxial cable used for HF installations. It has a centre core consisting of 7 strands of 0.19mm copper wire, a solid polyethylene dielectric and a screen of bare copper wire, having an optical coverage of 91 ± 4%.
The electrical specification is as follows:
Characteristic impedance: 75 ± 3Ω
Inner conductor DC resistance: 92mΩ/m
Capacitance: 67 ± 2pF/m
Delay: 5ns/m
Cost: < £1.50/m
Plugs: Pro Signal (6mm cable diameter)
Listener A
Nicely detailed and musically expressive, but tends to add a ‘cuppy’ effect to vocals and midrange information. Also a little opaque sounding, overall, in comparison with my reference Klotz MC5000 with MS Audio plugs. A decent cable.
Listener B
Not bad. Maybe still a little too much bass. Quite forward (in the) treble as well.
Listener C
Compared to the others this cable was a bit of a let down - a real slow plodder that stripped away a layer of detail right across the board. Mids were especially flat and too much bass weight was missing. A decent sound stage was welcome but in a way it actually hurt the music because it provided staging for what seemed like a collection of lifeless sounding vocals and instruments. It's not for me this one.
Cable 6 Huber & Suhner RG223/U-01
http://i780.photobucket.com/albums/yy88/barrydhunt/006_zpseuemc5cx.jpg
This is a precision coaxial cable made by the Swiss company Huber & Suhner AG. RG223/U cable consists of a single core of 0.9mm diameter silver-plated copper, with a polyethylene dielectric, followed by two silver-plated copper wire braided shields. The two shields are in electrical contact with one another throughout the length of the cable.
The suffix “– 01” denotes the fact that the tolerance on the cable’s characteristic impedance is ± 1Ω, whereas for most RG cables of 50Ω impedance it is ± 2Ω. Of course the characteristic impedance of a cable is totally irrelevant at audio frequencies. What the tighter tolerance reflects is the tighter tolerance on the cable dimensions and a low degree of deviation from a true coaxial geometry (eccentricity); hence the designation ‘precision cable’.
In employing a double braided screen (having an optical coverage of 96% and 94% respectively), the surface transfer impedance is very low (0.18mΩ/m at 30MHz [1]), so the cable is very effective at being immune to RF pick up. The electrical properties are as follows:
Characteristic impedance: 50 ± 1Ω
Nominal loop resistance: 34.4mΩ/m
Capacitance: 101pF/m
Nominal delay; 5ns/m
Cost: £2.50 - £13/m. (depending on manufacturer and supplier, this is for the ‘standard’ RG-223/U and when bought in bulk) This is the most expensive cable used here.
Plugs: Pro Signal (6mm cable diameter)
Listener A
Better in my system than the URM70 cable. This cable has a nicely ‘fluid’ quality to its musical presentation, also gives acoustic instruments appropriate leading-edge ‘attack’. Teases out subtle information in the mix which was missed by the URM70.
Listener B
Excellent - very close to our usual cable choice, perhaps very slightly more full and rich sounding. The only real criticism was that the height of the stereo image seemed reduced.
Listener C
To be honest it sounded very similar to the Maplin cables (vide infra). Possibly a slightly more aggressive reproduction of higher frequencies and a bit of bass detail is missing BUT, we're really splitting hairs here (ones that are a few thou thick!).
Listener D
These cables were, until recently, this listener’s standard audio interconnect. Only the Chord ‘Signature ‘ Tuned ARAY (at ~ £765 for a stereo metre!) bettered it.
And a latecomer to the party:
Cable 7 Maplin microphone cable (Code XS24B)
http://i780.photobucket.com/albums/yy88/barrydhunt/013_zpslxeuloz6.jpg
This is a single screened cable, which has 20 x 0.12mm copper conductors (conductor cross section area: 0.23mm², that is a diameter of 0.54mm) with PVC insulation, a special semiconducting screen layer, a lapped copper screen and two layers of PVC outer insulation (outer black). It is designed to be used as a flexible microphone cable. The semi-conducting screen layer is to minimise ‘handling noises’ by discharging triboelectric generated charges.
Capacitance: 135pF/m
Cost: < £2.50/m
Plugs: Neutrik ‘Profi’
Listener A
Don’t really like this cable, as it seemed to make everything sound rather ‘bland’, and consequently rather than music that I know to sound very dynamic and engaging, I found it washing over me in a rather ‘matter-of-fact’ way.
Listener B
Awful bass - one note and made bass guitar sound like a slack drum covered with a folded blanket(!) Hard to follow the different elements when the recording got busy. Loss of fine detail - there were some subtle, low level sounds missing from the recording.
Listener C
It offers pretty much the same as my QED cable. A good balanced cable: well controlled bass, smooth mids and clear grain free treble. Pace and timing are good and the soundstage is wide with reasonable depth. BUT, (this cable) does better my QED in just two areas! The treble is very good. Cymbal sounds are excellent - clear, sharp and with a nice decay (sorry, I used to play the drums!). The cable has no trouble coping with distorted electric guitar sounds either. The other difference is a wider soundstage. Not by a great deal but it is there.
And that's it; no game changer or anything special - just a good 'honest' cable that I would be happy to own (for around £40).
Summary
In order of preference, the Listeners ranked the cables as follows:
Listener A
Boston RG58 > Belden 9907 > Trompeter
and
Belden 8219 RG58 > Suhner RG223 > URM70 > Maplin
Listener B
Belden 9907 > Trompeter > Boston RG58
and
Suhner RG223 > URM70 > Belden 9219 RG58 > Maplin
Listener C
Belden 9907 > Trompeter and the Boston RG58
and
Belden 8219 RG58 > Suhner RG223 > Maplin > URM70
Listener D
Suhner RG223 > Trompeter > Belden 9907 and the Boston RG58
Most Listeners in general, liked both the Belden 8219 RG58 and Suhner RG223 cables, whereas all universally disliked the Maplin cable. However one listener was at odds with the others concerning the RG58 type cables (the Boston and the Belden 8219), being unimpressed by both of them.
All liked the Belden 9907 cable; so much so by one listener, that he insisted I make a set for him.
Both the Trompeter and URM70 cables found favour with some Listeners, but not all.
Conclusion (of sorts)
I started this exercise to see if cables had an intrinsic sound of their own (and thus would received equally by all listeners), or if the change in sound quality perceived when the cables were swapped in and out of listener’s systems would, itself, depend on the individual systems. So has the exercise shown this to be true, one way or another – and if cables do have an intrinsic sound, is it in any way related to the measured electrical specification of the cable?
Unfortunately the findings of the Listeners were sufficiently ambiguous, so the question(s) cannot be answered. All liked the Belden 9907 and the Suhner RG223 cable, so the finding of favour with these cables was system independent, and there are some technical reasons why this might be so (such as good immunity to RF pick-up). Likewise everyone disliked the Maplin cable. This might be due to the use of PVC for the dielectric; yet it is curious that these cables received (much to my surprise) a good reception when I submitted them for a blind cable comparison in London about three years ago, and was subsequently asked by a few AoS members to have sets made for them.
The two other RG58 type cables (the Belden 8219 and the Boston) perhaps demonstrate system dependence: both being well received by one Listener, yet disliked by another. It also shows that one RG58 type cable made by one manufacturer may sound different to that from another.
I should like to thank all those who took part in the exercise, for their patience and their honesty in reporting their findings. Sadly it has only been a partial success – but I hope you have found it interesting.
Notes
[1] Another quantity, which goes some way to describing the effectiveness of the cable shielding or screening and the immunity of the cable to RF pick-up, is the surface transfer impedance (Zt); but this is virtually impossible to measure at audio frequencies. Cable manufacturers rarely quote it, because to measure it requires specialised equipment. The transfer impedance can be calculated for a solid outer screen but such calculation is of little use to predict the surface transfer impedance of a woven braided shield.
At audio frequencies, the surface transfer impedance for a shield having an optical coverage of 100% is virtually identical to the DC resistance of the shield per unit length. However this is of little use when it comes to calculating the shielding effectiveness. In systems operating at radio frequencies, cables are usually terminated in their characteristic impedance (Zo) and the shielding effectiveness (in dB) can be calculated as:
20.log (Zt/2.Zo).
Unfortunately the impedance of the two items between which the cable connects, whilst known at audio frequencies (the source impedance of, say, ~ 100Ω and a load impedance of ~ 10kΩ) are unknown at radio frequencies, so it is impossible to calculate the shielding effectiveness of the cable in this situation.
The best that can be done is to note that the IEC specification for a single braid cable is 100mΩ/m at 30MHz. The shielding effectiveness of the various cables considered here can either be ranked by comparing their surface transfer impedance at low frequencies (~ the DC resistance per unit length of the screen), or to compare the surface transfer impedance at 30Mz with the IEC specification. This latter has only been measured for a few cables:
RG-58: 125mΩ/m
RG-174: 100mΩ/m
RG-223: 0.18mΩ/m (double braid),
and
URM70: 30mΩ/m
[2] If one ignores the loop resistance of the cable and assumes the load impedance is very much greater than the source impedance, then the low frequency roll-off point of the low-pass network thus formed, is largely determined by the capacitance and the output impedance of the source item. The roll-off frequency is then given by
[I]F = 1/(2πCR)
Where C is the total capacitance of the cable and R is the output impedance of the source.
For valve gear the source output impedance is ~ 1kΩ, whereas the load impedance is ~ 1MΩ, so the load impedance can be ignored. Similarly for solid-state gear, the source impedance is ~ 100Ω and the load impedance is > 10kΩ.
Usually both the cable capacitance and source output impedance are quite low, so placing the roll-off frequency well beyond the range of hearing. For example even if the output impedance was as high as 5kΩ, and the cable have a typical capacitance of 100pF/m, the cables would need to be16m long before the roll-off point fell to 20kHz.
Postscript
At the time of these tests, I speculated whether the Trompeter 75Ω triaxial cable might make a good digital cable. I tested the idea with one listener, with such favourable results that he now uses a 1.8m length between his CD transport and DAC.
With this interconnect the second, outer, screen is connected to the screen of the inner coaxial cable at one end only. This makes the cable directional and allows the user to experiment.
Introduction
There has been much discussion here of commercially available VFM interconnects, using cable as supplied to professional recording studios and broadcast institutions. These cables are available ‘off the drum’ and are inexpensive to buy. Some of these cables have been designed for use with hand-held microphones, so are very flexible and often have a semi-conducting layer between the insulating dielectric and the outer screen, to discharge any static charge caused by cable flexing (so-called ‘handling noise’).
This report concerns itself with the audition of audio interconnects, again using VFM cable, but this time most of the cables are those that would be used in the aerospace, computer and other industries.
As is well appreciated, the whole discussion of audio cables polarises the audio community, often causing much dissention and heated debate. It might be useful at this stage to briefly outline the philosophy and thinking of the two ‘camps’. They can be conveniently divided into “objectivists” and “subjectivists”.
Those belonging to the objectivist camp would argue that cables are entirely passive constructs and are electrically so simple (relative to the complexity of the items between which they connect) they cannot have much influence on the overall sound quality. Cables, according to objectivists are qualified by three simple values: the loop resistance per unit length (which depends on the metal used for the conductors), the capacitance per unit length (determined by the geometry of the cable and the dielectric insulation used) and the inductance per unit length (which again, depends on the cable geometry) [1]. Of these three values, the most important is the capacitance per unit length. When cables are used as interconnects, their capacitance will form a filter network with the impedance of the items between which they connect. [2]
Subjectivists on the other hand would rightly argue that any audio system is a concatenation of components, including the interconnecting cables. Thus every part of the chain is as important as every other part, regardless of the electrical complexity of the item. It therefore follows that cables contribute to the overall sound quality of the system just as much as the more complex items of electronics used. Subjectivists will claim they can readily hear the changes made by exchanging interconnects to the overall sound and these changes can be quite large; large enough to justify spending a considerable amount on interconnects if necessary.
At this stage I need to state my position. As someone who has had a scientific training and having worked in an electronics research laboratory for thirty years, I’m inclined to put myself in the objectivist camp and am somewhat sceptical of so-called cable sound. Yet there have been times (much to my surprise) where I have heard audible differences made through changing cables. The sound differences have always been small, but distinct and most importantly, repeatable. So I find myself being unable to explain why I can occasionally hear these differences, but willing to accept they can occur; and more importantly, that other listeners can perhaps more readily hear them.
The purpose of this exercise was thus to try and find out whether cables have an intrinsic sound of their own, or whether the changes in sound quality heard is system dependent.
Methodology
To this end I made up six sets of cables, using conventional coaxial cable as supplied to and used in the aerospace, computer and hydrology industry. These cables were then sent to four other listeners, having four quite separate audio systems and listening requirements. Each listener spent time with the cables (a few weeks) and recorded their observations. In three cases (as well as my own), the cables were used between a CD player or DAC and the preamp. One listener preferred to only use vinyl as the source material, so the cables were fitted between his phono stage and preamp.
I don’t claim the exercise was in any way scientifically rigorous – it was not blind (the cables were not identified, they were simply given a code, but on some the identity of the cable is printed on the outer jacket); the cables were not of identical length (most were 1 metre long), and the RCA phono plugs used were not common to all. None of the cables made used expensive cable and most were fitted with inexpensive RCA plugs of reasonable quality.
The cables are listed in turn along with their technical specification, typical cost per metre and a brief description of the RCA plugs fitted. Listener’s comments then follow; and for all the cables, the listener’s comments appear in the same order.
The cables
Cable 1 Belden 8219 Coax – RG-58A/U
http://i780.photobucket.com/albums/yy88/barrydhunt/008_zpsemdknvx4.jpg
This is an RG-58A/U type cable; consisting of a 20 AWG (19x32) stranded tinned copper inner conductor (0.94mm diameter), foam polyethylene dielectric and a tinned copper braid shield having 95% optical coverage. The outer insulation is PVC.
The characteristic impedance is 50Ω (though this is totally irrelevant at audio frequencies).
Nominal loop resistance: 42.3mΩ/m.
Capacitance: 87pF/m.
Nominal delay: 4.5ns/m.
Nominal surface transfer impedance: ~ 13.5mΩ/m [1]
Typical cost: < £1.50/m
Plugs: Neutrik ‘REAN’
Listener’s comments:
Listener A
Unquestionably my favourite cables of the bunch. Very detailed and open sounding, with no hint of harshness or artifice. Lets the music through relatively unhindered. I could live with them, although swapping back to my reference Klotz revealed a slight lack of spaciousness, ‘musical flow’, and ultimate resolution.
Listener B
Way too much bass. One note bass. Everything ill defined - hard to separate the different elements of the performance. Very little air or ambiance where it should be.
Listener C
This is one aggressive, forward sounding cable with plenty of slam. I would imagine that an extended listening session involving a lot of bright sounding CDs would see your ears sliced off! The aggressive bright sound gave the impression that there was a lot of detail retrieval but at what cost? Can't fault the bass though - powerful, detailed and with some real slam. Great soundstage too. Mids were a bit screechy and thin sounding - okay, but the more natural and fuller tones of (the Maplin microphone cable) were better. Clearly a cable that needs careful system matching but in the right set up could be a bit of a star. I suspect 'What Hifi' testers would love it.
was something of an enigma (can a cable be described this way...?) and potentially a real star. You certainly get a lot of detail retrieval and a fast punchy sound but at times I could barely stand it!!
Cable 2 Boston Insulated Wire (UK) - RG-58 hydrophone cable
http://i780.photobucket.com/albums/yy88/barrydhunt/005_zpsxnu9lazv.jpg?t=1432233356
This cable is designed for use with hydrophones. Essentially it consists of an RG58 coaxial cable, clad with a tough silicone outer sheath. Hydrophones use the piezoelectric effect, so it is important the hydrophone signal is not distorted by any change in capacitance of the connecting cable due to dimensional deformation caused by pressure of the water. Hence these cable need a thick (7.8mm OD), tough and waterproof cladding.
Being an RG-58 type cable, the construction and performance details are very similar to those for the Belden 8219 listed above.
Unfortunately the thick and tough, silicone plastic cladding makes this cable extremely difficult to work with as well as making the cables ‘springy’ and not particularly flexible.
I cannot find any information about Boston Insulated Wire (UK), Esher, Surrey, UK, so I assume they are no longer in business. Cost (if still available): probably ~ £2.50/m.
Plugs: Pro Signal (8mm cable diameter)
I like these plugs because the screen and outer insulation are gripped by a collet, preserving a true coaxial geometry. http://www.farnell.com/datasheets/320074.pdf.
Listener A
Very nice indeed! Sound remarkably similar to my current Klotz MC 5000 reference: Big, wide-open sound with naural musical flow, offers bags of insight and detail into recordings, and instruments portrayed with realistic weight and authority. They reproduce ‘Explosive sounding’ transients.
I suspect that the Boston are probably the best cables I’ve heard that you’ve made. There’s certainly little between those and my favourite Klotz MC5000s. I would like to hear them made into shorter lengths. as I only need a 0.5m pair to go from my DAC to preamp.
Listener B
[This] we didn't like at all - the bass was very overpowering, and it was difficult to follow the different strands that the various players were contributing to a piece.
Listener C
With pop and rock music there was good bass but the mid and top were a little suppressed. However on classical music the cable was found to be neutral.
Listener D
Playing Joan Baez 5 (Villa Lobos), the Boston was a little more edgy than the Chord ‘Signature’ Tuned ARAY (~£765/stereo metre!), violin softer with good ambience. Bass definition was not quite as good as the Chords.
Cable 3 Belden 9907 Coax
http://i780.photobucket.com/albums/yy88/barrydhunt/012_zpsyqkjtwrz.jpg
This cable was designed as a thin Ethernet (“Thin-net”) interconnection. Essentially it is the same as RG-58A/U, but the screen/shield consists of a bonded aluminium foil-polyester tape-aluminium foil sandwich providing 100% coverage, with a tinned copper braid having 93% coverage. The electrical properties are very similar to those of Belden 8219, thus:
Characteristic impedance: 50 ± 2Ω.
Nominal loop resistance: 50mΩ/m.
Capacitance: 86pF/m.
Nominal delay: 4.2ns/m.
Surface transfer impedance: ~ 19mΩ/m. [1]
The presence of the aluminium foil sandwich in the shield makes these cables a little stiff - any cable bend tends to remain ‘set’, so these cables are probably best described as ‘semi-flexible’. As such, it is important not to bend these cables too tightly otherwise a ‘kink’ will form.
Cable cost ~ £2/m (when bought in bulk)
Plugs: Neutrik ‘Profi’
Listener A
These have a more musically cohesive presentation, with a less ‘spotlit’ midrange and lack of ‘grain’ [than the Trompeter cables (vide infra)]. Again, as with the ‘Bostons’, voices and instruments have realistic weight and ‘scale’, although transients weren’t quite reproduced with the same sense of dynamic alacrity and ‘attack’ available with the ‘Bostons’.
Listener B
[We] didn't think that any [of the cables] were an improvement to the cheap Belkin cables that the system is almost completely wired up with normally, but of course, they are familiar to us. Having said that, the one that came closest was the grey Belden one with the Neutrik Profi plugs.
Listener C
They made a noticeable difference to the sound of my CD player, especially:
a) greater 'attack' and a more lively presentation (but NOT increased brightness) probably because the notes start and stop with greater precision.
b) greater clarity of bass drums and cymbals (essential for me!).
c) a bit more space around the instruments.
d) wider sound stage.
I am using them to connect my CD player to the amplifier. Superb accuracy, treble, depth, pace, masses of detail coming off the CD, and a little wider soundstage was a little wider too. Lovely!!!!
Listener D
[When listening to] Joan Baez singing ‘Brasilieras Baccianas’ (Villa Lobos) there was good focus - everything flows. Good integration with good depth and ambience. Voilins were smooth, treble forte control best of all.
Cable 4 Trompeter - TC-75-2 (75Ω triaxial cable)
http://i780.photobucket.com/albums/yy88/barrydhunt/010_zpspjhtavsa.jpg
This cable is used for Ethernet connection. It is a triaxial cable, that is, there is a central coaxial cable with an additional screen braid. The two screens are electrically insulated from one another. The inner central conductor consists of 7 strands of AGW34 tinned copper wire (overall diameter 0.47mm) with a solid clear polyethylene dielectric (diameter 2.96mm), followed by the first single braid screen of tinned copper wire, having a coverage of 93%. The inner jacket (separating the two screens) is clear polyethylene. A second single braid screen of tinned copper wire also has an optical coverage of 93%. The outer jacket is yellow PVC. The electrical properties of this cable are as follows:
Characteristic impedance of the inner coaxial cable: 75 ± 2Ω
Inner conductor DC resistance: 145mΩ/m
Capacitance of inner coaxial cable: 67.3pF/m
Nominal delay: 5ns/m
Trompeter have not published the DC resistance of either screen, so it is not possible to quote the loop resistance, or to estimate the surface transfer impedance.
To simplify matters the cables were constructed with both screens connected in parallel, that is, both screens were connected together at the plugs at each end. These cable were shorter than the others, being only 0.675m long.
Cable cost ~ £5/m
Plugs: Maplin (code JU06G) http://www.maplin.co.uk/p/maplin-7mm-metal-barrel-gold-plated-phono-plug-red-ju06g
Listener A
Ok, but nothing special. Rather midrange prominent presentation, with a slightly ‘grainy’ sounding top-end, and overall lacklustre way of music-making, certainly in comparison with the musically entertaining and addictive ‘Bostons’ above.
Listener B
This came second after the Belden 9907 cable and was preferred over the Boston RG58 in the opinion of Listener B.
Listener D
When listening to Joan Baez ‘5’ (the Villa Lobos track) there was moderate focus and good articulation. Violins were smooth with good ambience, but a bit more edgy on treble forte.
Listening to Elkie Brooks’s ‘Shooting Star’ the (Trompeter) cables had a firmer, smoother and full-bodied bass, compared to the Van Damme Ultra Black cables. Overall the Trompeter cables were very slightly preferred to the Van Damme.
Cable 5 BICC - Uniradio 70
http://i780.photobucket.com/albums/yy88/barrydhunt/007_zps3ixf00xa.jpg
This is identical to URM70 cable, a 75 Ω characteristic impedance coaxial cable used for HF installations. It has a centre core consisting of 7 strands of 0.19mm copper wire, a solid polyethylene dielectric and a screen of bare copper wire, having an optical coverage of 91 ± 4%.
The electrical specification is as follows:
Characteristic impedance: 75 ± 3Ω
Inner conductor DC resistance: 92mΩ/m
Capacitance: 67 ± 2pF/m
Delay: 5ns/m
Cost: < £1.50/m
Plugs: Pro Signal (6mm cable diameter)
Listener A
Nicely detailed and musically expressive, but tends to add a ‘cuppy’ effect to vocals and midrange information. Also a little opaque sounding, overall, in comparison with my reference Klotz MC5000 with MS Audio plugs. A decent cable.
Listener B
Not bad. Maybe still a little too much bass. Quite forward (in the) treble as well.
Listener C
Compared to the others this cable was a bit of a let down - a real slow plodder that stripped away a layer of detail right across the board. Mids were especially flat and too much bass weight was missing. A decent sound stage was welcome but in a way it actually hurt the music because it provided staging for what seemed like a collection of lifeless sounding vocals and instruments. It's not for me this one.
Cable 6 Huber & Suhner RG223/U-01
http://i780.photobucket.com/albums/yy88/barrydhunt/006_zpseuemc5cx.jpg
This is a precision coaxial cable made by the Swiss company Huber & Suhner AG. RG223/U cable consists of a single core of 0.9mm diameter silver-plated copper, with a polyethylene dielectric, followed by two silver-plated copper wire braided shields. The two shields are in electrical contact with one another throughout the length of the cable.
The suffix “– 01” denotes the fact that the tolerance on the cable’s characteristic impedance is ± 1Ω, whereas for most RG cables of 50Ω impedance it is ± 2Ω. Of course the characteristic impedance of a cable is totally irrelevant at audio frequencies. What the tighter tolerance reflects is the tighter tolerance on the cable dimensions and a low degree of deviation from a true coaxial geometry (eccentricity); hence the designation ‘precision cable’.
In employing a double braided screen (having an optical coverage of 96% and 94% respectively), the surface transfer impedance is very low (0.18mΩ/m at 30MHz [1]), so the cable is very effective at being immune to RF pick up. The electrical properties are as follows:
Characteristic impedance: 50 ± 1Ω
Nominal loop resistance: 34.4mΩ/m
Capacitance: 101pF/m
Nominal delay; 5ns/m
Cost: £2.50 - £13/m. (depending on manufacturer and supplier, this is for the ‘standard’ RG-223/U and when bought in bulk) This is the most expensive cable used here.
Plugs: Pro Signal (6mm cable diameter)
Listener A
Better in my system than the URM70 cable. This cable has a nicely ‘fluid’ quality to its musical presentation, also gives acoustic instruments appropriate leading-edge ‘attack’. Teases out subtle information in the mix which was missed by the URM70.
Listener B
Excellent - very close to our usual cable choice, perhaps very slightly more full and rich sounding. The only real criticism was that the height of the stereo image seemed reduced.
Listener C
To be honest it sounded very similar to the Maplin cables (vide infra). Possibly a slightly more aggressive reproduction of higher frequencies and a bit of bass detail is missing BUT, we're really splitting hairs here (ones that are a few thou thick!).
Listener D
These cables were, until recently, this listener’s standard audio interconnect. Only the Chord ‘Signature ‘ Tuned ARAY (at ~ £765 for a stereo metre!) bettered it.
And a latecomer to the party:
Cable 7 Maplin microphone cable (Code XS24B)
http://i780.photobucket.com/albums/yy88/barrydhunt/013_zpslxeuloz6.jpg
This is a single screened cable, which has 20 x 0.12mm copper conductors (conductor cross section area: 0.23mm², that is a diameter of 0.54mm) with PVC insulation, a special semiconducting screen layer, a lapped copper screen and two layers of PVC outer insulation (outer black). It is designed to be used as a flexible microphone cable. The semi-conducting screen layer is to minimise ‘handling noises’ by discharging triboelectric generated charges.
Capacitance: 135pF/m
Cost: < £2.50/m
Plugs: Neutrik ‘Profi’
Listener A
Don’t really like this cable, as it seemed to make everything sound rather ‘bland’, and consequently rather than music that I know to sound very dynamic and engaging, I found it washing over me in a rather ‘matter-of-fact’ way.
Listener B
Awful bass - one note and made bass guitar sound like a slack drum covered with a folded blanket(!) Hard to follow the different elements when the recording got busy. Loss of fine detail - there were some subtle, low level sounds missing from the recording.
Listener C
It offers pretty much the same as my QED cable. A good balanced cable: well controlled bass, smooth mids and clear grain free treble. Pace and timing are good and the soundstage is wide with reasonable depth. BUT, (this cable) does better my QED in just two areas! The treble is very good. Cymbal sounds are excellent - clear, sharp and with a nice decay (sorry, I used to play the drums!). The cable has no trouble coping with distorted electric guitar sounds either. The other difference is a wider soundstage. Not by a great deal but it is there.
And that's it; no game changer or anything special - just a good 'honest' cable that I would be happy to own (for around £40).
Summary
In order of preference, the Listeners ranked the cables as follows:
Listener A
Boston RG58 > Belden 9907 > Trompeter
and
Belden 8219 RG58 > Suhner RG223 > URM70 > Maplin
Listener B
Belden 9907 > Trompeter > Boston RG58
and
Suhner RG223 > URM70 > Belden 9219 RG58 > Maplin
Listener C
Belden 9907 > Trompeter and the Boston RG58
and
Belden 8219 RG58 > Suhner RG223 > Maplin > URM70
Listener D
Suhner RG223 > Trompeter > Belden 9907 and the Boston RG58
Most Listeners in general, liked both the Belden 8219 RG58 and Suhner RG223 cables, whereas all universally disliked the Maplin cable. However one listener was at odds with the others concerning the RG58 type cables (the Boston and the Belden 8219), being unimpressed by both of them.
All liked the Belden 9907 cable; so much so by one listener, that he insisted I make a set for him.
Both the Trompeter and URM70 cables found favour with some Listeners, but not all.
Conclusion (of sorts)
I started this exercise to see if cables had an intrinsic sound of their own (and thus would received equally by all listeners), or if the change in sound quality perceived when the cables were swapped in and out of listener’s systems would, itself, depend on the individual systems. So has the exercise shown this to be true, one way or another – and if cables do have an intrinsic sound, is it in any way related to the measured electrical specification of the cable?
Unfortunately the findings of the Listeners were sufficiently ambiguous, so the question(s) cannot be answered. All liked the Belden 9907 and the Suhner RG223 cable, so the finding of favour with these cables was system independent, and there are some technical reasons why this might be so (such as good immunity to RF pick-up). Likewise everyone disliked the Maplin cable. This might be due to the use of PVC for the dielectric; yet it is curious that these cables received (much to my surprise) a good reception when I submitted them for a blind cable comparison in London about three years ago, and was subsequently asked by a few AoS members to have sets made for them.
The two other RG58 type cables (the Belden 8219 and the Boston) perhaps demonstrate system dependence: both being well received by one Listener, yet disliked by another. It also shows that one RG58 type cable made by one manufacturer may sound different to that from another.
I should like to thank all those who took part in the exercise, for their patience and their honesty in reporting their findings. Sadly it has only been a partial success – but I hope you have found it interesting.
Notes
[1] Another quantity, which goes some way to describing the effectiveness of the cable shielding or screening and the immunity of the cable to RF pick-up, is the surface transfer impedance (Zt); but this is virtually impossible to measure at audio frequencies. Cable manufacturers rarely quote it, because to measure it requires specialised equipment. The transfer impedance can be calculated for a solid outer screen but such calculation is of little use to predict the surface transfer impedance of a woven braided shield.
At audio frequencies, the surface transfer impedance for a shield having an optical coverage of 100% is virtually identical to the DC resistance of the shield per unit length. However this is of little use when it comes to calculating the shielding effectiveness. In systems operating at radio frequencies, cables are usually terminated in their characteristic impedance (Zo) and the shielding effectiveness (in dB) can be calculated as:
20.log (Zt/2.Zo).
Unfortunately the impedance of the two items between which the cable connects, whilst known at audio frequencies (the source impedance of, say, ~ 100Ω and a load impedance of ~ 10kΩ) are unknown at radio frequencies, so it is impossible to calculate the shielding effectiveness of the cable in this situation.
The best that can be done is to note that the IEC specification for a single braid cable is 100mΩ/m at 30MHz. The shielding effectiveness of the various cables considered here can either be ranked by comparing their surface transfer impedance at low frequencies (~ the DC resistance per unit length of the screen), or to compare the surface transfer impedance at 30Mz with the IEC specification. This latter has only been measured for a few cables:
RG-58: 125mΩ/m
RG-174: 100mΩ/m
RG-223: 0.18mΩ/m (double braid),
and
URM70: 30mΩ/m
[2] If one ignores the loop resistance of the cable and assumes the load impedance is very much greater than the source impedance, then the low frequency roll-off point of the low-pass network thus formed, is largely determined by the capacitance and the output impedance of the source item. The roll-off frequency is then given by
[I]F = 1/(2πCR)
Where C is the total capacitance of the cable and R is the output impedance of the source.
For valve gear the source output impedance is ~ 1kΩ, whereas the load impedance is ~ 1MΩ, so the load impedance can be ignored. Similarly for solid-state gear, the source impedance is ~ 100Ω and the load impedance is > 10kΩ.
Usually both the cable capacitance and source output impedance are quite low, so placing the roll-off frequency well beyond the range of hearing. For example even if the output impedance was as high as 5kΩ, and the cable have a typical capacitance of 100pF/m, the cables would need to be16m long before the roll-off point fell to 20kHz.
Postscript
At the time of these tests, I speculated whether the Trompeter 75Ω triaxial cable might make a good digital cable. I tested the idea with one listener, with such favourable results that he now uses a 1.8m length between his CD transport and DAC.
With this interconnect the second, outer, screen is connected to the screen of the inner coaxial cable at one end only. This makes the cable directional and allows the user to experiment.