Coax cable 50 ohm for audio interconnects and speaker hookup

[Light Dependant Resistor]

Hi
I have always been impressed with the audio improvement that results from using coax cable in audio installations. For interconnects RG174 which is a 50 ohm cable has nice features, as it is flexible having a tight turn ability.

The subject of cables at audio frequency of its use was investigated by Cyril Bateman writing for Electronics World, and more recently his republished article finds speaker cables do comply with established transmission line behaviour. Cyrils article is the one titled Cables at AF ( Audio Frequency ) here: http://web.archive.org/web/20091027010126/http://uk.geocities.com/cyrilb2@btinternet.com/downloads_5.html

For speakers there are cheap grades of RG58 with copper solid core centres, so avoid them as they are too brittle and difficult to work with. Rather try and source RG58 that allows a bit of flexibility with silver stranded wire inner, For speakers just use the inner conductor - don't extend the shield to be ground. so you end up with four wires in a stereo speaker installation the ground shield that serves purpose still as a shield however not grounded is cut back to not get in the way. RG58 is also a 50 ohm coax. If you use the typical black variety it may need cable tieing where it gets close to speaker terminals.

Some time is needed to assess the complexity of Cyril's findings to then say what is best in each installation, but generally 50 ohm coax is an audio users best friend.

Cheers / Chris

[Light Dependant Resistor]

A few hours reading Cyrils article finds he advocates a speaker end zobel network then matching the cable in use, fascinating reading. Will give it a try and let you know results.

[StanleyB]

I wrote a short article describing the construction of a cheap but effective digital coax cable some 8 years ago. It was on another forum, and caused so much trouble with a couple of "established" cable sellers, that I got ejected from the site. Marco was one of my backers with regards to freedom to disclose such technical secrets. He ended up getting ejected as well. Since then I have shared the info to Mark from Mark Grant Cables, who I assume has made good use of it.

But 50 Ohms cable for speakers is a bad idea. The Zobel network at an amp's output is normally set for a 10 Ohms impedance. The combination of a 10 Ohms Zobel network, 50 Ohms speaker cable, and 8 Ohms speaker crossover input can form a nice T filter network that is going to dampen the top and bottom, and give you a raised mids.

[Reffc]

I wrote a short article describing the construction of a cheap but effective digital coax cable some 8 years ago. It was on another forum, and caused so much trouble with a couple of "established" cable sellers, that I got ejected from the site. Marco was one of my backers with regards to freedom to disclose such technical secrets. He ended up getting ejected as well. Since then I have shared the info to Mark from Mark Grant Cables, who I assume has made good use of it.

But 50 Ohms cable for speakers is a bad idea. The Zobel network at an amp's output is normally set for a 10 Ohms impedance. The combination of a 10 Ohms Zobel network, 50 Ohms speaker cable, and 8 Ohms speaker crossover input can form a nice T filter network that is going to dampen the top and bottom, and give you a raised mids.

Sorry to hear that you were badly treated Stanley. Some traders do unfortunately bring the trade into disrepute, and there can be too much of a symbiosis between those traders and the forum owners (but not on AoS which is a good thing).

RE the highlighted cables issues, exactly this! It's a bad idea. Also, to be effective as a shield, all shields have to be grounded. Speaker cables by and large are less susceptible to RFI (unless you live opposite a taxi rank!) so shielding simply isn't required, but keeping the spacing controlled and adding twist helps the cable not to act as an effective loop antenna.

Ditto, zobels nor any other circuits are really necessary. Low loop impedance is the most important thing to aim for. Adding any unwanted capacitance is a bad thing.

[Light Dependant Resistor]

Sorry to hear that you were badly treated Stanley. Some traders do unfortunately bring the trade into disrepute, and there can be too much of a symbiosis between those traders and the forum owners (but not on AoS which is a good thing).

RE the highlighted cables issues, exactly this! It's a bad idea. Also, to be effective as a shield, all shields have to be grounded. Speaker cables by and large are less susceptible to RFI (unless you live opposite a taxi rank!) so shielding simply isn't required, but keeping the spacing controlled and adding twist helps the cable not to act as an effective loop antenna.

Ditto, zobels nor any other circuits are really necessary. Low loop impedance is the most important thing to aim for. Adding any unwanted capacitance is a bad thing.

Discussing Zobels one must define where it is placed. in relation to amplifier Zobel networks I doubt many amplifier manufacturers ( possibly Naim as an exception who then market their speaker cable having similar characteristics to be used in conjunction ) would release product on the marketplace without a zobel network. A zobel assists the amplifier to drive speaker leads ( not speakers as such, even though they are there )

The shield when using coax as speaker lead as stated is not grounded for speaker use, rather forms a shield along its length without 0v reference. Twisted cable and coax that has tightly controlled spacing both have benefit in audio use, and it is inexpensive to assess the merits of one against the other relative to your audio system. For interconnects with suggested RG174 the shield is of course grounded.

Cyrils article is interesting as it points to equations that attempt to assist the speaker end of the speaker lead in the process assessing the speaker leads electrical properties, in theory three zobels can then be used in amplifier/speaker relationship, one for the amplifier so it can assist driving speaker cables, one at the speaker cable end as cyril discusses for assisting that cable end, and another to counteract the rising impedance of a voice coil located after the crossover network within a speaker.

[Reffc]

Discussing Zobels one must define where it is placed. in relation to amplifier Zobel networks I doubt many amplifier manufacturers ( possibly Naim as an exception who then market their speaker cable having similar characteristics to be used in conjunction ) would release product on the marketplace without a zobel network. A zobel assists the amplifier to drive speaker leads ( not speakers as such, even though they are there )

The shield when using coax as speaker lead as stated is not grounded for speaker use, rather forms a shield along its length without 0v reference. Twisted cable and coax that has tightly controlled spacing both have benefit in audio use, and it is inexpensive to assess the merits of one against the other relative to your audio system. For interconnects with suggested RG174 the shield is of course grounded.

Cyrils article is interesting as it points to equations that attempt to assist the speaker end of the speaker lead in the process assessing the speaker leads electrical properties, in theory three zobels can then be used in amplifier/speaker relationship, one for the amplifier so it can assist driving speaker cables, one at the speaker cable end as cyril discusses for assisting that cable end, and another to counteract the rising impedance of a voice coil located after the crossover network within a speaker.

RE Zobels, sorry, my presumption was that the inference was placing a zobel at the end of a speaker cable run (as these are commercially provided by some companies), not amplifier zobels. Yes, zobels within the output stage of an amplifier are eminently sensible. Agreed RE the comparisons between shielded/twisted, but having undertaken these comparisons, my own findings were that shielded speaker cables were of very limited value/possibly no value at all, and also, there's little to recommend using single core coax as single conductor speaker cable especially given that for anything other than short runs, most coax uses insufficient copper for the conductor to be used as a speaker cable imho. That is not what it is intended for, although overall screened balanced cables using 1.5mm cores would be fine. Speaker cables I think are understood quite well enough in their role as part of the circuit encompassing amplifier, speaker cable and loudspeaker, so selection of an appropriate cable for eat circuit is a relatively simple thing. I can't see any new research (or old for that matter) resulting in any earth shattering scientific discoveries that significantly affect SQ, although selection of the wrong type of cable can do (ie significantly high in capacitance, or too high a loop resistance)

[YNWaN]

in relation to amplifier Zobel networks I doubt many amplifier manufacturers ( possibly Naim as an exception who then market their speaker cable having similar characteristics to be used in conjunction ) would release product on the marketplace without a zobel network.

Actually, there are quite a few amplifier manufacturers that don't include a Zobel network in their amplifiers. In the UK (as you say) Naim are the most well known, but NVA don't use one either and nor do Spectral.

[Barry]

The characteristic impedance of RF cables (such as RG58, RG174 etc.) is totally irrelevant at audio frequencies. The only parameters of interest are the loop-resistance per unit length and the capacitance and inductance per unit length. Compared to the values of the components in the crossover, the LCR values of the speaker cable are small (regardless of whether coaxial cable or 'figure of eight' geometry is used).

(If you like using RG58, try using RG223.)

[Reffc]

The characteristic impedance of RF cables (such as RG58, RG174 etc.) is totally irrelevant at audio frequencies. The only parameters of interest are the loop-resistance per unit length and the capacitance and inductance per unit length. Compared to the values of the components in the crossover, the LCR values of the speaker cable are small (regardless of whether coaxial cable or 'figure of eight' geometry is used).

(If you like using RG58, try using RG223.)

Exactly so. The loop resistance if too high can have a significant impact on damping factor (longer runs for example), hence the 5% rule of thumb (ie cables total DCR should be within 5% of the speakers minimum impedance).