Why on earth do speaker wires sound different?

[Guest]

Given adequate sized conductors, why do speaker wires sound different?

If we use different wires from the ones used during design are we undoing the hard work of the speaker designer?

[Gordon]

Given adequate sized conductors, why do speaker wires sound different?

Good question but I don't have a ready answer or any rock solid objective evidence. One thing I would say is that, following on from the other thread where the interface between amp and speaker was done to death, we might think we have a thick enough wire but perhaps we don't? What is adequate?

A few random thoughts:

If we believe that dynamic range is high and that speaker impedances are low then transient peaks for example require lots of current. Given say a 250 watt amp driving a 4 ohm resistive [unlikely] load the max peak current is clearly 250/4 or 62.5 amps. A reactive load will need more current. A modest amplitude DR [defined as peak to mean here, not peak to quietest] of 40 dB would run at 100th this current on average. Even so, in proportion to the signal current required, the percentage loss is the same.

Even ring mains wire [2.5mm square TWE] is rated at 30 amps. 62.5 amps in 0.1 ohms is 6.25 volts - the 250 watt peak requires a peak voltage of 10 root 10 or about 31.6 volts so we have a 20% loss of peak voltage in the feed wire starving the speaker of what it should get and ruining the damping factor the while.

Remember also that the current loop comes back through the return conductor which also drops similar volts so the ground terminal of the speaker isn't really ground. What is the speaker seeing then? What happens to transients? Are they not a bit affected? Are they not compressed? And shouldn't this be audible to those with keen ears? A little loss of sparkle perhaps or of that last little amount of transient punch?

Moral? wire resistance has to be very very small, preferably a lot less than 0.01 ohms per run, which means very thick and as short a possible and it's worse the lower the nominal impedance is. That's an objective start.

1 metre of 2.5mm sq [ring mains TWE again] has a DC resistance of about 0.006 ohms per metre. So an average run of about 4 metres say has a DC resistance of 0.024 Ohms. A 62.5 amp peak will now drop about 1.74 volts in 31.6 or 5.5% so a bit of peak compression even here. At higher frequencies [ie during fast transients] this increases due to eg skin effect. The radius of the conductor here is 0.9 mm and the skin depth at 20 kHz is 0.5 mm which means the inner half radius isn't really conducting and so effective impedance increases significantly, approx 5 times. What chance transient precision then [even if the speaker can do it anyway]?

Discuss!!

If we use different wires from the ones used during design are we undoing the hard work of the designer?

What did the designer in the quote use to drive his baby through these wires anyway, surely we should all be using exactly what he did? How thick are wires inside the box?

[Dave2002]

Come on Gordon, you know it's because some wire has been stroked by 101 Virgins incanting Shakespeare sonnets in unison, and for some batches they got the words wrong.

[Gordon]

Come on Gordon, you know it's because some wire has been stroked by 101 Virgins incanting Shakespeare sonnets in unison, and for some batches they got the words wrong.

[clive heath]

FWIW I too use ring main as speaker wiring with the solid core as the -ve or return and the two outer sheathed L,N solids paired as the signal feed. The electrical centre of the two paired outer conductors is coincident with the solid core approx. I fondly imagine that this reduces inductance. Furthermore if you bi-amp and use the centre solid wire of, say, the HF signal as the return of the LF signal and vv then (capacitance being proportional to the inverse of distance) you've not got a lot of capacitance per metre as the wires are several cm apart. I am somewhat in awe of the expertise displayed on these topics so feel free to tell me this is nonsense! I just like the sound.

[MrGongGong]

How thick are wires inside the box?

Exactly ........
but what you really need are some of these

http://www.russandrews.com/product.asp?lookup=1&region=UK&currency=GBP&pf_id=3361&customer_id=PAA2484079813272WSXJEFRQDHZKWCRL

that's for ONE

I'm also using one of these on my kettle

http://www.russandrews.com/product.asp?lookup=1&region=UK&currency=GBP&pf_id=1559&customer_id=PAA2484079813272WSXJEFRQDHZKWCRL

and the tea tastes , well, just more highs , enhanced lows and a clean finish !

(but seriously , you do need a reasonably hefty amount of metal in a speaker lead BUT not all the deep frozen "special mains supply from the power station so that the mains isn't contaminated by copper as opposed to pure gold" woo nonsense

[Gordon]

FWIW I too use ring main as speaker wiring with the solid core as the -ve or return and the two outer sheathed L,N solids paired as the signal feed. The electrical centre of the two paired outer conductors is coincident with the solid core approx. I fondly imagine that this reduces inductance. Furthermore if you bi-amp and use the centre solid wire of, say, the HF signal as the return of the LF signal and vv then (capacitance being proportional to the inverse of distance) you've not got a lot of capacitance per metre as the wires are several cm apart. I am somewhat in awe of the expertise displayed on these topics so feel free to tell me this is nonsense! I just like the sound.

Agreed, broadly!! BUT beware that the E of the TWE is not as thick [it isn't normally required to carry mains current] and in your situation still has to carry the same current return as the L + N - if I understand your configuration properly. The main inductive element, ie a rising impedance with frequency, comes from the skin effect; if the feed and return paths are tightly coupled, as they are in this case, there should be very little inductance as the currents are balanced. Bi-Amping and Wiring is good to get lots more copper involved. Not sure what the crossover of the thinner wire returns in your BiWiring will do. Normally capacitance is not an issue in these circumstances of low impedance.

Anyway, glad you like the sound!

[Ferretfancy]

Back in the 1980s, when this subject first came into prominence, a group of BBC engineers approached several cable companies and asked if they would submit their products for double blind testing. They all refused.

[MrGongGong]

Back in the 1980s, when this subject first came into prominence, a group of BBC engineers approached several cable companies and asked if they would submit their products for double blind testing. They all refused.

I wonder why ?
Apart from the obvious need for hefty cable
the rest is pure homeopathic woo

[jayne lee wilson]

Exactly ........
but what you really need are some of these

http://www.russandrews.com/product.asp?lookup=1&region=UK&currency=GBP&pf_id=3361&customer_id=PAA2484079813272WSXJEFRQDHZKWCRL

that's for ONE

I'm also using one of these on my kettle

http://www.russandrews.com/product.asp?lookup=1&region=UK&currency=GBP&pf_id=1559&customer_id=PAA2484079813272WSXJEFRQDHZKWCRL

and the tea tastes , well, just more highs , enhanced lows and a clean finish !

(but seriously , you do need a reasonably hefty amount of metal in a speaker lead BUT not all the deep frozen "special mains supply from the power station so that the mains isn't contaminated by copper as opposed to pure gold" woo nonsense

Hmmm... no, never went THAT far. Better not say what I do use (they do look). All chosen via home trial though! Does that make it OK?

Kettles are very good for running mains cables in...

[Nevalti]

Good question but I don't have a ready answer or any rock solid objective evidence. One thing I would say is that, following on from the other thread where the interface between amp and speaker was done to death, we might think we have a thick enough wire but perhaps we don't? What is adequate?

A few random thoughts:

If we believe that dynamic range is high and that speaker impedances are low then transient peaks for example require lots of current. Given say a 250 watt amp driving a 4 ohm resistive [unlikely] load the max peak current is clearly 250/4 or 62.5 amps. A reactive load will need more current. A modest amplitude DR [defined as peak to mean here, not peak to quietest] of 40 dB would run at 100th this current on average. Even so, in proportion to the signal current required, the percentage loss is the same.

Even ring mains wire [2.5mm square TWE] is rated at 30 amps. 62.5 amps in 0.1 ohms is 6.25 volts - the 250 watt peak requires a peak voltage of 10 root 10 or about 31.6 volts so we have a 20% loss of peak voltage in the feed wire starving the speaker of what it should get and ruining the damping factor the while.

Remember also that the current loop comes back through the return conductor which also drops similar volts so the ground terminal of the speaker isn't really ground. What is the speaker seeing then? What happens to transients? Are they not a bit affected? Are they not compressed? And shouldn't this be audible to those with keen ears? A little loss of sparkle perhaps or of that last little amount of transient punch?

Moral? wire resistance has to be very very small, preferably a lot less than 0.01 ohms per run, which means very thick and as short a possible and it's worse the lower the nominal impedance is. That's an objective start.

1 metre of 2.5mm sq [ring mains TWE again] has a DC resistance of about 0.006 ohms per metre. So an average run of about 4 metres say has a DC resistance of 0.024 Ohms. A 62.5 amp peak will now drop about 1.74 volts in 31.6 or 5.5% so a bit of peak compression even here. At higher frequencies [ie during fast transients] this increases due to eg skin effect. The radius of the conductor here is 0.9 mm and the skin depth at 20 kHz is 0.5 mm which means the inner half radius isn't really conducting and so effective impedance increases significantly, approx 5 times. What chance transient precision then [even if the speaker can do it anyway]?

Discuss!!

What did the designer in the quote use to drive his baby through these wires anyway, surely we should all be using exactly what he did? How thick are wires inside the box?

Mmmm, thank you. Mostly I felt the whoosh as your comments went straight over my head. The calculation arriving at 62.5Amps was interesting though - mainly because I almost understood it. That is a hell of a high figure. Can mortal amplifiers provide current that high? Where do they get it from? Capacitors I presume?

I checked my amp's manual but it does not quote a figure for current, it did say this though, ".....when operated with both channels delivering maximum power into 4 ohm loads, will consume all the available power in a normal household circuit, therefore a dedicated electrical circuit may be necessary with this situation". 'Blimey' seems to be the appropriate word. I never guessed that my music could delay the making of my dinner. I must remember to turn the volume down if I'm hungry.

Wikipedia says:

A common rule of thumb is that the resistance of the speaker wire should not exceed 5% of the rated impedance of the system.
The table below shows recommended lengths based on this guideline:

Maximum wire lengths for two conductor copper wire[4]
Wire size...................2 Ω load...... 4 Ω load......6 Ω load......8 Ω load
22 AWG (0.326 mm2) 3 ft (0.9 m) 6 ft (1.8 m) 9 ft (2.7 m) 12 ft (3.6 m)
20 AWG (0.518 mm2) 5 ft (1.5 m) 10 ft (3 m) 15 ft (4.5 m) 20 ft (6 m)
18 AWG (0.823 mm2) 8 ft (2.4 m) 16 ft (4.9 m) 24 ft (7.3 m) 32 ft (9.7 m)
16 AWG (1.31 mm2) 12 ft (3.6 m) 24 ft (7.3 m) 36 ft (11 m) 48 ft (15 m)
14 AWG (2.08 mm2) 20 ft (6.1 m) 40 ft (12 m) 60 ft (18 m)* 80 ft (24 m)*
12 AWG (3.31 mm2) 30 ft (9.1 m) 60 ft (18 m)* 90 ft (27 m)* 120 ft (36 m)*
10 AWG (5.26 mm2) 50 ft (15 m) 100 ft (30 m)* 150 ft (46 m)* 200 ft (61 m)*

...........which, with my limited understanding, seems to tie in with what you said. That makes (schoolboy physics) sense to me but what about fancy insulation or twisting the wires together or silver plated wire or solid silver wire or keeping the two wires well apart or suspended or bi-wiring etc?

For over 20 years now, I have used very cheap Linn K20 cable (fairly thick multi-strand). My occasional auditions of other, fancy cables have either provided no audible benefit or they have been significantly worse than the cheap Linn cable. Some people spend a fortune on speaker cable, I was just trying to understand a little of the theory to see if I was missing out on a simple upgrade.

Linn K20
Cable 2-Core dumbell wire
Core 4 mm sq.
Strand 56 strands 0.3 mm copper
Insulation Grey PVC
Inductance 0.005 ohm per metre
Capacitance 16pf per metre at 1KH (typical value)
Impedance 0.93 mh per metre 1KH (typical value)

[Dave2002]

I checked my amp's manual but it does not quote a figure for current, it did say this though, ".....when operated with both channels delivering maximum power into 4 ohm loads, will consume all the available power in a normal household circuit, therefore a dedicated electrical circuit may be necessary with this situation". 'Blimey' seems to be the appropriate word. I never guessed that my music could delay the making of my dinner. I must remember to turn the volume down if I'm hungry.

There are people who are serious enough about this to install a separate main circuit for hi-fi kit!

There are also people who will pay a fortune (see Mr GG's earlier post) for mains cable, and say it makes a difference. I once tried to "discuss" this with someone who had that belief, and suggested that the mains wiring behind the socket from the consumer unit was very likely to dominate, so the cable from the socket to the equipment was probably an expensive waste of money.

Running continuously could present problems, but for classical music surely sufficient energy should be stored for a few seconds of serious high level work inside the amp.

[OldTechie]

Mmmm, thank you. Mostly I felt the whoosh as your comments went straight over my head. The calculation arriving at 62.5Amps was interesting though - mainly because I almost understood it. That is a hell of a high figure. Can mortal amplifiers provide current that high? Where do they get it from? Capacitors I presume?

Ooops - Gordon forgot a square root in his quick sum - it's actually 8 amps, 32 volts (rounded, rms) or roughly 12 amps/50volts peak. Bur as Gordon says when things are not in phase because of the usual reactive load the numbers can be higher.

The advantage of bi/tri-wiring is that if the crossover in the speaker has some horrid impedance variation with frequency, the resistance in the cable won't allow compromises in the design of the lf crossover to mess up the hf drive and vice versa. So really good cross-overs probably don't need them, but provide them because many of the customers will have been told it is essential.

There is a problem with skin-effect, where high frequency signals are inclined to travel in the outer of the conductor thus increasing the resistance. Multi-strand helps a bit.

It seems to me that if you can hear a problem introduced by a sensibly sized piece of cable such as you have, then you should move the amplifier closer to the speaker.

[cloughie]

Being sceptical I do think that cable manufacturers and stockists have made money from this hype.

[Phileas]

Before discussing why speaker cables sound different, it might be worthwhile first proving that they do sound different.