Why some speaker cables sound different - in layman's terms
Shadders posed some interesting questions. I am sure he has an opinion on the original post?
/* Style Definitions */
table.MsoNormalTable
{mso-style-name:"Table Normal";
mso-tstyle-rowband-size:0;
mso-tstyle-colband-size:0;
mso-style-noshow:yes;
mso-style-parent:"";
mso-padding-alt:0cm 5.4pt 0cm 5.4pt;
mso-para-margin-top:0cm;
mso-para-margin-right:0cm;
mso-para-margin-bottom:10.0pt;
mso-para-margin-left:0cm;
mso-pagination:widow-orphan;
font-size:12.0pt;
font-family:"Times New Roman";
mso-ascii-font-family:Cambria;
mso-ascii-theme-font:minor-latin;
mso-hansi-font-family:Cambria;
mso-hansi-theme-font:minor-latin;
mso-ansi-language:EN-US;}
To clarify the situation I used to work for my brothers company DNM and the concepts I put forward are the result of his in depth analysis of the way cables interact with amplifiers – which I happen to agree with.
I am no longer involved with DNM and do not sell any brand of cable!
The root of this problem lies in the radio frequency (R.F.) range. DNM has taken a close interest in this since the late 1980’s.
An amplifier that uses negative feedback (and most of them do) is constantly comparing output with input. Any difference between the two signals is corrected by the feedback loop, so that distortion is reduced in the audio frequency range.
A weakness of this system is that it only works well at audio frequencies, because the feedback correction process is not infinitely fast. At the top of the audio range, amplifier distortion always rises sharply. At even higher frequencies, beyond the audio range, distortion becomes even higher running to many percent or even tens of percent. More complex amplifiers with ultra-low distortion have the greatest increase in distortion outside the audio range.
To visualise what is happening it helps to understand how fast the amplifier applies its feedback correction signal; what frequency does this process correspond to? All feedback amplifiers have a maximum rate-of-correction that can be associated with a specific frequency ( feedback control frequency) and there is some variation of this between amplifiers of different design. However the frequency range of interest is the medium-to-high radio frequency range.
Such high frequencies are not heard directly, the amplifier uses them to make its internal corrections and yet, in that range the amplifier has high distortion i.e. it is non-linear. The consequence of this is that the amplifier generates lower frequencies from its feedback action-- a form of demodulation.
The best way to describe this process (well understood in the radio world) is to imagine 100 sine waves at the extreme top end of the audio range being cancelled by the amplifier. The distorted feedback signal cannot cancel perfectly, most are cancelled but some are added to and boosted in amplitude. The new signal generated by this imperfect feedback process is always much lower in frequency than the original, so it becomes audible at the top of the audio frequency range.
Up to now I mentioned the amplifier's internal feedback process, so you may ask where does the audio cable enter the picture? The amplifier "reads" its output signal but the "reading" is unable to distinguish between the internal circuit and the external circuit, along the cable to the load. So when a cable is connected to the amplifier it compares its output (including the complex reflections in the cable) with its input. At medium-to-high radio frequencies the unterminated cable is not passive, but very active. At the feedback control frequency the cable's transmission line characteristics are highly visible to the amplifier, its sound changes as it responds to the signal
changed by the cable. Each different cable design produces its own unique change in the amplifier's sound and this process explains a large part of the "sound of cables".
DNM understood the details of this process while trying to minimise feedback errors in their amplifier designs. Research found that an amplifier could have a near-perfect feedback system, but when a cable was plugged in to connect the amplifier to its load (another amplifier or a loudspeaker) some of the feedback accuracy was destroyed.
The most effective solution to this problem was to accurately terminate the cable so that its R.F. reflections are not be seen by the amplifier's feedback system. DNM’s research resulted in a passive device called a high frequency termination network (HFTN) To understand this better you should visit the DNM website.
Shadders posed some interesting questions. I am sure he has an opinion on the original post?
To clarify the situation I used to work for my brothers company DNM and the concepts I put forward are the result of his in depth analysis of the way cables interact with amplifiers – which I happen to agree with.
I am no longer involved with DNM and do not sell any brand of cable!
The root of this problem lies in the radio frequency (R.F.) range. DNM has taken a close interest in this since the late 1980’s.
An amplifier that uses negative feedback (and most of them do) is constantly comparing output with input. Any difference between the two signals is corrected by the feedback loop, so that distortion is reduced in the audio frequency range.
A weakness of this system is that it only works well at audio frequencies, because the feedback correction process is not infinitely fast. At the top of the audio range, amplifier distortion always rises sharply. At even higher frequencies, beyond the audio range, distortion becomes even higher running to many percent or even tens of percent. More complex amplifiers with ultra-low distortion have the greatest increase in distortion outside the audio range.
To visualise what is happening it helps to understand how fast the amplifier applies its feedback correction signal; what frequency does this process correspond to? All feedback amplifiers have a maximum rate-of-correction that can be associated with a specific frequency ( feedback control frequency) and there is some variation of this between amplifiers of different design. However the frequency range of interest is the medium-to-high radio frequency range.
Such high frequencies are not heard directly, the amplifier uses them to make its internal corrections and yet, in that range the amplifier has high distortion i.e. it is non-linear. The consequence of this is that the amplifier generates lower frequencies from its feedback action-- a form of demodulation.
The best way to describe this process (well understood in the radio world) is to imagine 100 sine waves at the extreme top end of the audio range being cancelled by the amplifier. The distorted feedback signal cannot cancel perfectly, most are cancelled but some are added to and boosted in amplitude. The new signal generated by this imperfect feedback process is always much lower in frequency than the original, so it becomes audible at the top of the audio frequency range.
Up to now I mentioned the amplifier's internal feedback process, so you may ask where does the audio cable enter the picture? The amplifier "reads" its output signal but the "reading" is unable to distinguish between the internal circuit and the external circuit, along the cable to the load.So when a cable is connected to the amplifier it compares its output (including the complex reflections in the cable) with its input. At medium-to-high radio frequencies the unterminated cable is not passive, but very active. At the feedback control frequency the cable's transmission line characteristics are highly visible to the amplifier, its sound changes as it responds to the signal
changed by the cable. Each different cable design produces its own unique change in the amplifier's sound and this process explains a large part of the "sound of cables".
DNM understood the details of this process while trying to minimise feedback errors in their amplifier designs. Research found that an amplifier could have a near-perfect feedback system, but when a cable was plugged in to connect the amplifier to its load (another amplifier or a loudspeaker) some of the feedback accuracy was destroyed.
The most effective solution to this problem was to accurately terminate the cable so that its R.F. reflections are not be seen by the amplifier's feedback system. DNM’s research resulted in a passive device called a high frequency termination network (HFTN) To understand this better you should visit the DNM website.
Hi,
In response to this post – I examined the DNM web site and hence the Jim Lesurf article used as evidence to support the DNM product claims.
At medium-to-high radio frequencies the unterminated cable is not passive, but very active.
Cables are never active, they are always passive, whatever the frequency. You may be trying to state that a cable that has no load has an unusual frequency response ?
The distorted feedback signal cannot cancel perfectly, most are cancelled but some are added to and boosted in amplitude.
This is incorrect for an amplifier designed to be stable. As long as the phase and amplitude of the feedback signal is sufficient not to introduce positive feedback, then the amplifier will not amplify these signals.
I examined the reasoned statement of the post and completed some calculations.
The inference is that the cable connected to an amplifier will introduce RF into the system that affects the amplifier feedback. Each cable has a different LCR characteristic (we are ignoring conductance) and hence the RF fed back into the amplifier will be different, thus causing one cable to sound different to another.
Using a spectrum analyser set to 0Hz to 30MHz (as per article on DNM web site), and using a +3dB gain dipole – the horizontal polarisation RF noted as a peak stored, was -85dBm on average. Removing the gain of the dipole resulted in -88dBm RF across the 30MHz bandwidth.
In the J Lesurf article from DNM website – the speaker chosen has a 17ohm impedance at 100kHz, and rises to a peak 500ohms at some frequencies – hence the speaker impedance across this 30MHz bandwidth is taken to be 50ohms.
Cables in the article have a varying impedance from 5.5ohms to 337ohms maximum, where for the J Lesurf article this was 127.75ohms on average.
I have assumed a zobel network in the amplifier before and after the (output inductor in parallel with resistor) as is standard in most solid state amplifiers.
Hence for analysis we are using :
Cable impedance at RF 127.75ohms
Speaker impedance at RF 50ohms
RF power on average for horizontal polarisation in a 30MHz bandwidth is 50uW (speaker cables are lying flat on the floor – hence horizontal polarisation).
Assumed Resistors for Zobel are 4ohms, and series Resistor is 2ohms (since inductor is high impedance)
For the 50uW induced into the speaker cable – I assume this per positive terminal to speaker, and that this 50uW is wholly transferred into the cable, then the amount of power at the output of the amplifier internally where the output devices sum (usually feedback point taken from here) is 444nW (444 x 10^-9 Watts, or 444 nano Watts)
This 444nW for the 4ohm resistor introduces 1.33mV noise.
This 1.33mV noise is across 30MHz – we assume Additive White Gaussian Noise (AWGN).
As such – the RF noise per Hz is 44.4pV (44.4 picoVolts = 44.4 x 10^-12)
A 1kohm resistor at room temperature produces 4.1nV AWGN (4.1 nanoVolts = 4.1 x 10^-9)
As such, a single resistor produces nearly 100x greater noise than the injected RF power that is stated to cause errors in an amplifier feedback circuit.
Since the RF noise and Resistor noise is AWGN, the magnitude of AWGN of the MANY resistors in an amplifier is vastly greater than the injected RF. As such, the RF noise does not affect the amplifier as stated.
Furthermore, a cable may have different frequency response to the next – as per J Lesurf article Figure 9, but the cables difference as per the comparison plots show extremely small changes in behaviour – the same behaviour with a slight shift in frequency at the 8MHz onwards region.
As such, the negligible RF energy injected into the amplifier which does not affect the amplifier, may be slightly affected by the cable change, but this again is a negligible change applied to a negligible RF energy.
Of note, is the J Lesurf article in July 2010 Hifi News where he has provided evidence, that despite mains borne RF causing his amplifier protection circuit to indicate in error, that distortion in the amplifier was occurring, when there was none, no one had ever noticed the very high RF interference.
Regards,
Shadders.
Blimey! Shadders has got SKILLS! 
I think I'm going to go have a lie down...
Yes, excellent post, as usual 
- Login to post comments
So, after having read much in this thread and reading and watching videos that comments in this thread link too, I think it is safe to assume that cables are a very personal thing subject to the equipment already in use. A more expensive cable isn't necassarily going to get you a better sounding system - it depends on the system. Correct?
I'd say that's a pretty good summary of what has passed before. Mind you, this being a very touchy subject with some forum members, I can only speak for myself.