Platenspeler.Com |
 |
Platenspeler.Com |
|
There is no subject so subjective as discussions on cabling. Ask ten people what they heard in a hifi setup when changing interconnects and you'll be surprised. But what if these people know what interconnects they've been listening to: Most listeners will probably tell you that the interlink with the highest status and price sounded best.
I therefore decided to have a look at the technical side of interconnect cables in order to avoid errors in my own setup.
To avoid any confusion or misunderstanding: I will only focus on the electrical characteristics of audio interconnects; The "subjective" judgements about how interconnects sound or what materials are better is left up to the reader.
My recommendation (2 Cents): Try to be critical to yourself, if you don't hear it, it probably isn't there...
The general model for a cable is mostly something like:
Do not pay attention to the values in this picture, it's the components and
their relation that matters. A cable can be considered a small network: It has
a serial resistance, a parallel capacitance between the conductors and it acts
as a serial inductance (even if we do not roll-up our interlinks or speaker
cables).
A cable acts as a low-pass filter that is mainly active in the HF range. Depending
on the intended use and length required we should either select low-capacitance
or low-resistance cable (or both).
There is at least three types of cable to be discussed:
I will discuss the first three on this page, especially since the last one
requires different kind of math and background to be judged.
Interlinks transport audio signals between a source (e.g. CDplayer or pre-amplifier) and a power amplifier. The figure below describes the generalised schematics for interconnect cable setup.
For audiointerconnect cables, it is important that capacitance is kept as low
as possible. The signal source mostly has an output impedance below 1kOhm, and
often even less when the device is designed to connect to an amplifier over
some length. Amplifiers have an input impedance of 20kOhm or better (50k may
be typical). Therefore, the serial resistance of audio interconnects is less
importance than other cable parameters, manufacturers such as vdHul therefore
are successful with carbon based cables.
Inductance is also not of a large influence in interconnects, and simplicity's sake therefore one can leave it out of the formula's (simulators prove us right). The formula for calculating the -3 dB point is as follows:
F(-3dB)= ( R_out + R_in )/( 2 * pi * ( C_cable + C_in ) * R_out * R_in )
For longer lengths however, low capacitance and a lower resistance is advicable. Below a small table with parameters for popular types of interconnect cable.
| Cable | resistance/m | Capacity/m | Inductance/m |
| vdHul First Ultimate | 36 (!!!) | 61 pF | |
| vdHul D-102mk3 | 0.055 | 37 pF | ??? |
| 5 meter cheap interconnect | 0.055 ?? | 160 pF | |
Let's simulate one or two of these interconnects with a SPICE simulator using the following parameters for the pre-amp and the power amp: 1kOhm ouput impedance and resp. 50 kOhm input impedance and 30pF input capacitance.
I want to point out regarding this sheet that the values for the interlink itself
are quite OK, it's my setup with a tube pre-amp with just 1kOhm output impedance
that is killing for high-frequencies. 100 Ohms would be much better as is shown
in the chart too.
For example, the pre-out of my Marantz PM14mkiiKI is just 250 Ohms at 1.8 Volts. The "main in" input on the other hand is 100 kOhm. With these values, it is relatively easy to find a nice long interconnect.
What the model shows clearly as well is that the frequency response is not dependent on inductance parameters for line level interconnects. Changes in this parameter do not influence the frequency response too much. Capacity is the most important parameter. And for longer interconnect cables, the relation between output impedance of the source and input impedance of the power amp are very important.
Speaker cable is an essential element in the audio chain. At the output of the amplifier the amplified signal is found mostly with very low distortion figures. At this output one (or more) speakers are connected that each have a typical (average) serial resistance of 4 to 8 Ohms but in practise this load varies between 3 and 10 Ohms over the full frequency range of the speaker. Power amplifiers will be able to offer the output voltage levels more or less independent from the connected load thus compensating with lots of current (and a low internal output resistance).
Because of this, the serial resistance of speaker cable plays an important role but next to this inductance needs to be kept low.
Capacitance is less of a problem since the output impedance is very low, but inductance needs to be kept to a minimum for this type of application.
For a lot of people, power cables are just power cables and nothing more. After all, there is kilometers of wire between the power plant and our house, and in our walls there is just a few meters inferior copper cable of 2.5 square mm to our hifi set. So what's the use of hooking up your hifi with one meter of exotic/expensive power cable? And how would a different power cable help to improve your sound? Why not use a power filter if there's so much pollution on the net? I do agree on most of these comments: After all most power supplies will know how to deal with HF pollution. And if things were so easy, manufacturers of power amps would have bundled a better power cable for their high-end products (not?).
On the other side: There's many people that tell me that better power cables improve the sound in their setup. And if it works for them then why bother.
According to some people including manufacturers, power filters are not the best solution for power amplifiers and a better power cable with "built-in filter" would be the better solution. After all, there is no problem in power cables behaving as a low-pass filter, such in contrary with NF interconnects where loss of high frequencies is not desirable. Maybe the explanation could be that some power cables have a higher capacitance and inductance because of twisting and braiding of the wires. This may help to kill HF noise before it reaches your Hifi components.
Initial measurements did not provide me with much information, the NET-3 cable I made has a capacitance of approx 0.5 nF, the Conradsel about 1 nF. A regular heavy computer power cord also had a value of approx. 0.5nF so that did not bring me any further.
The figure above shows the schematics for a power cable, with I think realistic
values for capacity and inductance. OK, so what would desirable parameters be?
Well let's put this model in a simulator.
First of all, I did a capacity sweep simulation. All values stayed the same
as in the schema above but the capacity of the cable varied between 1nF and
1000nF. As shown in the chart above, the capacity of the power cable is an important
factor, but you need a huge capacity in the powercord in order to have some
effect. And even then, maybe a parallel filter in the wall socket would do the
same trick (and it often does).
Secondly, an inductance sweep simulation showed how this parameter has a huge effect even for small values of the "coil".
Serial resistance must be as low as possible, parallel capacity may be higher since this will help in getting HF noise shortened out in the cable and serial inductance must not be too high but in the order of 0.5 mH or lower. In my case therefore I do not worry about the parallel capacity since the netfilter I use for the CDplayer and DVD-player uses a capacitor of 220nF between phase and 0, and the power cable of my amp will "see" this parallel capacitance as well. (Well don't expect to turn your 1.5 meters mains cable into a filter capacitor anyway).
So what's left is an increase in serial inductance that could help in getting HF noise reduced. A value of 10 uH would do just fine I guess, especially for power amplifiers. Hhmmm, unfortunately my multimeter does not support measuring inductance (this needs fixing so close to the Xmas season right).
© Maarten@platenspeler.com, May 2002
<>