Why aren't speaker cables Coaxial?

Stupid_Ben

Just out of interest I was wondering and couldn't work out...

Why are (even high end) speaker cables not coaxial? We always use coaxial cable in the lab to get rid of noise on our measurements.

I figure that 50/75 ohm standard coax would create an impedance mismatch with the typically low ohmic speakers, and output of the audio amplifier. However I calculate that with an outer diameter to inner diameter radius of 1.1, the impedance would be 2.5 Ohms as per below


[itex]Z_0 =\frac{138\Omega}{\epsilon_r}log_{10}(1.1)=2.5 \Omega[/itex]


[itex] \epsilon_r=2.33[/itex] which I think is typical.


This ratio seems feasible to me, to be made into a thin flexible wire.

Come to think of it my guitar cable is coax, and so are some headphones cables.


look forward to hearing your comments on this.

Ben

Studiot

Good afternoon Ben.

Noise is only a problem if its level is comparable to that of the signal. This is measured by the parameter 'signal to noise ratio'.

Since the noise in a given situation is pretty fixed thae only variable is the signal level.

Thus noise is only significant for low level signals like microphones and pickups.

For signals at loudpeaker levels the nosie necomes an insignificant proportion of the total signal.



go well

Stupid_Ben

Thanks Studiot I hadn't thought about that. I guess speakers need a fair amount of signal to get them to move!

Hence the excuse behind the enormousness speaker wires audio buffs spend vast sums of money on.

yungman

Also coax has shunt capacitance also. As small as it is, still in the range of 50 to 100 pF. Not to mention it would be very stiff.

The important thing about speaker cable is the size of the conductor. I am into audiophile, I actually experiment with speaker cables. I ended up using 4 pairs of 12 gauge Monster type of cable for each speaker. I add one pair at a time and amazingly it make a difference in the sound.

Small little capacitance at audio freq don't seems to be important, so are 2 pair or 4 pair of those big cables!!! I ended up spending close to $1000 on all the inter connects in my stereo system and they are not of top quality. Human ears are funny!!! It is not about the highs and the lows of the sound, it open up the perceived sound stage. It make the sound more three dimensional and more airy.

f95toli

The answer is actually a bit subtler than that. Which type of cable is best depends on the source and load impedance. For situations where you are primarily interested in the "current" (low impedances) one should -in general- use twisted pairs; whereas in high(ish) impedance ("voltage") situations (and at high frequencies) coax is better.
Note that microphone cables are also (usually) twisted pairs.


Note that this does not only apply to speakers (or audio). In the lab one should use a mixture of twisted pairs (preferably shielded) and coaxial cables; depending on the application (hence, if you are ONLY using coaxial cables in your lab, you are either working with high impedances or you are doing something wrong).

There are also all sorts of ifs and buts to this; which is one reason why there are also other types of cables (notably triax).

Stupid_Ben

Hi yungman,

Thanks for the interesting comment, I can't say I every actually tried playing with big cables but it is interesting to know that you find it makes a difference. I think that the highs and low frequencies make the harmonics and resonances which makes a sounds "open up " . Though I imagine that it less down to transmission of specific frequencies, more to do with a temporal response perhaps.... BTW just out of interest did you try a blind test? the 'placebo effect' can be very strong

Stupid_Ben

Hi f95toli,

I think I agree with you on most things. However one 'but' that I think might be worth mentioning . For low frequency, the characteristic impedance of coax makes no difference, but it's shielding is much better than twisted pair. That's why we like to use it, also it radiates much less than twisted pair at high frequencies.


If I'm wrong please tell me !

f95toli

No, its not. It really depends on what kind of interference you are trying to shield from (inductive/capacitive pickup, RF interference etc). Also, for low-level signals one should of course use shielded twisted pairs (grounded at one or both ends, depending on the situation). Coax would be pretty useless if one was for example to try to do a simple 4-point resistance measurement of a low-ohmic load (a pretty common thing to do).

See e.g. Ott's "Noise Reduction Techniques in Electronic Systems" or a similar textbook.

But, yes. For high frequencies one would generally speaking use coax (but not always, network cables are twisted pairs).

Stupid_Ben

I don't remember ever having seen shielded twisted pair...

Anyway I'll have a look at the book you recommend next time I'm at work.

Studiot

50 ohms or less coax is high impedance?

Whereas 'figure of eight' or ribbon is 300 ohms.

f95toli

Shielded twisted pairs are very common. Standard LAN (network) cable is a good example; shielded CAT5e is very common (contains 4 pairs in a foil shield)

skeptic2

I have always been curious about this and I wonder if you've ever done any blind testing. What kinds of sounds is it most noticeable with? How much power are you delivering to the speakers? Which characteristics of the monster cables do you think are responsible for the difference in sound, less resistance, less inductance or what? Don't the wires inside the speaker cabinet from the terminals to the speaker itself have enough resistance and inductance to nullify the effect of monster cables?

As you refer to cable pairs, I assume you have a stereo system rather than a 4 channel system. Since stereo and our ears are at best a two dimensional system, how is the third dimension perceived?

f95toli

I was refering to the impedance of the source/load; the cable impedance is more or less irrelevant unless you are measuring at high freqiencies (although the cable capacitance/inductance can matter, so the impedance has an indirect effect).
And yes, 50 ohm is a high(ish) impedance in many situations.

Studiot

No of course the cable impedance is not irrelevant.

For the record.

The standards for audio and telephone work are set at 600 ohms

Most coaxial laboratory instrumentation uses 50 ohms, as does computer coaxial connections.

Radio (including TV) uses 75 or 150 or 300 ohms.

In all cases it is important to ensure the cable and loads are matched to avoid reflections.

f95toli

But reflections are only an issue if your measuring at a frequncy where the wavelength is of the same order of magnitude (say lambda/4) as the length of the cables. In the lab this is usually only an issue RF/MW measurements (over say 10 MHz). It is complettely irrelevant if you are for example measuring a reistance with a multimeter.
Low-noise measurements are much, much trickier near DC than at high frequencies.

yungman

I was doing the test together with my wife. It was quite obvious. Believe me, I was a absolute non believer before. The reason I even tried this was because I bought a pair of JM Lab to replace the pair of Kef and I did not hear any significant improvement and I was getting desperate and looked into it. The better the quality the system, the more critical it is. Yes we went back a fore to verify.

I was absolutely surprised.

AlephZero

If anybody can give a rational scientific explanation as to what is so good about speaker cable that costs $1000 per meter, and which must for best performance be connected "the right way round" (it is marked to show which end to connect to the amp and which end to the speakers), go right ahead.

Personally, I find flat 30-amp mains power cable works just fine. I don't really need more power than 1800 watts per speaker in my house, so I'm not going to exceed the rating of the cable!

Cheap and nasty connectors with high and/or inconsistent internal resistance are a differnet matter. Fractions of an ohm are significant for high power low impedance wiring like loudspeakers. But copper wire is just copper wire, whether it comes from your local electrical store or insulated with snake oil and costing 10,000 times as much, IMHO.