Theoretically, there would be a timing difference in when the left and right signals actually reach the speakers. Whether this would make an audible difference or not, I can not say as the signals are both travelling so fast that your ears may not be able to detect the slight delay in one channel.
Regardless, I would go with equal lengths on the cables for at least 3 reasons that immediately come to mind: 1) You MAY be able to hear the timing diference, 2) Some manufacturers claim that their cables don't "work" in very short lengths. I tried to buy some 6ft MIT 750 cables once from a store and was told that MIT would not sell them in lengths of less than 8 ft. 3) Resale. It may be the furthest thing from your mind right now, but trust me, you are going to want to sell/trade these someday and will get zero interest/value if they are of unequal lengths.
Hope this helps. |
As a general rule cable lengths short be as near identical as possible and as short as possible.
|
Yup, go with the same length. The simple answer is that the damping of the two cables is different if they are different lengths, also your high frequencies may roll off differently. (And it's really, really tough to sell two cables of different lengths! :-))
Greg |
If you're using a cable that's so resistive or (with speakers where it can matter) so reactive that an extra 12 feet causes an audible difference in sound, my advice would be to change cables. But if you are intentionally using your cables as a tone control, then yes they have to be the same length.
The timing difference really is too small to matter. Think about it: You can also create a timing difference by moving one speaker slightly closer to you than the other. Adding an extra 12 feet of wire to one channel would have the same timing impact as moving a speaker a few millimeters. (I'm estimating here, but I'm pretty sure that's the right order of magnitude.) Now, did you use a tape measure that accurate when you placed your speakers? And what happens when you move your head? If this mattered, audio would sound very weird. |
ELSman: Another point you may consider, if you have not already, is that wire to an ESL has special issues. An ESL acts like a capacitor (C) in the circuit (unlike conventional speakers)and has a transformer as part of the speaker which is an inductor(L). This can cause an L/C resonant circuit and may result in an unwanted peak at high frequencies. You do not want high inductance in your speaker wire to add to this problem. Inductance is largly determined by the area between the conductors (in the speaker wire). A few folk I know favor coaxial design.
I am not a big ESL person but there are several at this site, Detlof, for one, comes to mind. Maybe they can offer better suggestions if you are not already familiar with the issue.
Anyway, this may be as important as the length issue.
Cheers
I remain, |
Since the signal travels at the speed of light, there shouldn't be a timing difference but it's still a good idea to keep the lengths the same. |
Great response and thank you for the knowledge. While we are on this topic what cables would compliment electrostatic speakers. Clueless and some of you have pointed out electrostatic speakers may require special considerations. |
Actually the propagation speed for the average cable is probably about two-thirds the speed of light.
Velocity (m/s) = Sqare root (1 / (Li*Ci)). Li = linear inductance of cable (H/m) and Ci = linear capacitance of cable (F/m). |
There are no absolutely newtral cables. The active resistance plays the smallest role especially in the wires of 10AWG or thicker. The reactance will play more important role(especially inductance). If your amp is high current and can take high capacity loads than you will not experience any audiable differences. |
I am using a ML No. 336 and Martin Logan Prodigy electrostatic speakers. |
For ML 336 it's no problem at all. |
Inductance, iirc, is related (for a "normal" wire) to the diameter of the conductor. The capacitance of the cable is related to the distance between conductors.
By normal wire, I mean one that is not a spiral like the old Krell/Symo stuff for example...
Some mfrs do things to add capacitance (shunt capacitance) via the relationship between the wires so as to offset the self inductance of the cable. Dunlavy's patents might make for good reading on this sort of idea.
Personally, I'm against this sort of idea since it rarely seems to yield a cable that sounds good to my ears. What seems to happen is that the capacitance wrecks havoc with the output waveform of feedback amps. For non-feedback amps (the few out there) the effect is different. So, IMHO, these sorts of cables never sound alike on two different amps or speakers. Unpredictable.
I try to minimize the inherent inductance and capacitance effects and feel that results in the most predictable and stable cables.
And again, the internal geometry of any cable will play the major role in its sound, the wire, materials, and the like being more secondary or even less important. That's not to say that you can not hear any of these effects when comparing identically built cables with one item changed, you often can.
So, the idea is that if you use two different lengths, usually the basic L & C values will be different enough to cause some differential in what you may hear. So, go for two
identical lengths, and DON'T COIL the excess cable!
(it makes an inductor) :- )
_-_-bear |
By coiling the excess cable are you conserned about magnetic field generations affecting the sound quality? In addition my problem will be the system is in our living room and my wife will kill me if any more cable is exposed. I already have my receiver antenna sticking out of her plastic tree. |
NO, not magnetic field generation! By coiling the cable you are making, well, a coil! This increases the series inductance of the cable, rolling off the highs.
What you do with your wife and your antenna, is another matter for a different group, I think! ; -)
_-_-bear |
I use two different lengths in my room and actually they are two entirely different cables and there is little to no difference at high volumes. At very low volumes one speaker is a little louder than the other. |
that should be a good closing statement for this thread |
By coiling you create an extra inductance which is so small that it can be just neglected for audio freequencies.
On the other hand the difference in volume level tells that loads to an amplifier are different. To decrease this difference you should use thick 10AWG or even thicker speaker wires/cables. |
|
Now I remember too, Bwihite about that article.
Moreover, It's almost impossible to find a HT system that uses equal speaker wire runs for mains, sub, center and surrounds(ABSURD!) |
In a situation similar to yours, e-man, I hedged my bets by making the shorter wire about two-thirds the length of the longer one. That's quite a bit longer than I needed, but I didn't want to have the difference between the two be too large. (It also leaves me with a length that's not too short to sell, if I wanted to.) I loosely gathered the excess into a sort of figure-eight shaped bow. Someone smarter than me said this would be better than a coil-- maybe Bear can comment. I don't know if I've come out ahead in terms of sound quality or not, but over extended listening I have not noticed any imbalance or deleterious effect. If I chose to approach this like a hobbyist and experimented and investigated, I might or might not discover some subtle differences, but it's easy for me to leave this alone. |
That inductance is measurable at audio frequencies.
If you look at the output of a typical solid state amplifier with an inductor in the output you can easily measure the change in the way a square wave is reproduced with and without the inductor when driving a capacitive load. Most speakers have a capacitive load angle somewhere.
So, it is highly probable that increasing the series inductance that the amp sees WILL have a measureable and probably audible effect within the "audio range."
Published tests have shown that this small series inductance found in cables will cause a small but measurable drop by 20kHz. This is different than the square wave effect I mentioned above, but does show a measured effect within the "audio range."
In general it is a good idea to minimize these effects in a system that you are trying to optimize. If you are not trying to put together such a system, it won't matter at all.
_-_-bear |
I did a quick experiment to see what the effects of coiling up 8' of a 13' speaker wire up would be. I took a 13' speaker cable (12 gauge oxygen free copper) and measured the effective resistance over 20 Hz - 300 kHz. I then coiled up 8' of the cable which made the effective length of the cable to be 5'. The coil was about 4 loops in a 6" diameter. I then measured the effective resistance again. Each line of the data is in this format:
Frequency,
Resistance of straight wire (milli-ohms),
Resistance of wound wire (milli-ohms),
% increase of resistance:
20 36.38 36.65 0.74
30 36.72 36.92 0.54
40 36.7 36.93 0.62
50 36.72 36.99 0.73
80 36.79 37.23 1.18
100 36.87 37.43 1.50
120 36.94 37.64 1.86
150 37.06 37.96 2.37
200 37.72 38.5 2.03
250 37.45 39.03 4.05
300 37.63 39.55 4.85
400 37.92 40.48 6.32
500 38.2 41.34 7.60
600 38.46 42.17 8.80
800 38.95 43.7 10.87
1000 39.41 45.14 12.69
2000 41.06 51 19.49
4000 43.66 59.9 27.11
5000 44.84 63.77 29.68
6000 45.81 67.28 31.91
8000 47.66 73.9 35.51
10000 49.47 80.14 38.27
12000 51.32 86.16 40.44
15000 54 94.62 42.93
20000 58.36 107.76 45.84
25000 62.5 119.8 47.83
30000 66.55 131.05 49.22
40000 74.25 151.6 51.02
50000 81.9 171 52.11
60000 89 188 52.66
75000 99 213 53.52
100000 118 251 52.99
300000 255 505 49.50 |
I should mention that the data I posted is for one side of the speaker cable only. The data for the other side is similar. Therefore, the numbers can be doubled for a quick approximation. For instance, at 10 kHz, instead of a 40.44% increase in resistance, you will actually have an 80.88 percent increase, etc.
When I did my experiment, the coiled wire had a higher inductance and capacitance. For instance, at 10 kHz, the straight cable had 3.96 uH inductance and 227 pF capacitance while the coiled cable had 5.26 uH inductance and 249 pF capacitance. Instead of listing all of these values, I have only provided the "effective resistance." I would be happy to email a spreadsheet with all of the data to anyone who wants it. |
Wow, nice Bufus, I never would have thought it made that much of a difference (furtively slinking away to uncoil his cables). :-)
Greg |
Okay... so what is worse, coiling one speaker cable or having one side shorter? |
Before we move on to deciding which is worse, coiling one side or having one side shorter, I want to mention one more thing about the data. I made a mistake with my percentage calculations. For instance, notice at 10 kHz, the "effective resistance" increased from 49.47m to 80.14m. That is actually about a 62% increase instead of the 38% that I listed. At 300 kHz the jump from 255m to 505m is about a 98% increase !!! This series of measurements is interesting because it points out that side-by-side speaker wire is generally not that great. It is not as linear over the listening range as we would like. All of our other equipment costs alot of money and is linear. So why would we think that spending more money for speaker cables is a waste? How much more to spend is the question. Hey, if anybody has cables they want measured, send them to me. It will give us some data points for comparison. |
Thank you all: I am convienced, the purchase of even length speaker cable is well worth the money when optimizing our systems. A special thanks to Bufus for takeing what we suspect and coming up with just facts. |
Well, Bufus is not telling you the whole story.
What he did say is essentially correct though!
The fact is that the resitance of the cable is in the milliohms, and that a double (100%) increase in resistance is really rather small. So, it's not the resistance that you have to worry about too much.
IF there is anything to worry about it is the reactance and the effect that it will have on how your amp sees the load. In some cases (highly capacitive speaker cables come to mind) I'd actually want some series inductance with certain loads.
But, for optimizing a system, I'd have to say keep both sides the same as much as is practical.
And, coiling *one side* of a speaker cable is likely to not be as good as not coiling one side...
In case you are unsure about what the effect of "out of band" reactance is on a typical amp, try measuring a square wave with a capacitive load on it! It usually doesn't look very pretty - and doesn't usually sound so pretty either.
_-_-bear |
