Speaker wire length uneven

I just purchased Belden 10 gauge for my left, center and right. Do I need to match wire length for the left and right. Or better to make the shortest possible runs. The runs would be approximately 8, 12 and 18 feet.

I also purchased Soundking 12 gauge for the four surrounds. The minimum runs would be 35, 45, 55 and 70 feet. Do I need to make length per pair? Or better to make the shortest runs?

Will 70 feet be a problem? I assume 12 gauge is plenty.

I see "in-wall" speaker wire. Is there something special regarding the insulation and fire safety for this type. I will be pulling the wire to keep it in the walls.

i wouldnt worry too much about different lenght runs for your speakers.. 12ga should be fine for thelonger runs as im only using 16ga and the longest run i have is around 40 feet... if you notice any different volumes from right left speakers (unlikely for fronts bur rears MAYBE.. just a maybe) you should be able to set your system to compensate for that.. if you do notice any differences.. only 1db from shortest to longest is about all the compensation you would need... my rears vary from 25 to 40.. and i have them set dead on even each side.. and each side seems even to me.. im ure a meter or other sophisticated equipment would give me a difference.. but my ears dont.. so im good hehe

Na, it's not a real big deal. The length is fine.

In general I think it's a good idea to keep all your wire length the same. If you're using a surround sound system the signal from the reciever is sent to the speakers at the same rate. If the wire to your left channel is longer than to your right the signal will arrive their first. A true anal audiofile will tell you to run the same length to all channels.

Quote:

---

Originally Posted by Urloony

If the wire to your left channel is longer than to your right the signal will arrive their first.

---

Knowing the speed at which the signal travels through the wire is pretty close to the speed of light, could you please show me the math you used to arrive at this conclusion, particularly when dealing with the lengths of wire in question?

Here's a hint: The signal moves faster through the wire than it does in the air from the speaker to the listener's ear.

Wow I actually did the math earlier today, thinking about digital microphones.

Speed of light is EXACTLY c = 299,792,458 m/s in a vacuum. Copper signal propagation is a bit slower, but lets just use the common number.

I have arbitrarily decided a 15 degree phase difference for 100khz should solve any and all objections as to what is outside of audibility. It flirts with the current SACD spec.

[15(degrees) * c] / [100,000 * 360(degrees)] = 124.5meters

Stating this in english, it takes around 400 feet of copper cable before the timing difference of propagation delay creates a lag of 15 degrees to a 100,000kHz sine wave. A good cable geometry for such a cable would be a wide, flat cable, with horizontally spaced multiple conductors for send and return paths, in the case of a 2 conductor bi-directional current handling cable, like that used for speakers. This geometry simultaneously minimizes capacitance, inductance, and if scaled correctly, series impedance due to sheer bulk material available.

I'm giggling waiting for confirmation/detraction...

Unless you know how to change the laws of physics the cable length to speaker pairs should be as close to the same as possible. It may not make a huge difference but if you have a good system the timing ques may be slightly different.

Hey Nighttrain, maybe my post above made the eyes glaze over. Yes, there IS A DIFFERENCE. But the human brain does not have infinite resolution. It is why I suggested choosing a point of reference outside of our likely resolution. There are some claims that Ultrasonic frequencies up to 100kHz are somehow audible to humans. Starting there, I suggested we pretend that 15degrees of phase shift would be a subtle shift at such an extreme end of the range. 15degrees at 100kHz equates with 3degrees at 20kHz. Do you want to talk about being able to hear that? Many people don't even hear that tone as a tone anymore.

The reason I was "doing the math" above is because someone requested it. It was very smart of them to want to know the actual effect.

Human intuition is a wonderful thing. It lets us think "outside of the box". Intuitively, if I make you walk twice as far to one store, while I walk to the one just around the corner, I will get home from shopping quicker than you. It seems if you wire the left channel with a wire 6ft long and the right channel with a wire 12ft long, that there will be a "huge difference" like in the shopping trips. But its more like Albertson's and Safeway are at distances one mile away, but one only 1 inch more than the other. It won't affect our shopping times in a way we'd be able to notice.

I qualified my post with the importance of avoiding inductance over a significant length of wire. ALL CURRENT CARRYING CONDUCTORS HAVE SOME INDUCTANCE. A magnetic field is created and destroyed by the AC current (in this case the complex one of music). This ramping up and down of the field takes some TIME, and that lag is partially phase shift, and partial loss of initial current to the field strength. If you take each cycle on its own, the slope of the field strength rise over time intersecting the cyclic rate of the input AC waveform, results in a net impedance and consequent reduction in transmission of current.

There are better shapes than others for avoiding inductance. Round is worst. Flat is best. Practical limitations say you can't have an infinitely wide foil of one atom thickness, but wider is better, assuming enough cross section. A GREAT cable might be one that uses a copper type braid as seen in a "shield", with a foam core hollow center. Many speaker wires follow this geometry, some with Litz insulation.

I will close with a reminder that the path length acoustically to the ears should be as identical as possible within the inevitable "sweet spot" that stereo reproduction entails. The speed of sound in air at room temperature is around 86,000 times SLOWER than the speed of light in a vacuum. That is a law of physics. So a change of wire length of 6ft, is like a change of speaker position by the width of a human hair. You move your head around far more than that much WHEN YOU BREATHE...

Quote:

---

Originally Posted by Toga

ALL CURRENT CARRYING CONDUCTORS HAVE SOME INDUCTANCE.

---

Yes

Quote:

---

Originally Posted by Toga

There are better shapes than others for avoiding inductance. Round is worst. Flat is best.

---

More important is the geometry of the pair.

Quote:

---

Originally Posted by Toga

A GREAT cable might be one that uses a copper type braid as seen in a "shield", with a foam core hollow center.

---

No. More significant is the proximity of the two conductors to each other. Terman's equation will still apply, but the delta mu component will be zero if the conductors are perfectly cylindrical.


Inductance is the relationship between the current in the conductor, and the total energy stored in the magnetic field around the conductor.

Coaxial cable constrains that inductance, limiting magnetic storage to the region between the inner wire and the outer braid.

A round wire has a default dc inductance of 15 nH per foot. A tube of conductor has none of this, and broadcasts a field only outside the cylinder.

Flat is better if the return conductor is very close. If they are far apart, the conductor will skin heavily, sending current to the outer edges to the exclusion of the middle.

Prop speed is V_prop=1/sqr(L*C)

Cheers, John

Hi John!

Please review frequency of effect for skinning, and you will see it is well outside the audio band. Heck, run some Bode plots of spectral energy at your speaker input leads, and see where you are in terms of HF droop. I don't know how this became the faddy buzzword that it is, but there is no way that LAN performance in the GIGAHERTZ would be possible if skin effect was as horrible a bugaboo as claimed. As for distribution of magnetic field lines, a flat conductor creates maximum separation in the density of field lines, leading to the lowest "L" (this is why plain old circuit boards are now running 800mHz+ FSB on PCs). You could make a big loop with your return wire and this would still be true. Check the proportions for coil and transformer windings for ideal section, and the implied worst case scenario for synergistic field density.

Btw, a hollow cylinder is a silly shape to pursue. If skin effect is really difficult to avoid anyway, and yet back EMF from woofers (being the most massive driver in a multi-way system) occurs at near-DC frequencies, USE THE FREE COPPER in a solid cross-section cylinder. I'm starting to think all these high-impedance wire buffs don't like taught articulate bass, and don't like the effect of strong back EMF suppression of mid drivers. They favor wires that add 0.5 Ohm to over an Ohm to drag their damping factor down into SET range. Thus their speakers ring like undamped bells, and they whisper words like “bloom” and “warmth” with their faces all alight in wonder, when all they did was turn every decent dynamic speaker into the equivalent of a marimba.

You might as well get a 20Watt 1Ohm resistor and hang it off your SS amplifier output in series and pretend its a tube amp. That is what Bob Carver did...

Whatever… :P

Quote:

---

Originally Posted by Toga

Hi John!

Please review frequency of effect for skinning, and you will see it is well outside the audio band.

---

You referred to flat as "better", and how it doesn't skin..

You are referring to the simplistic model of skinning we were taught in school..that of a TEM wave propagating normal to the surface of an infinitely wide copper surface..with the corresponding depth of penetration exponential, and prop velocity with phase shift..That model is inaccurate for round wires, and for conductors that are not infinite...oh, and also, for magnetic and electric fields that are generated by the movement of charged particles within the wires...current..

For any arbitrary conductor, the reaction of that conductor to a change in current, is to create time varying magnetic fields within, and that in turn generates eddy currents....those eddy currents, when added to the impetus current, changes the overall current density profile, giving us the "skin effect", or more aptly called, current density profile re-distribution. It requires solving the Bessels for the system, as the TEM model is very, very inaccurate for audio frequencies and wires typically used for speakers.

For a flat conductor, the current density truncates at the edge of the conductor..this results in a magnetic field profile that is not the same across the entire flat conductor..because it is not uniform across the conductor, when the current slews, eddy currents will cause the current to crowd the edges, leaving the center starved. This effect will be noticed moreso for ribbon inductors, as there is significantly more field enhancement at the center of the ribbon. Proximity of the return ribbon will reduce this effect, the consequence being capacitive increase.

Skin effect, by definition, is the reduction of the internal inductance of a wire, due to NET current leaving the center, and shifting towards the outer surface of the conductor..all solid conductors, regardless of geometry, have 15 nHenries of internal inductance..A tubular one, however, approaches zero by geometry.

If you find a text that will further extend my understanding of skin effect, please let me know.. However, I must warn you, if they use the classic explanation of isotropic material field propagation, they are incorrect..

Quote:

---

Originally Posted by Toga

I don't know how this became the faddy buzzword that it is,

---

I do..try "high end audio guru".

Quote:

---

Originally Posted by Toga

but there is no way that LAN performance in the GIGAHERTZ would be possible if skin effect was as horrible a bugaboo as claimed.

---

Skinning is real, this is why waveguides were polished gold plated during WW2 research into radar, and is still a practice today, with mm and micrometer wavelengths.

Quote:

---

Originally Posted by Toga

As for distribution of magnetic field lines, a flat conductor creates maximum separation in the density of field lines, leading to the lowest "L" (this is why plain old circuit boards are now running 800mHz+ FSB on PCs).

---

High speed on PC boards is not possible without considering the traces as transmission lines..this is done with ground planes..I believe this is also why HDD cables now have 80 wires..for keeping impedance matched. I ran into this problem back in '83, when I was working ECL circuitry in military hybrids.

Quote:

---

Originally Posted by Toga

Btw, a hollow cylinder is a silly shape to pursue. If skin effect is really difficult to avoid anyway, and yet back EMF from woofers (being the most massive driver in a multi-way system) occurs at near-DC frequencies, USE THE FREE COPPER in a solid cross-section cylinder.

---

Actually, the hollow shape will afford the lowest inductance possible..understanding this, it is possible to create any wire, with any reasonable impedance, while keeping stored energy at a minimum. And, since the effect is well understood, it is possible to design and make a wire having 4 or 8 ohms impedance, with effective guages anywhere from #4 to #24.

Cheers, John

I'm going to go lay down...


Peace, pogue

Why isn't welding cable hollow?

*giggling* :D

Quote:

---

Originally Posted by Toga

Why isn't welding cable hollow?

---

I'm afraid I do not understand your question..

Standard off the line welding machines operate at the line frequency of 50 or 60 Hz.

At 60 Hz, the skin depth is approximately two inches. This means, that as the currents required get larger, there comes a point of diminishing returns for total volume of copper conductor. This was found out decades ago by the power companies..they found that once a copper conductor reached about four inches in diameter, further increasing of the diameter produced less advantage than they expected..so, for conductors that have to be larger than 4 inches, they go to a tubular construction...this eliminates the weight of the center copper that does not contribute to conductivity. As an added bonus, they realized that they could water cool the copper pipes, and get even higher current density.

If you examine HF welders in the 5 Khz to 7 Khz range, you will find the wires are made of braided copper, with an external sheath to allow forced air or water cooling of the lines.

If you also consider the application, you will see that they are simply worried about the resistive drop along the wire, not with the energy storage of the wire w/r to inductance. The fact that the welding current lags the voltage has no meaning to the welding process.

So, this is why I do not understand your question..The applications are quite different, with different concerns..

Oh, btw...I'm sure you were only referring to 50/60 Hz and 5Khz/7Khz welders...as your question would be rather silly for 10 Mhz induction welders...

Cheers, John

ahhhh...

lol @ 10MHz welders...

;)

I'm so sorry about your head!

:O

Quote:

---

Originally Posted by Toga

lol @ 10MHz welders...

;)

---

http://www.ewi.org/technologies/plastics/induction.asp

Google is such a wonderful thing..it only takes a minute to find a link with the word "induction welding".

Were you not aware of the technology?

Cheers, John

Actually I was aware of the technology (and yes google can be our friend at times). I was laughing because so many times audio technology is held hostage to bandwidth considerations that are not appropriate. A really good example is negative feedback being a "bad" thing. See this sobering analysis done at the dawn of the "cable" age:

http://www.audio-muziek.nl/audiotechniek/cables.htm

Check the stats on the "welding cable". Somehow people can't do the MATH on skin effect. The truncation to 0.5mm of cylindrical depth from outer diameter at 20kHz may reduce the effective gauge smaller than 12ga., but the IMPEDANCE RATIO is only 1.5 to 1. The majority of the cross-sectional area is WITHIN the skin. This means that for a cable that loses a tiny fraction of a dB for a given length, say 4 meters, it will lose half again as much IN ADDITION to the tiny dB loss. A fraction of a fraction does not a mountain make. Incidentally, give some thought to the ringing that is required in reconstruction filters that try to make 20kHz happen at the edge of the Nyquist theorem for CD, and what a mess that makes of phase in conversion as well. Thus 4 degrees of phase shift from a cable is nothing to fret over, because it gets “swamped� in phase errors of the medium. This is true of tweeters near rolloff, upper end of analog tape, microphones, phono cartridges, and standard bandwidth stable amplifiers. Chasing chimeras with science without handling the tools effectively is like throwing a loaded pistol at someone because you don't know how to shoot it.

If you disagree with this particular "professor" at the link, enlighten me how.

:)

Quote:

---

Originally Posted by Toga

Actually I was aware of the technology (and yes google can be our friend at times). I was laughing because so many times audio technology is held hostage to bandwidth considerations that are not appropriate.

---

Ah, I was confused by your post..

Quote:

---

Originally Posted by Toga

If you disagree with this particular "professor" at the link, enlighten me how.

---

Very well...first, this...

""Skin-depth phenomena have only a slight effect on copper wires at audio frequencies. The skin depth for copper at 20 kHz is about 0,5 mm. Thus, wires larger than No. 14 gauge will have a resistance, at 20 kHz, slightly higher than the d.c. value. The ratio of the 20 kHz to d.c. resistance is given in Table I.""

He has used the old normal to surface, TEM based isotropic propogation model for skin depth...this is entirely inaccurate for current travelling through a conductor..

The skin depth is absolutely related to the <b><i>rate of change of the current</i></b> within the conductor...it is an eddy current based reaction of the conductor to a change in magnetic field caused by the internal currents...NOT from a TEM wave impinging on the surface at a right angle....as such, if the load resistance at the far end of the wire were to halve, the skin depth will change...ya gotta use the bessels to solve this..that exponential TEM model just doesn't do it...I thought I already told you that?

Next, look at table 2...50 uHenries per meter???? Where did that come from..

If you use terman's equation, the inductances are far lower, see my chart..BTW, I verified terman's equation for #24, #18, and #10 guage wires...the equation worked just fine.

So, I think that table 2 should be fixed...It wouldn't have made it past me had I been a referee for publication..

Here's #12 inductance vs spacing, from Terman's equation..

I also note, not a damn thing in that article about localization cues....nothing....I guess if one has no concerns about a virtual soundstage image, then one can ignore that two channel aspect...

Sorry, I had to very quickly glance through the link, so I just picked out the two most glaring inaccuracies..

Cheers, John

PS...Hey, Pogue...would you like me to send you a bottle of aspirin???

PPS...I enjoyed your posts, btw..

Graphs! :D

I appreciate from the example given you are claiming almost 2 orders of magnitude error (thanks for taking the time to elucidate your claims with a visual). Wasn't he talking about 10 METERS, which is more like 33 feet? And don’t you have to double that for the return trip? That gets us into closer agreement.

I’m a little concerned about your scale; While clearly showing logarithmic behavior, the area of interest (0 to 2 inches) has no data points. Are we to assume a linear slope (on the graph with log scale) to near zero?

I'm curious what resolution for timing differences at which frequencies you or research you can point to claim to be audible for sound field position discrimination by the ear/brain. Remember than neural signal pathways fire at rates around 40Hz (at least for pain receptors). My understanding that pulse density or frequency from each hair cell in the cochlea is interpreted in concert with adjacent cells in the spectrum to find both level and midpoint frequency of occurrence. There has to be an upper end of resolution relating to firing rates and such, even given the system is specialized. Got some numbers?

A good mental comparison would be to the GPS system. Military accuracy is down to a few feet, affected by "jitter" in atmospheric transmission and the transmitter timebase accuracy as well. An error signal is "added" to muddle consumer GPS units to lesser accuracy. It would be good to establish arrival time difference perception limits for humans, as this would set a standard for wire measurements and models (and other audio equipment errors, such as channel to channel component tolerances, or even DSP latency issues for pipeline sharing.

Oh hey, did you read the part about confirmation of his models with differential input amplifier MEASUREMENTS? The o’scope may be clearer than what is on paper. If you have the time/patience, please explain what measurement techniques you use to verify your models.

Quote:

---

Originally Posted by Toga

I appreciate from the example given you are claiming almost 2 orders of magnitude error (thanks for taking the time to elucidate your claims with a visual). Wasn't he talking about 10 METERS, which is more like 33 feet? And don’t you have to double that for the return trip? That gets us into closer agreement.

---

No...his table 2 states 3,5 meters and 50 uH. And, it is useless to consider only 1 wire, as there is always a return current path. The terman equation I am using is inductance for the pair, and it includes the internal inductance of the conductor, the 15 nH per foot per wire.

My graph for 100 inches is about 1 uH per foot, 3.28 per meter vs his 50...magnitude and a half error..

Quote:

---

Originally Posted by Toga

I'm a little concerned about your scale; While clearly showing logarithmic behavior, the area of interest (0 to 2 inches) has no data points. Are we to assume a linear slope (on the graph with log scale) to near zero?

---

Well, it didn't concern me....:-)It wasn't an area of my interest, so I didn't bother...attached is one for .1 inch to 5.

Quote:

---

Originally Posted by Toga

I'm curious what resolution for timing differences at which frequencies you or research you can point to claim to be audible for sound field position discrimination by the ear/brain. Remember than neural signal pathways fire at rates around 40Hz (at least for pain receptors). My understanding that pulse density or frequency from each hair cell in the cochlea is interpreted in concert with adjacent cells in the spectrum to find both level and midpoint frequency of occurrence. There has to be an upper end of resolution relating to firing rates and such, even given the system is specialized. Got some numbers?

---

ask and you shall receive..second graph

Quote:

---

Originally Posted by Toga

A good mental comparison would be to the GPS system. Military accuracy is down to a few feet, affected by "jitter" in atmospheric transmission and the transmitter timebase accuracy as well. An error signal is "added" to muddle consumer GPS units to lesser accuracy. It would be good to establish arrival time difference perception limits for humans, as this would set a standard for wire measurements and models (and other audio equipment errors, such as channel to channel component tolerances, or even DSP latency issues for pipeline sharing.

---

You forgot the blue shift as the signal drops into the gravity well..it is 11 km per day error..

Quote:

---

Originally Posted by Toga

Oh hey, did you read the part about confirmation of his models with differential input amplifier MEASUREMENTS? The o’scope may be clearer than what is on paper. If you have the time/patience, please explain what measurement techniques you use to verify your models.

---

It's 5:01 here, tomorrow is another day..don't mind the typo's...it's too fast.

Cheers, John

Um, your second graph shows calculated arrival times, but does it show PERCEPTION? Can you "hear" 1 inch off at 10 feet? There should be a shaded area where you can't hear the placement difference anymore...

Quote:

---

Originally Posted by Toga

Um, your second graph shows calculated arrival times, but does it show PERCEPTION? Can you "hear" 1 inch off at 10 feet? There should be a shaded area where you can't hear the placement difference anymore...

---

You are indeed correct.

The placement error I am setting up to measure will actually be a distribution, and I'll set 1 sigma as the error band. I am confident it will be dependent on angular placement, frequency content, and the transient nature of the envelope.

First part of the test will be a single source, moved side to side behind a screen, and the human will determine where it is, with head fixed.. this will give the sensitivity distribution, the "shaded area" as you call it.

The second part will be the binaural simulation of a moveable source, with same shaded area stuff. However, this part has both ITD and IID components to add in, as well as angular position, freq, and envelope shape..

Last I saw, researchers have shown human sensitivity to 1.5 uSec ITD's...however, that was with simplistic wveforms...I believe more complex stimulus, such as a cowbell, or a voice, provide more information for the brain to use to localize..so humans may indeed be sensitive to ITD below 1.5 uSec...that, btw, is quite difficult to comprehend...I was startled to find that 1.5 uSec ITD was detectible by humans..that is when I ran the excel ss that gave the graph I posted.

Cheers, John