[kinku]

I came across this website( http://diyaudioprojects.com/Power/Low-Inductance-DIY-Speaker-Cables/) while I was trying to improve the bass from my speaker with my new amplifier. I did not have to search around much for the improvement.But some thought remained even after the things changed.
From my search the self inductance of the cable has more impact on high frequency response than capacitance . Capacitive reactance affect lower frequencies more.
But when you try to reduce the inductance by twisted pairs running closely carrying electricity in opposite directions,capacitance goes up. For example in that DIY project using inexpensive wire inductance is better than many expensive brands(may be that is why they stopped publishing those values out).
For the 12Gauge lovers like me, the wikipedia (http://en.wikipedia.org/wiki/Speaker_wire)if it is correct ,will have a wire of inductance about 0.1-0.2uH/feet. Even if you use about 8 feet of wire ,which becomes 16 feet long inductive reactance(0.2x16=3.2 )you can see that you developed a 3.2uH self inductance which attenuate about 2-3 % of high frequency signal . I do not know if that is detectable by normal listening. but again it is only one channel.What about combining both sides?
On the contrary the stary capacitance by braiding wires affect lower frequencies.The losses of more than 1% at 30 Watts develop at a reactance about 1000pf of total length(ref: wiki chart).But again if you are using 10 feet of that DIY cable you are just (65.4x20=1300)in a safe zone with low power applications(30watts).Going down to 8 feet may save some more bass energy.
So something came up to my mind.May be I need to know more. But the million dollar question is
Can I use the braided DIY cable for biwiring my tweeter side and 12AWG zip cord to my woofer side? The resistance per feet are close match (0.10 vs 0.16 ).What do you all think.

[Speedskater]

As you wrote Capacitance and Inductance are opposite sides of the same coin.  If you increase one the other decreases.

Inductance
With low impedance tweeters, low self-inductance cables are good.

Resistance
With some woofer alignments, end to end cable resistance can affect woofer response. But less resistance may not be better.

Capacitance
High cable capacitance can make some poorly designed amplifiers try to turn into oscillators.

[underdawg]

One rule of thumb I use is woofers prefer solid core, and tweeters like stranded.This idea is based on testing and using vmps speakers as the test platform.Brian at vmps knew alot about drivers, crossovers, and cable.

[kinku]

Underdawg what you said matches with theory.Since 12AWG zip cords has low stray capacitance as such compared to most premium configurations it works well for a woofer.
What you all think about bi wiring with zip cord to woofer and braided wires to HF?

[cheap-Jack]

Hi.

From my search the self inductance of the cable has more impact on high frequency response than capacitance . Capacitive reactance affect lower frequencies more.

Don't you mind direct me to whatever research papers to show me capacitance affect LF more than HF???

c-J

[underdawg]

Well I could go more in depth in theory but I prefer listening for some friends of mine that come to listen. I assure you they thought the same as I, solid core for woofers and stranded for tweeters. To be honest I have a hard time with results as cable should be cable, but it all depends on what you hear not theory so much.

[kinku]

C-J please read the wiki link and it explains theory very well.
I tried using 12 awg ZIP FOR BIWIRING my speaker and braided 16 AWG for tweeter. All I can say is it is different for sure. The back ground of music felt more clear and I could seperate each instruments more precisely  with the new wiring.But that made music more analytical.I would listen again ......
Some one else has thought about it well before I bought a Bi wirable speaker it seems. That happens to be Jon Risch
See the link to his web page http://web.archive.org/web/20091021233731/http://geocities.com/jonrisch/biwiring.htm
Is there any way to contact him?

[jneutron]

For the 12Gauge lovers like me, the wikipedia (http://en.wikipedia.org/wiki/Speaker_wire)if it is correct ,will have a wire of inductance about 0.1-0.2uH/feet. Even if you use about 8 feet of wire ,which becomes 16 feet long inductive reactance(0.2x16=3.2 )you can see that you developed a 3.2uH self inductance which attenuate about 2-3 % of high frequency signal . .......On the contrary the stary capacitance by braiding wires affect lower frequencies.The losses of more than 1% at 30 Watts develop at a reactance about 1000pf of total length(ref: wiki chart).But again if you are using 10 feet of that DIY cable you are just (65.4x20=1300)in a safe zone with low power applications(30watts).Going down to 8 feet may save some more bass energy.

The inductance of a typically spaced #12 awg zip is 180 nH per foot.  This number is for the pair, not per conductor.  For a 10 foot zip, there will be 1.8 uH.

The product of inductance and capacitance for a speaker line is relatively constant for a specific geometry of cable.  For a zip, the equation is approximately: (upper range)

LC = 1034 * 4 (6), or LC = 4136 (6204).   L is nH per foot, C is pf per foot.

For a cable of .18 uH per foot (180 nH per foot), the capacitance per foot of cable will be C = 4136/180, or 23 pf per foot (35 pf).

For a 10 foot zip cable, it will have 1.8 uH total, and 230 to 350 pf total.

So something came up to my mind.May be I need to know more. But the million dollar question is
Can I use the braided DIY cable for biwiring my tweeter side and 12AWG zip cord to my woofer side? The resistance per feet are close match (0.10 vs 0.16 ).What do you all think.

Go for it.  The inductance of the cable will be the base inductance of a wire pair divided by the number of pairs within the braid; capacitance will be the number of pairs times the base capacitance.  If you use two pairs, the inductance per foot will be 80 nH and capacitance will be 50 pf per foot.  If you use 4 pairs, the inductance will be 40 nH per foot, the capacitance 100 pf per foot.

The net result will be that the multi braid used on the tweeter side will have less prop delay as the characteristic impedance of the cable will be getting closer to the impedance of the load.   The downside is that at very high frequencies which are below the amplifier's open loop unity gain, if the tweeter unloads (becomes high impedance), there is a possibility albeit small, that the amplifier will oscillate or be marginally stable.  If this becomes the case, a simple zobel at the speaker will solve that issue.  I'd use a zobel at 30 Khz and 8 ohms....simple first order crossover.

If you really want to test the setup for effect, biwire one speaker, connect both speakers to one amp channel, and listen to the image presented.  If the image is stable and exactly centered within the soundfield, then the biwiring you tried had no discernable effect.  If the hf content seems to drift towards the biwired speaker while bass and midrange remain centrally locked between the speakers, then you are hearing a difference.

John

[kinku]

quote]The inductance of a typically spaced #12 awg zip is 180 nH per foot.  This number is for the pair, not per conductor.  For a 10 foot zip, there will be 1.8 uH.

The product of inductance and capacitance for a speaker line is relatively constant for a specific geometry of cable.  For a zip, the equation is approximately: (upper range)

LC = 1034 * 4 (6), or LC = 4136 (6204).   L is nH per foot, C is pf per foot.

For a cable of .18 uH per foot (180 nH per foot), the capacitance per foot of cable will be C = 4136/180, or 23 pf per foot (35 pf).

For a 10 foot zip cable, it will have 1.8 uH total, and 230 to 350 pf total.

[/quote]
Thanks John.That is a very detailed explanation. I may need a little help here.
How do you calculate LC 4136 in that equation and what are the values in bracket?


Please find the impedence plot attached.
Monitor Audio has built in crossover network in bi WIRE configuration .So will that protect against tweeter starting oscillation in output circuit?
Is the possibility of oscillation is related to multi strand wire or it is from wire being of low inductance and hence reactance?
What are the symptoms of oscillation ?
My amp is a tube with output transformers are they more prone to this issue than solid state device?
Using an 8ohm tap over 4 ohm or 4ohm over 8ohm provide any benefit in this situation?
Some of my questions are silly 

[jneutron]

Thanks John.That is a very detailed explanation. I may need a little help here.
How do you calculate LC 4136 in that equation and what are the values in bracket?

The equation is actually derived from transmission line theory.  The full equation is LC = 1034 times epsilon times  mu..
epsilon being relative dielectric permittivity, mu being relative permeability.  Although free space mu is 1, when a cable such as zip is modelled, the fact that the magnetic field is not confined to space of the dielectric as a coaxial or stripline is means a slight mod to the equation.  In cases such as that, it is useful to refer to EDC, or effective dielectric coefficient.  This accounts for the geometry in this calc as well as impedance of the line and prop velocity of the signal.

For ALL wire pairs, the epsilon-mu factor can never be less than 1.  For coax and striplines,  it is exactly the dielectric constant of the insulation.  For all other configurations where either the electric field or the magnetic field can spread outside the conductors, it will be a bit higher.  Zip runs are typically 4 to 6, widely spaced conductors can go up to 10.

In my previous post, I provided calculations based on the use of 4 as the cable EDC, and brackets were used to indicate the same calculations using an EDC of 6.

Monitor Audio has built in crossover network in bi WIRE configuration .So will that protect against tweeter starting oscillation in output circuit?
Is the possibility of oscillation is related to multi strand wire or it is from wire being of low inductance and hence reactance?
What are the symptoms of oscillation ?
My amp is a tube with output transformers are they more prone to this issue than solid state device?
Using an 8ohm tap over 4 ohm or 4ohm over 8ohm provide any benefit in this situation?
Some of my questions are silly 

Oscillation due to weird cables is a result of what the amplifier sees at high frequencies at it's output terminals.  One condition of amplifier stability when high negative feedback is used is that the output load does not force a large phase shift of the signal being fed back before the amplifier internally runs out of gain at hf.  (if you break the amplifier feedback loop and put signal into the input, at some rather high frequency the amplifier's gain will be 1, that is the unity gain frequency.).  Sufficient phase shift can cause the feedback to go positive, and if it does this before the amplifier runs out of gain, it will oscillate.

In a low inductance cable, capacitance will be high. If the cable has a line impedance of 8 and the load is 8, the amp will only see resistance.  If the load becomes very high above the audio band, then current draw will be lower and the amp still sees the cable capacitance.  So as the load "lets go" at hf, the only thing left is capacitance, which can cause phase shift.

Nowadays, most amps aren't designed fast enough to have this problem. 

Bad oscillation at hf will result in smoke either in the amp, tweeter, or crossover.  Low level oscillation would just compromise the sound.

SS vs tube, it really only matters if the design allows excessive phase shift before unity gain frequency.

Sorry for the length of the post.

jn

[jneutron]

For the 12Gauge lovers like me, the wikipedia (http://en.wikipedia.org/wiki/Speaker_wire) if it is correct......

Interesting site content. I fixed the link here, the "if" ended appended to the link addy, so wiki gave a weird page...

Oddly enough, the L and C reactance is not a loss.  The cable L and C are actually energy storage mechanisms.  As a consequence, they will conspire to delay the energy that is being sent to the speaker.  It is never a loss.

jn

[JerryLove]

I would think the.. err... inductors, that are in the crossover, would have a far higher inductance than the cable leading to them.

[kinku]

Thanks John.I actually measured the capacitance of my zip cord bought from monoprice 12AWG and it is 18pf /Feet!I twisted up two zip cords of 16 AWG and capacitance measured is 44. Considering possible measuring errors  with ordinary scale and cheapo DMM it is still very close to the values you provided!!!!!
So my next question would be considering the lowest impedence for high frequency from the plot being around 6Ohms which output terminal should I use to avoid the oscillations possible with high frequency. The 4Ohm or 8 Ohm?

[jneutron]

I would think the.. err... inductors, that are in the crossover, would have a far higher inductance than the cable leading to them.

You are absolutely correct.  We are not speaking about a lot of inductance if one considered only the lumped number.

But given the length of the wire and the horrible mismatch we create with a line z of 150 give or take, a load of 4 to 8 nominal, and a source of zero give or take, the actual line delay intrudes on localization parametrics midband.  By altering the ratio of L to C, we change the settling time for a specific load Z.  My recommendation to run the cabling somewhere near nominal speaker z in midband is geared towards minimizing this delay.

What I speak of is something the industry will have to catch up to.  There's plenty of time, I'm in no hurry.

I note that Cyril Bateman has actually used a reflection bridge to see the delays I speak of.  It is unfortunate his articles are too technical for mainstream publication, the content is very good.

jn

[kinku]

John you are suggesting that with a small risk of making amp unstable from oscillations there might be a quality advantage in using a low inductance cable to tweeter and the best available low capacitance cable being the zip cord, it will benefit the woofer by preventing the delay in passing signals. Did I got it right?
I do not want to ask but seems like you ignored my last question about binding post ,which impedence post should be less risky? 4 or 8Ohms.I always found 8 Ohms as best sounding post from Amp.

[jneutron]

Thanks John.I actually measured the capacitance of my zip cord bought from monoprice 12AWG and it is 18pf /Feet!I twisted up two zip cords of 16 AWG and capacitance measured is 44. Considering possible measuring errors  with ordinary scale and cheapo DMM it is still very close to the values you provided!!!!!
So my next question would be considering the lowest impedence for high frequency from the plot being around 6Ohms which output terminal should I use to avoid the oscillations possible with high frequency. The 4Ohm or 8 Ohm?

Your measurements may be accurate.  The EDC is dependent on the insulation material and the distance between conductors..no problem.

I honestly do not think you will have an oscillation problem regardless of what wires you choose...so don't worry about it.

For any speaker run, figure the characteristic z of a zip as being in the 150 to 200 ohm range ( at audio frequencies it is necessary to include the resistance per foot into the impedance equation and the impedance goes up as a result).  When you parallel several together, they will have an effective z of 150-200 / number of cables.  To get 50 ohms for example, use 4 zips.  For 25, use 8.

Since the resistance is also paralleling, just add the area of the conductors to get a good effective guage.

area vs guage in circular mils:

12----6530
14----4110
16----2580
18----1620
20----1020
22----640


If you want 12 guage, calculate the number of cables you need by division.

using 16:

6530/2580 = 2.5, so 3 16 guage zips gives you better than 12 guage.  Z of about 50 to 70 ohms.

18:
6530/1620 = 4, so 4 #18awg equals a #12.  z of about 37 to 50 ohms.

20:
6530/1020 = 6.4, so 7 #20awg betters #12, z of about  22 to 29 ohms.

22:
6530/640 = 10.2, so 10 #22 awg is about #12, z of 15 to 20 ohms.

Given how amazingly cheap it actually is to create #12 awg cables with characteristic impedances running from 150 to 15 ohms, I see no reason NOT to try it.  It's not rocket science, and it's pennies..

Me, I'd go down to number 20 or 22 construct and be done with it.  If you don't hear any difference (you may not), you've not spent a lot.

jn

[jneutron]

John you are suggesting that with a small risk of making amp unstable from oscillations there might be a quality advantage in using a low inductance cable to tweeter and the best available low capacitance cable being the zip cord, it will benefit the woofer by preventing the delay in passing signals. Did I got it right?

For the midband signals (500-3 or 4 K), it may make a difference.  But remember, it may not either.  It is so easy to fool ones' self into believing either way.  I've posted the technically correct reasoning, but the system and human variables are staggering.
I've never heard about low capacitance for woofers before, I'd suspect it plays little in bass.  But if the woof goes into 1-2k region, who knows.

 

I do not want to ask but seems like you ignored my last question about binding post ,which impedence post should be less risky? 4 or 8Ohms.I always found 8 Ohms as best sounding post from Amp.

sorry, missed that.
Honestly...go with what you like.  If 8 sounds better to you, it's obviously the best choice.  It'd be a 50-50 chioce for me as well..

jn

[JerryLove]

I note that Cyril Bateman has actually used a reflection bridge to see the delays I speak of.  It is unfortunate his articles are too technical for mainstream publication, the content is very good.

I think that, in terms of affecting the industry, the approach is backwards.

Do an impulse plot of pink noise.

Change cable. Repeat plot. Illustrate (and qualify) real, frequency dependant, delay to the room. Could even do something similar to a waterfall plot.

Everything else would be attempting to explain... at present the goal should be to illustrate that a real difference exists in what gets to the ear. *that* should be relatively simple.

I'd like to see some real, measured numbers from a microphone caused by real-use cable configurations.

[jneutron]

I think that, in terms of affecting the industry, the approach is backwards.

You are welcome to your opinion. 

What I speak of is actual engineering, both test and results.  As I stated, the difference may or may not be noticeable in the wild.
I also must say, anyone who states that "actual engineering or physics says cable changes we discuss cannot alter sound"..doesn't really understand the problem.  It is unfortunate, but the level of understanding of even signal energy prop delay in a simple cable is terribly misunderstood in the "industry".  Darn near everybody believes a speaker cable propagates the signal at the inherent cable velocity...totally incorrect of course.  It is 2 to 4 orders of magnitude slower for low load Z.  A 10 foot zip cord cannot have the signal power travelling at 2 ns per foot unless the load Z is in the 200 to about 400 ohm range...to state that it can ignores the limits of relativity.  Superluminal velocities would be required.

The industry needs to meld localization theory with actual t-line theory, and actually learn how to measure high slew currents in low z systems.  That will take a while.  And since I do not make a living in the audio industry, I can be patient..  I apply it to my work, but that is a somewhat different level...

Do an impulse plot of pink noise.

Change cable. Repeat plot. Illustrate (and qualify) real, frequency dependant, delay to the room. Could even do something similar to a waterfall plot.

Everything else would be attempting to explain... at present the goal should be to illustrate that a real difference exists in what gets to the ear. *that* should be relatively simple.

I'd like to see some real, measured numbers from a microphone caused by real-use cable configurations.

So then, it is your assertion that an impulse plot of pink noise will clearly show a midband delay of anywhere from 5 to 50 microseconds? (actually, we are sensitive to 1.2 uSec interchannel, but that is a discussion beyond the scope of the forum in terms of image stability)..

And, if a switch is made from normal to biwire, that the composite soundfield modifications which result from driver to driver timing changes and consequent polar pattern changes in the 5 to 50+ microsecond range will be seen by a microphone?

Sounds like a rather good microphone with some really hairy signal math..

jn

ps.  Until the "industry" actually learns this stuff, nobody's gonna accurately measure anything along these lines.  That is why I mentioned the audibility test before.  If it discerns an audible difference in the mono setup, there may be one at the stereo level.  If not audible in the mono situation, there cannot be one in the stereo case.

[kinku]

Thank you John.
So regardless of AWG of wire a single zip wire will have an impedence between 150-200?
Is it per feet ?
I was thinking of using Goertz M1 (http://www.goertzaudio.com/contents/en-us/d16_MI_Speaker_Cables.html)   instead of DYI wire since it has the lowest inductance value I know.I just noticed that the Impedence is 4. So I was planning to use Goertz M1 for HF and ZIP cord for LF since the capacitance of Goertz cable is 500pF /feet . But from this discussion it seems like Zip cord even with low capacitance offers higher impedence.
But going back to wikipedia the capacitance affect LF (copied from wikipedia)

 This table shows the capacitive reactance in ohms (higher means lower loss) for various frequencies and capacitances; highlighted rows represent loss greater than 1% at 30 volts RMS:


The voltage on a speaker wire depends on amplifier power; for a 100-watt-per-channel amplifier, the voltage will be about 30 volts RMS. At such voltage, a 1% loss will occur at 3,000 ohms or less of capacitive reactance. Therefore, to keep audible (up to 20,000 Hz) losses below 1%, the total capacitance in the cabling must be kept below about 2,700 pF.
Ordinary lamp cord has a capacitance of 10–20 pF per foot, plus a few picofarads of stray capacitance, so a 50 foot run (100 total feet of conductor) will have less than 1% capacitive loss in the audible range. Some premium speaker cables have higher capacitance in order to have lower inductance; 100–300 pF is typical, in which case the capacitive loss will exceed 1% for runs longer than about 5 feet (10 feet of conductor).
I am not sure at this point should I be looking at impedence of the wire vs each reactance seperate?