AudioNervosa
Specialists => Audiologists => Topic started by: rollo on January 16, 2018, 10:42:32 AM
-
What is the desired ga. for a speaker cable less than ten feet long ? How does a different ga. affect the sound ? Then winding of such ? Litz, Parallel, twisted ??
charles
-
That's a mighty big question for so few words. :D
I use 5ga SC - Top that! I've yet to hear better sounding SC in 10 years of comparisons. It's twisted cable inside, but it's shielded in a stiff tube that holds the bend like solid metal.
The conductor alloy has a big effect on the SQ. Pure copper rings (solid copper rings badly) and the trick is to voice the ring with various cable building techniques like you describe. Alternatively, alloy conductor can prevent or damp the ring, making a very quiet cable without special construction technique. coughcoughjpscough
The old rule of thumb is 16ga is adequate for speakers. I have 16ga whole house speaker wiring. It works but I don't care how well.
Larger diameter cable = less resistance = better electrical damping = better driver control = better transient response = more detail. It's stiffer and more expensive.
I'd like to try these (http://www.laventure.net/tourist/cables.htm) DIY pure silver cables. Scroll down to the "Wire Gauge Selection" section for discussion of speaker cable gage philosophy.
-
Thanks professor.
charles
-
What is the desired ga. for a speaker cable less than ten feet long ? How does a different ga. affect the sound ? Then winding of such ? Litz, Parallel, twisted ??
charles
Hi Charles,
Here is some information I posted on my forum sometime back. It deals with various gauges of wire, DC resistance, inductance and inductive reactance. I updated my post.
There has been a lot of confusion over the years concerning twisted wires and whether that is the best type of speaker wires. One could also consider internal speaker wiring as well. Some are dealing with long speaker wires from one room to another. I used "red" and "black" as the speaker wires.
Frankly, I parallel 18 gauge wires from amplifier to single pair jacks on my test two ways. In my case I found 10 wires per speaker lead sounded best. The larger number of paralleled wires not only increased the total gauge of each lead, thus less DC resistance, but also vastly reduced the inductance of each lead.
Inductive reactance is frequency dependent. For comparison, below is the DC resistance, the straight wire inductance, and inductive reactance for 5 feet of single wire and parallel wires at 20khz:
Single Single Single 10 parallel
18 gauge wire 13 ga. ~9 ga. 18 ga.
.0325 .0104 .0066 .0065 DC resistance
2410 nh 2232 nh 2162 nh 482 nh
.30 ohms .28 ohms .27 ohms .06 ohms Inductive reactance at 20khz
Notice the 10 parallel 18 gauge wires equate to an approximate 9 gauge wire in DC resistance, but the inductive reactance is only ~1/5th that of the 9 gauge wire.
(You may have heard the past arguments over the small wire parallel with the large wire improving the highs. The highs were increased because the inductance/inductive reactance of the two wires is about 1/2 that of a single wire.)
Remember, this is for a single 5 foot lead, not both leads. Multiply this by 2 for the total 10 feet speaker lead length. I would say .54 ohms is quite a bit in series with a 4 ohm, or 8 ohm speaker, although the impedance of a speaker at 20khz is higher. One could easily be about half a db down at 20khz, a couple of tenths at 10khz and even a tenth at 5khz. Whether one notices a difference will depend upon some factors. I and friends perceived a difference between 8, 10, and 12 parallel wires per lead. 10 Sounded best in my system with amp output Z of approximately 0.4 ohms and 4 ohm speaker. Your mileage may vary.
I used two conductor "doorbell" wires from ACE hardware. Both wires inside the jacket were connected in parallel. I separated the red speaker lead wires from the black speaker lead wires to minimize the capacitance, to just a few pf.
Concerning twisting speaker wires, to cancel inductance/inductive reactance, here is a post reply I made.
In order to eliminate (wire) inductance, both red and black wires would have to occupy the same space, which they cannot. Insulation thickness between wires also hinders attempts to lower inductance. Twisting is spiralling.
A simple comparison example from: http://www.bdloops.com/twisted_wire_test.html
- 6 x 12 Twisted pair loop (7 twists per foot*) with 20 twisted lead-in, jacketed, 16AWG. 64.1 µH
uH is micro (1 millionth of a henry) 38 feet total length in large loop. My 10 foot 10 conductors of 18 gauge parallel conductor example only has some 964 nano henries (nano = 1 billionth of a henry).
Wrapping them together will create capacitance, as you stated, which we also do not want, especially when higher frequencies involved.
Here is a quote from another post on another forum.
Inductance cancellation of twisted wires will depend upon the wire gauge (distance from center to center, thus lack of inductance cancellation), insulation thickness for sure (again influencing center to center, thus more lack of inductance cancellation), and frequencies involved.
Here is an example of twisted.
Wire size 18 gauge, dia is 1.023 mm
Separation between centers of two wires: 3 mm
Insulation between wires: ~1mm
Inductance: 18 nh/inch, 1080 nh/5 feet, 2160 nh/10 feet
4 mm separation of centers; Inductance is 21 nh/inch, 1260 nh/5 feet, 2520 nh/10 feet
What about the heavy two wire cables, 6, 8, or even 10 mm between centers? 6 mm is 3000 nh (3 uh) for 10 feet, 10 mm is 3624 nh for 10 feet.
However, for 10 parallel wires the inductance is only 964 nh for 10 feet.
It is actually a little more complicated but you get the gist.
It is advantageous to parallel 4 or more wires vs twisting/braiding for both less inductance and less capacitance. I would say 3 minimum parallel wires basically equals inductance of twisted braided wires, depending upon gauge of speaker wire twisted.
The main disadvantage is the WAF.
Cheers and hope this helps.
Steve
-
Hi Steve.
If I want to build myself the ideal 10' speaker cables
(low Ohms, industance, etc...),
what (In simple words) should I do.
Wire size, twisted, solid, starnded, how many strand ? ? ?
I am ready to experiment.
Guy 13
-
Hi Steve.
If I want to build myself the ideal 10' speaker cables
(low Ohms, industance, etc...),
what (In simple words) should I do.
Wire size, twisted, solid, starnded, how many strand ? ? ?
I am ready to experiment.
Guy 13
Hi Guy,
I have been super busy, so a little late in response. Best to use my example and work from it.
I should first asked what is the amplifier and speakers you are using, what output Z and speaker Z.
Cheers
steve
-
Steve, how does one know the amp output impedance without a bench of electronic tools? Is there a way with just a VOM?
-
Hi Steve.
If I want to build myself the ideal 10' speaker cables
(low Ohms, industance, etc...),
what (In simple words) should I do.
Wire size, twisted, solid, starnded, how many strand ? ? ?
I am ready to experiment.
Guy 13
Hi Guy,
I have been super busy, so a little late in response. Best to use my example and work from it.
I should first asked what is the amplifier and speakers you are using, what output Z and speaker Z.
Cheers
steve
Hi Steve.
Better late, then never !
The amplifier is a Decware SE84C+ and the speakers are Audio Nirvana 8 Ohms Open Baffle.
The distance from - to is maximum 10 feet and the end at the speakers are soldered to the terminals
and the other end at the amplifier connected to the amplifier's binding post.
I now use a Signal Cable 14ga. copper.
In the past, I've tried some copper 12ga. copper electrical extension cable with rubber exterior enveloppe.
Didn'r hear any difference between the two.
I have some 14ga, OFC solid wire purchased from Danny Rithchie, but I,ve not
tired it, because they are too ridgid.
Guy 13
-
Steve, how does one know the amp output impedance without a bench of electronic tools? Is there a way with just a VOM?
Hi Scott,
There is a way which involves measuring a sine wave (usually 400 or 1khz, but any lower F is ok for our purposes, say 60hz) with out a load. In triode mode the SE amp only puts out ~1.8 watts. For that amp, a 10 watt 2 to 5 ohm rheostat, or 2 watt 2 to 5 ohm resistor should work.
I would not worry about the voltage meter loading the amplifier's output.
Now adjust the rheostat until the output voltage measures 1/2 the no load voltage. At such point the output impedance (Z) equals the load/pot Z.
One probably won't have a rheostat, so one could measure the No load output voltage, and then place a low ohmage fixed resistor, as mentioned above, across the output.
Now measure the voltage with the load connected. Just as before, a voltage divider, a voltage ratio is created, so a resistance ratio is created.
In general, for any newbies out there, this example.
Say we measure 3 volts no load. Now we add a fixed 1.5 ohm load resistor, and measure 2 volts output.
We now know that 2 volts is across the 1.5 ohm load, so 1 volt is across the output impedance of the amplifier, to equal 3 total volts.
1 volt across the amplifier output Z and 2 volts across the load. That is a 1 to 2 ratio. We know the value of the load resistance, 1.5 ohms. So 1/2 of that is 0,75 ohms. That is the approximate value of the amplifier output Z.
The last question is, how do we know the amplifier is not clipping, which could influence the reading? I would listen to a tone (say from YouTube) that is loud but not clipping. That should be close enough.
I hope this helps Scott.
steve
-
Hi Steve.
Better late, then never !
The amplifier is a Decware SE84C+ and the speakers are Audio Nirvana 8 Ohms Open Baffle.
The distance from - to is maximum 10 feet and the end at the speakers are soldered to the terminals
and the other end at the amplifier connected to the amplifier's binding post.
I now use a Signal Cable 14ga. copper.
In the past, I've tried some copper 12ga. copper electrical extension cable with rubber exterior enveloppe.
Didn'r hear any difference between the two.
I have some 14ga, OFC solid wire purchased from Danny Rithchie, but I,ve not
tired it, because they are too ridgid.
Guy 13
Hi Guy,
This should be interesting since an "8" ohm speaker divided by the amp's output Z (damping factor) is probably different than my setup. Your previous values of wire may be too large for one to perceive any difference. It could also be optimum as well. We just don't know.
Are you up for an experiment Guy? If so, try installing just one solid 18 gauge wire in each speaker leg. Sounds extreme, but it should tell us if the "sound" can be altered in your particular setup. If it does, has the sound become thin, what about increased and/or bloated bass? How about the lower midrange/upper bass?
If you notice a sonic difference and it sounded worse than your larger wires, try two 18 gauge wires in parallel for ~15 gauge. Does it sound better, worse, or no difference? If it sounds better than one 18 gauge wire, try a 3rd for ~13.5 total gauge. A 4th will be ~12 total gauge. Keep adding on until the sound starts to become worse. (You may end up purchasing many inexpensive twin banana plugs, and chaining them together for larger and larger cable. Eventually, you want to rid of the banana plugs and just connect using the wire.)
For paralleling wire, here is a calculator.
https://www.wirebarn.com/Combined-Wire-Gauge-Calculator_ep_42.html
I think ACE hardware might have something in stock for testing purposes.
Just a thought Guy. If you decide not to experiment, no sweat, as it does become time consuming, and price does go up.
Cheers
steve
-
Hi Steve,
thanks for the info.
Yes, I will do the experimenting according to your info.
However, here on planet Vietnam,
not that easy to get solid core wires,
plus the wire's gage are in different value
plus all my audio stuff is packed and sealed,
ready to go on the other side of the planet,
however, I will do the test once I'm in my home country
with everything readily available.
Thanks again for your precious help.
Guy 13
-
Hi Steve,
thanks for the info.
Yes, I will do the experimenting according to your info.
However, here on planet Vietnam,
not that easy to get solid core wires,
plus the wire's gage are in different value
plus all my audio stuff is packed and sealed,
ready to go on the other side of the planet,
however, I will do the test once I'm in my home country
with everything readily available.
Thanks again for your precious help.
Guy 13
Thank you Guy. I know it will take time and shipping over seas can cause all sorts of problems. So after you arrive, if you have problems, no sweat concerning no experiment.
If shipping all works out nicely (that is a big if), and you do the experimenting, thank you for the time and efforts, we will all look forward to your results.
Cheers
steve
-
Guy13,
You can also use many strands of Ethernet cable like CAT5. They are made of solid core usually 24awg and insulated with PTFE or similar material. Each cable has 8 strands of 24awg solid core usually. To make the cables smaller, you can remove the outer jacket. One 8 strand cable of 24awg is 21awg is paralleled. Two eight strands cables in parallel is 20awg. For every doubling of the number of strands the awg decreases by 1.
The Ethernet patch cable version is stranded wire, so you want what they call plenum Ethernet cable. Might be easier than finding bulk solid core wire.
Just a thought.
-
Guy13,
You can also use many strands of Ethernet cable like CAT5. They are made of solid core usually 24awg and insulated with PTFE or similar material. Each cable has 8 strands of 24awg solid core usually. To make the cables smaller, you can remove the outer jacket. One 8 strand cable of 24awg is 21awg is paralleled. Two eight strands cables in parallel is 20awg. For every doubling of the number of strands the awg decreases by 1.
The Ethernet patch cable version is stranded wire, so you want what they call plenum Ethernet cable. Might be easier than finding bulk solid core wire.
Just a thought.
Hi HAL,
I will look into my audio - computer junk yard - shoe box - spider web corner to see what I've got.
Thanks a lot for that info.
I would have never thought bout that myself.
That's why I am here on Audio Nervosa.
Guy 13
-
Guy13,
You can also use many strands of Ethernet cable like CAT5. They are made of solid core usually 24awg and insulated with PTFE or similar material. Each cable has 8 strands of 24awg solid core usually. To make the cables smaller, you can remove the outer jacket. One 8 strand cable of 24awg is 21awg is paralleled. Two eight strands cables in parallel is 20awg. For every doubling of the number of strands the awg decreases by 1.
The Ethernet patch cable version is stranded wire, so you want what they call plenum Ethernet cable. Might be easier than finding bulk solid core wire.
Just a thought.
Hal, I was wondering how you figured 8 parallel 24 gauge wires only equates to one 21 gauge wire, and every doubling of same gauge wires increases the gauge size by 1? Am I reading your post correctly?
If not, please correct me and accept my apology.
According to my sources, the Radio Amateurs Handbook, RCA Radiotron Designers Handbook, and this link:
https://www.wirebarn.com/Combined-Wire-Gauge-Calculator_ep_42.html
Every doubling of same gauge wires equates to approximately (~) a single 3 gauge larger wire in terms of resistance.
For example, doubling the wires from one 24 gauge to two 24 gauge wires equates to ~ one 21 gauge wire in terms of resistance. Doubling again to 4 wires equates to ~ one 18 gauge wire's resistance. Doubling again to 8 wires equates to ~ one 15 gauge wire's resistance. The key we are looking for is the resistance of parallel wires equating to one larger wire.
Cheers
steve
-
There is a correction. For every doubling of surface area the awg goes up by a factor a 2 not 1. I had read it was one, but the awg charts shows it is 2.
So to go from 24 awg to 18 awg is only a 4 wire bundle and one piece of Ethernet cable all wires in parallel is 16awg.
-
There is a correction. For every doubling of surface area the awg goes up by a factor a 2 not 1. I had read it was one, but the awg charts shows it is 2.
So to go from 24 awg to 18 awg is only a 4 wire bundle and one piece of Ethernet cable all wires in parallel is 16awg.
I see where you are coming from Hal, but resistance is better for my experiments and Guy's and others who will perform future experiments with their speaker wires.
1. DC wire resistance is easily available in the gauge charts.
2. Paralleling wires will automatically reduce the inductance at higher frequencies.
3. What we are interested in is both 1 and 2 since the DC resistance affects the damping factor and the inductive reactance affects sound at higher frequencies. We need to optimize both total DC wire resistance and inductive reactance to the whole speaker.
I don't mean anything by disagreeing, just wish to keep things meaningful, straight forward, and simple.
cheers
steve
-
It is so nice to see two guys discussing a differing point of view in rational terms without resorting to insults and name calling like in some other areas of the net whose name shall not be mentioned....... :roll:
That's why I'm still here after all these years. :thumb:
-
It is so nice to see two guys discussing a differing point of view in rational terms without resorting to insults and name calling like in some other areas of the net whose name shall not be mentioned....... :roll:
That's why I'm still here after all these years. :thumb:
You are right TMAZZ
this place is civilized compared to many other places that I have visited
or being member.
I am anxious to see new members joining or civilized forum.
Guy 13
-
One final thought. It is possible that one will not experience any sonic difference if the system, room, or both are masking the inner detail etc. Bass heavy rooms, high db modes, dull sounding rooms, poor part(s) in general can mask so one won't perceive any sonic difference.
I should have posted this information sooner as newbies may not understand these principles as our veterans clearly understand.
cheers
steve
-
One final thought. It is possible that one will not experience any sonic difference if the system, room, or both are heavily masking the inner detail etc. Bass heavy rooms, high db nodes, dull sounding rooms in general can mask and make the tonal quality so full that one won't perceive any sonic difference.
I should have posted this information sooner as newbies may not understand these principles as our veterans clearly understand.
cheers
steve
So true. How many times have we solved one problem only to find that we had several more that the first problem was covering up? :roll:
-
One final thought. It is possible that one will not experience any sonic difference if the system, room, or both are heavily masking the inner detail etc. Bass heavy rooms, high db nodes, dull sounding rooms in general can mask and make the tonal quality so full that one won't perceive any sonic difference.
I should have posted this information sooner as newbies may not understand these principles as our veterans clearly understand.
cheers
steve
So true. How many times have we solved one problem only to find that we had several more that the first problem was covering up? :roll:
I have long held the opinion that the room is THE most important interconnect in a good to great system. Ya gotta be able to hear the source material as recorded.
Know what I mean?
-
Agree with the room being the important factor -- which one of you guys can explain that to my wife?
-
Agree with the room being the important factor -- which one of you guys can explain that to my wife?
That would be none of us. :roll:
-
Agree with the room being the important factor -- which one of you guys can explain that to my wife?
That would be none of us. :roll:
:rofl: :thumb: