lol, I showed measurements of two different types of wires yet the post about silver plated copper gets the attention.
I don't think these kinds of measurements are feasible using a sound card based measurement system. The differences you see are no doubt the result of unidentified confounders.
- The results were repeatable, I tested multiple times
- they were very similar in two different places (Halair's setup had the same thing happening, which made me test it)
That's enough for me to pick the one that performs best in measurements.
My point about more testing before wasn't aimed at the results you got in your circumstances. I was thinking that to draw a conclusion about wire in general would need to be done across different devices with different amplifiers to see if the same differences were measured there. 14 gauge multi-stranded copper is very close to most of the speaker cable I have, perhaps the other stuff you had was of more audiophile construction and could have induced ultrasonic oscillation or some other unknown "feature" of the wire construction.
Resistance is most important for insertion loss. The perfect wire doesn't exist as all have a certain measure of R, L and C. Most manufacturers tend to specify either inductance or capacitance not both. Lower capacitance usually means higher inductance, very low inductance cables didn't work well in the pass paper. Professionals are more concerned about capacitance as with very long runs the parameters create a low pass filter that comes closer to the audio band. The lengths in a home situation shouldn't cause issues with any reasonable quality cable.I don't think resistance is the only reason, all three parameters are going to determine the final outcome. That part is dependent on the type of load and amplifier. I agree on using wire with low enough resistance, which I had used as a guide line to buy my wire.
If I had not seen Halair's measurements I wouldn't have had any clue to check this at all. If he had not used two different type of wires on his left and right channel I wouldn't have seen it.
You tell me (not directed to fluid specifically
Have a look at the graphs at the bottom of this page of Canare Speaker cable. They graph the response of the cable due to the filter created by the cable parameters. It takes a long cable to make much of a difference but you can see that in some wires it can create a resonant peak around 100K. The 4S8 looks pretty good to me.
Canare Corp.: Speaker Cables: Speaker Cable (Single)(4S8)
For me I would buy good quality cable thick enough to do the job that comes in a package you like. I prefer professional style cable like Canare or the Thomann brand. Very standard construction nothing weird, no thin strands, no braids or audiophile claims. Capacitance in these is low enough not to worry about.
This is what I used and I would buy it again
Attachments
For cable debate : )
In past a GBR university had a nice flash application on their web site that could output response for speaker cable based core diameter plus cable length and core spacing but unfortunate it isn't available anymore. Take it with a grant of salt but as far as remember core spacing and diameter had big influence on inductance loop so that last octave of audio band could get a little bit low pass downward slope especially when spacing was more than few millimeters.
At True Audio website ( https://trueaudio.com/post_009.htm ) it tells we can ignore inductance and capacitance for the short runs in home environment below 10 meters and for resistance the problem is general SPL loss and also amp/speaker integration where voltage amp start to behave as current amp meaning any little wiggle in impedance curve start to show up in frq response curve.
Silver wire and conductivity / oxygen free wire / Resistance too high or low and many other wire-busters are nice explained at Roger Russell website here Roger Russell, McIntosh Lab., Scripto and Norma Pencils and leads (pick the link "Speaker Wire" under "Speaker Related Sites").
Myself use twisted solid core copper and also shielded whenever i can find it based its a simple route and my head can't think of a signal passing multiple strands will reach the end at same time and in same response condition when we have with RLC plus corrosion plus bends and other nasty's to do. Also like if copper is tin plated to protect copper from corrosion and in that skin depth dictates HF to run in outer diameter only, it should take the copper route because that has lowest impedance, where tin conductivity is only 15% of copper so tin is probably only 14,99% more conductive than a PVC jacket : ). Have few times taken audio band limited DATS unit and benched own homemade solid core cables and so far they look good and as long they do that my opinion is probably difficult to change unless some scientific proof gets on the table. Using shielded speaker cable is a wish to decrease antenna effect that could inject nasty's into input stage of power amp via its feedback network. Admit solid core copper is not very pratical in use, it can break over very easy at the steady soldering point if we start vibrate the lead too much and also sound can sometimes in a setup be less pleasing than a stranded multicore, but that is usually forgot and forgiven when speaker system is in situation corrected.
Can understand you wesayso in seeking info because its not just a simple plug/unplug trying different wires inside your arrays, hope input is little bit help and if you haven't seen the Roger Russell site before maybe it can bring some value.
In past a GBR university had a nice flash application on their web site that could output response for speaker cable based core diameter plus cable length and core spacing but unfortunate it isn't available anymore. Take it with a grant of salt but as far as remember core spacing and diameter had big influence on inductance loop so that last octave of audio band could get a little bit low pass downward slope especially when spacing was more than few millimeters.
At True Audio website ( https://trueaudio.com/post_009.htm ) it tells we can ignore inductance and capacitance for the short runs in home environment below 10 meters and for resistance the problem is general SPL loss and also amp/speaker integration where voltage amp start to behave as current amp meaning any little wiggle in impedance curve start to show up in frq response curve.
Silver wire and conductivity / oxygen free wire / Resistance too high or low and many other wire-busters are nice explained at Roger Russell website here Roger Russell, McIntosh Lab., Scripto and Norma Pencils and leads (pick the link "Speaker Wire" under "Speaker Related Sites").
Myself use twisted solid core copper and also shielded whenever i can find it based its a simple route and my head can't think of a signal passing multiple strands will reach the end at same time and in same response condition when we have with RLC plus corrosion plus bends and other nasty's to do. Also like if copper is tin plated to protect copper from corrosion and in that skin depth dictates HF to run in outer diameter only, it should take the copper route because that has lowest impedance, where tin conductivity is only 15% of copper so tin is probably only 14,99% more conductive than a PVC jacket : ). Have few times taken audio band limited DATS unit and benched own homemade solid core cables and so far they look good and as long they do that my opinion is probably difficult to change unless some scientific proof gets on the table. Using shielded speaker cable is a wish to decrease antenna effect that could inject nasty's into input stage of power amp via its feedback network. Admit solid core copper is not very pratical in use, it can break over very easy at the steady soldering point if we start vibrate the lead too much and also sound can sometimes in a setup be less pleasing than a stranded multicore, but that is usually forgot and forgiven when speaker system is in situation corrected.
Can understand you wesayso in seeking info because its not just a simple plug/unplug trying different wires inside your arrays, hope input is little bit help and if you haven't seen the Roger Russell site before maybe it can bring some value.
Hahaha, well I have seen and read Roger's website. In his own speaker he uses "Cardas wire with all connections soldered".
So I'm not sure that helps, he didn't consider parallel series wiring to be of benefit.
I liked what I saw from the thomann "the sssnake" wire, and may opt for the one step larger size (spk225) as speaker wire and also use it as the main wire leads inside the enclosure. That leaves me to choose the smaller leads to the drivers. So far it seems I'm still choosing 0.25 mm (in multiples) wherever I look. Solid core not being practical in my application.
BYRTT, was it something like this: https://www.ampbooks.com/mobile/amplifier-calculators/wire-capacitance/calculator/
The distance between both wires is the main factor they consider, apart from the size of the wire.
So I'm not sure that helps, he didn't consider parallel series wiring to be of benefit.
I liked what I saw from the thomann "the sssnake" wire, and may opt for the one step larger size (spk225) as speaker wire and also use it as the main wire leads inside the enclosure. That leaves me to choose the smaller leads to the drivers. So far it seems I'm still choosing 0.25 mm (in multiples) wherever I look. Solid core not being practical in my application.
BYRTT, was it something like this: https://www.ampbooks.com/mobile/amplifier-calculators/wire-capacitance/calculator/
The distance between both wires is the main factor they consider, apart from the size of the wire.
Hahaha great you been everywhere.
Alright did some digging into HDD subfolders to look for that information brain think it had seen in past : )
Here is a good one on maximum frq for 100% skin depth and max amps for chassis wiring verse power transmission, for example even AWG24 is only 0,577 amps for power transmission into chassis wiring its good for 3,5 amps ( American Wire Gauge table and AWG Electrical Current Load Limits with skin depth frequencies and wire breaking strength ).
Did also find the remembered calculator it was a Scottish university and it wasn't flash based but use Java and the broken thing was its not certified by Java's new owner, to get it running had to into Java control panel on "Security" tab then "Exception Site List" fill in below string and it worked at my computer after closing the browser and go out there once more.
They actual had more interesting stuff that i remembered into this subject, see headlines here ( https://www.st-andrews.ac.uk/~jcgl/Scots_Guide/audio/Analog.html ) and after fixing above security issues pick headline "Java Applet for Twin-Feed properties" read the guide and see the fun sweeping arrows up and down.
Alright did some digging into HDD subfolders to look for that information brain think it had seen in past : )
Here is a good one on maximum frq for 100% skin depth and max amps for chassis wiring verse power transmission, for example even AWG24 is only 0,577 amps for power transmission into chassis wiring its good for 3,5 amps ( American Wire Gauge table and AWG Electrical Current Load Limits with skin depth frequencies and wire breaking strength ).
Did also find the remembered calculator it was a Scottish university
and it wasn't flash based but use Java and the broken thing was its not certified by Java's new owner, to get it running had to into Java control panel on "Security" tab then "Exception Site List" fill in below string and it worked at my computer after closing the browser and go out there once more.
Code:
https://www.st-andrews.ac.uk/~jcgl/Scots_Guide/audio/Java/twin.html
Last edited:
Clearly most amps would prefer a purely resistive load, so minimizing the inductance and capacitance as much as possible should be the goal, but as was mentioned if you try to minimize one you usually increase the other.
PTFE/Teflon has a low dielectric constant and very stable dielectric properties over time and temperature. That should make it one of the best materials to use for the jacket material to help minimize capacitance. Air is probably the next best realistic material, unless you can put the wires in a vacuum
To reduce inductance you need to reduce the loop area (keep the power and return wires close together) and increase the wire diameter. Unfortunately, reducing the loop area also increases the capacitance, but you could experiment with the spacing to see if/how it affects the frequency and impulse response.
https://www.easycalculation.com/engineering/electrical/parallel-wire-impedance-calculator.php
As you said a few times, the amp, wire, and speaker are a system.
That is the same wire that I have used in my cables that connect the speakers to the amp. I used the thinner gauge inside the cabinet to make it easier.
For interference rejection the Canare cable has 4 wires in a "Star quad" arrangement that is sometimes used in microphone cable. By connecting the two similar coloured wires together less interference is induced into the wires due to the winding inside the jacket. Thomann has a four conductor cable of 2.5mm2 that could be used in the same way. Cheaper to buy it that way if you want 4 wires and are going to strip the outer sheath off anyway.
Now if only I could make sense of it, maybe then I could make an educated guess
Both wires I linked were the few that actually show a bit more actual data in the datasheet. Most wires datasheets only show their size.
I have no clue what to look for as far as amp/wires/speaker goes. All I know is there should be some balance there.
For instance: silver plated copper wire is often referred to as bright. Without knowing what it really does it's kind of hard to guess.
With my speaker wire tests I did see difference in SPL level at the top end. As said, those differences would vanish with DSP, except when the phase gets disrupted outside of the FDW I use. And why would that happen?
Most phase related deviations over time were directly related to reflections/diffraction. What is it that a wire could do different to show similar behavior. In other words, what combination of resistance, inductance and capacitance would be the right choice.
As hinted by samuelj some of these numbers should only influence an area way outside the pass band of the speakers. Yet it does seem to influence the area of interest, like the Polk wire mentioned in the paper by Mr. Pass.
I am by no means an expert on electrical properties. The graph shown with the difference in group delay of the wires I tested is not a normal view you'd see. It's like looking at it under a microscope:
Normally, with usual time-scale it would look like this:
That's the same wire on both plots. Though if using another wire can get this under that same microscopic view:
It results in a similar difference as the whole series/parallel debate. My gut tells me to go for the one that measures smoothest. Without asking the question: can you hear that?
So far that route has given me quite good and pleasing results. Half the time I read about people that say: we can't measure what we hear. And when we do get differences turn up in measurements others are likely to say: but can we hear it?
I think we probably will have difficulty hearing some of the differences we measure, but every little detail that makes the overall behavior smoother is going to help to get a good or better end result. At least that's my point of view. Which means I can't close my eyes for these differences.
The fact that we measure with one microphone (that has no brain attached to it) and hear with two ears (that has that strange processor attached called our brain) doesn't exactly help.
It still puzzles me how it all can work inside our head. I often wonder if I were to invite singing identical twins. Place each of them at a 30 degree angle, one left and one right, as seen from my listening position and make them sing simultaneously, would I hear only one voice in the phantom center?
The ability of our brain to group and sort sound fragments is nothing short of dazzling. One of the important features of that sorting and grouping is timing.
The writing is too small, are those milliseconds? Over one millisecond group delay at 12 or so KHz and a bit less at 17 or so KHz?
I have had similar issues with Litz wire: As soon as the copper strands start to corrode, so the oxide insulates strands electrically from one another and also prevents proper soldering with modern non-aggressive-fluxing-agent solder, sound deteriorates, not just something to become detected only with a 24/96 system but with simpler equipment and my ears, too. Crazy stuff, just what electronics is all about.
0,1 to 0,6 ms in the microscopic views. I will never make the mistake again thinking a wire is a wire, as long as it's the right size for the job.
So just to be clear, IF one were able to put up with the wiring difficulties of 14gauge solid copper wire, it would be superior to the stranded copper wire of equivalent gauge because the oxidation of the copper strands deteriorates the effectiveness of the strands?
If this is true, I'd be using the solid core for the speaker internals only, not Amp to speaker. But maybe?
If this is true, I'd be using the solid core for the speaker internals only, not Amp to speaker. But maybe?
Last edited:
Over the past few weeks I've been re-doing some of my EQ work on the arrays. Actually, I'm not quite done yet.
Participating in fluid's thread and seeing some measurements up close have brought some interesting comparisons to my own set of measurements and this time I'm set to optimize the single position measurements the best I can.
Just as a test setup (extra test zone in JRiver), nothing set in stone but so far I'm certainly not disappointed in the results I got...
Every day or so I'm listening to what I have while I continue to try and optimize the output of the FIR filters.
I'm making it a little easier on myself to split the processing, resulting my Mid Side EQ being much simpler to read.
In the past I've shown a couple of house curves. The one I still like and follow quite closely is the curve as presented by JBL/Toole:
The one marked "Trained Listeners" presented as the "_ _ _" line is very similar to my "mid" part. What I did to get there wasn't just picking out a graph and go with that. I setup the arrays with FIR correction and used some slight tweaks while listening to a lot of different program material and re-measured every now and then. That became the new standard and over time I would tweak again. After the tweaking stopped, most songs just worked, I recorded the end result. Guess what... it basically looked a lot like that graph.
Just coincidence?
Before using the above mentioned method I had tried the steady downwards curve from Sean Olive (dropping 1dB/octave, I think Linkwitz preferred something similar), did sound alright, but just not quite there. I've also tried the B&K curve, it was a little light on the bottom end for my (in room) taste. I tried the Bob Katz preferred curve and still didn't feel I was quite there. That's when I decided to simply grow into what I liked. All of them brought something in one song or another.
The JBL trained listeners curve seems to work on a lot of material (and different genre's) for me. (*)
With my mid side EQ the sides still get some small deviations from that curve.
(which incidentally makes them look more like the 1dB/octave downward curve)
In time I'll post some graphs to show how it all comes together. Still a lot of work ahead, but I'm getting there.
(*) = Disclaimer: what works for me, might not work for you. Tonality is highly preference based.
Just sharing what worked for me, with strongly reduced early (0-20 ms) reflections. How does one even judge a room curve? Is it done by 'molding' a standard 500 ms measurement into shape? A moving mic measurement?
Personally I use the FDW to tell me the shape, however I can confirm the un-gated measurements still show me the same basic shape. Courtesy of the array's ability of averaging out reflections, the absorbing damping panels at first reflection points and the Haas Kickers/ambient channels filling in the shortcomings of the room.
Participating in fluid's thread and seeing some measurements up close have brought some interesting comparisons to my own set of measurements and this time I'm set to optimize the single position measurements the best I can.
Just as a test setup (extra test zone in JRiver), nothing set in stone but so far I'm certainly not disappointed in the results I got...
Every day or so I'm listening to what I have while I continue to try and optimize the output of the FIR filters.
I'm making it a little easier on myself to split the processing, resulting my Mid Side EQ being much simpler to read.
In the past I've shown a couple of house curves. The one I still like and follow quite closely is the curve as presented by JBL/Toole:
The one marked "Trained Listeners" presented as the "_ _ _" line is very similar to my "mid" part. What I did to get there wasn't just picking out a graph and go with that. I setup the arrays with FIR correction and used some slight tweaks while listening to a lot of different program material and re-measured every now and then. That became the new standard and over time I would tweak again. After the tweaking stopped, most songs just worked, I recorded the end result. Guess what... it basically looked a lot like that graph.
Just coincidence?
Before using the above mentioned method I had tried the steady downwards curve from Sean Olive (dropping 1dB/octave, I think Linkwitz preferred something similar), did sound alright, but just not quite there. I've also tried the B&K curve, it was a little light on the bottom end for my (in room) taste. I tried the Bob Katz preferred curve and still didn't feel I was quite there. That's when I decided to simply grow into what I liked. All of them brought something in one song or another.
The JBL trained listeners curve seems to work on a lot of material (and different genre's) for me. (*)
With my mid side EQ the sides still get some small deviations from that curve.
(which incidentally makes them look more like the 1dB/octave downward curve
)In time I'll post some graphs to show how it all comes together. Still a lot of work ahead, but I'm getting there.
(*) = Disclaimer: what works for me, might not work for you. Tonality is highly preference based.
Just sharing what worked for me, with strongly reduced early (0-20 ms) reflections. How does one even judge a room curve? Is it done by 'molding' a standard 500 ms measurement into shape? A moving mic measurement?
Personally I use the FDW to tell me the shape, however I can confirm the un-gated measurements still show me the same basic shape. Courtesy of the array's ability of averaging out reflections, the absorbing damping panels at first reflection points and the Haas Kickers/ambient channels filling in the shortcomings of the room.
Last edited:
You pick what you like
... I haven't tried enough different wire/speaker/amp combos to be of any help. I can only say anything useful about my own combination. Be advised, My FR curve is independent from whatever wire I use. Switching wires on a non EQ system just might have even more influence!I am also reinvestigating the room curve right now. Can't seem to settle on anything. The B&K is too dark for me at the top end. Curious to know where you got the image above. Can you post a link to the paper? And, when you say Olive, are you talking about this presentation?
Audio Musings by Sean Olive: The Subjective and Objective Evaluation of Room Correction Products
(the link to Olive's ppt is in the third paragraph, and the subjectively preferred room curve is on a slide in the presentation)
I am intrigued by the image you posted, especially the relative flatness above 2kHz in the predicted curve. With the narrowing directivity of the TC9, I'm wondering if some sort of flattening above 3-6 kHz is better than a continuously dropping response. At least I feel the need for it.
Audio Musings by Sean Olive: The Subjective and Objective Evaluation of Room Correction Products
(the link to Olive's ppt is in the third paragraph, and the subjectively preferred room curve is on a slide in the presentation)
I am intrigued by the image you posted, especially the relative flatness above 2kHz in the predicted curve. With the narrowing directivity of the TC9, I'm wondering if some sort of flattening above 3-6 kHz is better than a continuously dropping response. At least I feel the need for it.
The link was here http://www.diyaudio.com/forums/full-range/242171-making-two-towers-25-driver-full-range-line-array-145.html#post4482816
http://www.aes.org/tmpFiles/elib/20151013/17839.pdf
The predicted curve is just that, a prediction, it might make more sense when you have read that paper.