OK, let's wade into some of these:
Indeed there are lots of 'tech papers'. None of the competent ones that I am aware of (i.e. those that actually have a basis in the physics of signal transmission -alas, probably about 90% of the 'papers' out there on the subject are utter twaddle) appear to support your contentions. Perhaps you could indicate the ones that do; I would be most interested in reading them, since, despite my notorious cynicism about wire, I have an interest in the subject.
Unfortunately the TNT 'naked truth' page has zero in the way of technical detail. 'First / second order effects?' By whom? By what standard? None are properly referenced with any data at all. You are simply expected to take the author's word. Take skin effect for example, of which much is made. A perfectly ordinary piece of 10ga zip cord can be calculated & measured as having about 0.1dB loss at 20KHz through skin effect. Here's an article with a scientific basis that actually contains such calculations: Skin Effect Relevance in Speaker Cables | Audioholics To put it mildly, you'll have more variability through the drive unit that that. Assuming you can hear it at all, the likelihood of which is about zero, even for the most bat-eared teenagers.
The second link does contain a handful of basic lumped parameters, which is a start, but it goes rapidly pear-shaped by spouting a bunch of unsubstantiated claims on audibility of losses &c. Incidentally, since it sets stock by RFI etc., it does rather contradict your claims that braids etc. 'suck' since these can (can) have superior rejection of RFI than straight antennas. How significant this is depends on circumstance of course. By the way, not all speaker wires using Cat5 are either braided or twisted. Some designs using Cat5 require the individual wires to be untwisted for instance. Do these also 'suck' and if so, why?
'The best TNT project' -according to whom, and to what criteria? They're different, but without comparative data, this means nothing. Sorry. Wire gauge? Resistance? Capacitance? Inductance? And that's before you consider the rest of the system.
If you prefer it, fair enough. Not knocking that at all, to each their own. But to write off other types purely on the basis of your own preference rather than any scientific validity does nobody any favours.
Regarding the 'power audio cable' reference, I hope you mean 'speaker cable' not 'power cable.' Satellite cable is completely unfit for carrying mains currents / voltages & would be a severe fire hazard if used in such a way. To do so would be utter madness, and potentially life-threatening.
Indeed there are lots of 'tech papers'. None of the competent ones that I am aware of (i.e. those that actually have a basis in the physics of signal transmission -alas, probably about 90% of the 'papers' out there on the subject are utter twaddle) appear to support your contentions. Perhaps you could indicate the ones that do; I would be most interested in reading them, since, despite my notorious cynicism about wire, I have an interest in the subject.
Unfortunately the TNT 'naked truth' page has zero in the way of technical detail. 'First / second order effects?' By whom? By what standard? None are properly referenced with any data at all. You are simply expected to take the author's word. Take skin effect for example, of which much is made. A perfectly ordinary piece of 10ga zip cord can be calculated & measured as having about 0.1dB loss at 20KHz through skin effect. Here's an article with a scientific basis that actually contains such calculations: Skin Effect Relevance in Speaker Cables | Audioholics To put it mildly, you'll have more variability through the drive unit that that. Assuming you can hear it at all, the likelihood of which is about zero, even for the most bat-eared teenagers.
The second link does contain a handful of basic lumped parameters, which is a start, but it goes rapidly pear-shaped by spouting a bunch of unsubstantiated claims on audibility of losses &c. Incidentally, since it sets stock by RFI etc., it does rather contradict your claims that braids etc. 'suck' since these can (can) have superior rejection of RFI than straight antennas. How significant this is depends on circumstance of course. By the way, not all speaker wires using Cat5 are either braided or twisted. Some designs using Cat5 require the individual wires to be untwisted for instance. Do these also 'suck' and if so, why?
'The best TNT project' -according to whom, and to what criteria? They're different, but without comparative data, this means nothing. Sorry. Wire gauge? Resistance? Capacitance? Inductance? And that's before you consider the rest of the system.
If you prefer it, fair enough. Not knocking that at all, to each their own. But to write off other types purely on the basis of your own preference rather than any scientific validity does nobody any favours.
Regarding the 'power audio cable' reference, I hope you mean 'speaker cable' not 'power cable.' Satellite cable is completely unfit for carrying mains currents / voltages & would be a severe fire hazard if used in such a way. To do so would be utter madness, and potentially life-threatening.
Last edited:
This is probably the only correct statement made. Neither of the articles quoted make any attempt to quantify possible effects of cable geometry ,etc. Jon Risch, bless his soul, was basically a crack pot. Unfortunately, while cable geometry, materials and isolation permeability can make huge differences at megaHertz frequencies, they are meaningless are audio frequencies.
Yes, interconnect cables CAN have pronounced effect on the sound of a system because the impedance values are relatively large. For speaker cables, not so much because of the very low impedance values involved.
For speaker cables, the LCR characteristics are the only properties that are important. As long as these values are reasonable, the cables will all sound the same. The important one is R. As long as there is less than 1dB drop in the cables, the cables are big enough. For the power levels involved with single driver speakers, a single pair of 24ga CAT5 wires is adequate. If you cross connect the whole CAT5 cable, you get ~17ga, which will handle something like 100w.
The big deal about the braided CAT5 cables is that all of that braiding will reduce L to near zero, which will improve the high frequency performance of the cable. Two problems. The improvement is above the audio band. C is driven so high the it can cause those little SET tube amps that tend to be used with braided cables to go into oscillation.
Enough! Go to WalMart and buy some 16ga zip cord. You really can't do any better.
Bob
IMO, it follows that anyone trying really hard to get that last micro-ounce of magic out of various cabling should then have no excuse for the rest of their system not being completely up to snuff, including the listening room. Just about any other performance aspect I can think of thoroughly swamps anything to be "gained" from cables. If anything, measurable and audible differences in cables will usually result from a terrible "design" or implementation and the most basic stuff, 16ga zip as Bob Brines mentions, really is as good as it gets. From there, or any other reasonable cabling, you may pretty much only introduce losses.
this seems like a useful "rule of Thumb"
multi-parallel twisting of the Flow and Return maximises the Capacitance between the Flow and Return. It also minimises the inductance.
Braiding of the Flow and Return increases the gaps between the conductors and that decreases the capacitance.
The increased gaps also increases the inductance.
It seems you have the parameter changes back to front.
Perhaps so, as I am not an EE. However, I parroted the standard reasoning for braiding -- to improve the top end.
Bob
EDIT: ad said i trust more my ears than theorical papers. According to my ears Sats "equalize" better than Cats and are *mutch* easier/faster to setup/manage.
The Shotgun is called WF100, but there are many industries that manufactures low-loss sats...
About dielectric: i've never used air ones but only foam (I don't think there's audible differences, anyway).
About impedence: I'm still trying to understand if there's sonical differences between 50 and 75 ohm cables...
The Shotgun is called WF100, but there are many industries that manufactures low-loss sats...
About dielectric: i've never used air ones but only foam (I don't think there's audible differences, anyway).
About impedence: I'm still trying to understand if there's sonical differences between 50 and 75 ohm cables...
Last edited:
I don't. I don't trust my ears, or more specifically, I don't trust my brain not to be fooled. I do trust science however. Wire conforms to the laws of physics rather than wishful thinking, and there is unfortunately a lot of the latter about.
As I asked above:
-What Cat5?
-What / which specific configuration?
-What / which specific geometry[ies]?
-Why, specifically, does it 'suck'?
Your claim that Cat5 as speaker wire has a high pass effect is perfectly accurate inasmuch as all (as in 'all') speaker wire ultimately forms a high pass filter. This is part and parcel of physical reality. The frequency at which this occurs is a function of its electrical properties. Now, there are a huge number of different speaker wire designs employing Cat5 out there, all with different geometries and therefore different (sometimes wildly different) LCR properties. These can run from multiple twisted or braided wires, to a couple of conductors pulled from a bunch, untwisted, and run parallel to each other. Are you trying to claim that all of these 'suck' and that they 'suck' for the same reason?
As I asked above:
-What Cat5?
-What / which specific configuration
-What / which specific geometry[ies]?
-Why, specifically, does it 'suck'?
Your claim that Cat5 as speaker wire has a high pass effect is perfectly accurate inasmuch as all (as in 'all') speaker wire ultimately forms a high pass filter. This is part and parcel of physical reality. The frequency at which this occurs is a function of its electrical properties. Now, there are a huge number of different speaker wire designs employing Cat5 out there, all with different geometries and therefore different (sometimes wildly different) LCR properties. These can run from multiple twisted or braided wires, to a couple of conductors pulled from a bunch, untwisted, and run parallel to each other. Are you trying to claim that all of these 'suck' and that they 'suck' for the same reason?
Last edited:
Transmission line theory can be applied also here. In Ethernet application the Cat5 cable is driven and terminated by 110 ohms in order to avoid any reflections. In audio, the source is (near) zero and the termination is a few ohms. So there are reflections from the far end. I suspect that the amount of reflections make a difference between a good and a bad cable, and it can be properly described by their physical properties. Essentially, Cat5 cable is not used according to its designed purpose (I also use rigid Cat6 as speaker cable), and it was empirically found usable for audio.
It would be even better if the article were correct. Alas, they use the exponential approximation equation instead of the correct bessels. Skin effect is about one third the effect using the exp equation would suggest. IOW, even less an issue..
Proximity isn't considered either, which would tend to making the effect a bit worse. But at the system impedances and frequencies, I'd not be worried.
Well at least he liked denali moose tracks, so he can't be all bad.
I'd worry more about ground loop topology.
As you've stated that you are not an EE, what source of information are you using to make this blanket statement?
For the same insulation, raising or lowering L or C will not alter the high or low frequency performance of any cable.
All it can do is change the characteristic impedance of the cable.
L cannot be reduced to near zero. For one twisted pair, it's about 100 nH per foot. For 4 paralled as in a cat5e cable with all solids tied and all stripes tied, it will be 25 NH per foot. It will scale as 100/# of pairs.
Capacitance climb as direct multiplication of # of pairs.
Impedance will scale as initial Z/# of pairs. 4 pairs will be 25 ohms, 8 pairs 12.5, assuming 100 ohms to start.
If you take multiple cables wired as per all solids tied/ all stripes tied technique, braiding will do absolutely nothing to the characteristic impedance, the inductance, or the capacitance of the system.
btw, I am an EE.
jn
OH, ps.. amps oscillate using low impedance cable when the speaker impedance unloads before the amp reaches open loop unity gain, that will compromise phase margin.
Last edited:
You are correct.
Given the low Z of the source and load compared to the typical twisted pair Z, the line will experience a settling time dependent on the mismatch coefficients. Values can run into the 10 to 20 uSec range, which rise into the audible range of localization. Audibility will depend heavily on how steady the load impedance can remain through the audio band. For single drivers, it can also be dependent on the eddy dissipation parameters and how badly they are impacted by velocity based flux dragging through the pole tip and outside pole ring.
jn
You are attempting to mix rf based engineering with audio. You need a plan B.
What is of concern is the settling time of the system. That is over 3 orders of magnitude longer than the single pass prop velocity number.
The "size" of the wire has no significance within the context.
If only speaker drivers maintained a constant impedance across the band. If it did, the settling time would be frequency independent, and therefore of no consequence to soundstage imaging.
As it is, speakers vary their impedance quite a bit across the audio band. Single drivers are far worse in that they also drag high frequency energized voice coils across the pole gap surfaces, resulting in a modulation of the impedance based on the velocity. edit:remember, the terminal voltage of an inductor is L dI/dt + I dL/dt. I have NEVER seen a reasonable paper out there discussing the second term, most do not even understand that it exists.
The use of "wavelength" in the context of the discussion is of no use. We are not discussing single pass energy at the characteristic impedance of the cable, but energy which must match the source voltage and the load current requirement. It can take many hundreds to a thousand passes for the load to achieve the correct current, this based on the tightness of the source impedance and the mismatch of the speaker.
edit: I can never get this linky stuff right. Here's one graph from my diy gallery...note how the settling time is load dependent.
jn
Last edited:
Couldn't tell ya. To minimize settling time issues, the cable should be near the median impedance, perhaps 16 to 20 ohms.
You might be hearing the consequences of insufficient gauge.
jn
In fairness to Audioholics they regularly point out re skin effect that the simple lumped parameter approach there gives pessimistic results -albeit not in that particular page, so fair point. Not that it matters much since however you wish to calculate it, it's a non-issue at audio frequencies. Rather like electrical TL theory (no issue in the high KHz / MHz regions, obviously).
Re 'soundstage smearing', since it has no technical meaning, chalk up another pass. Assuming any effects exist, it could be any one of a number of factors. Incidentally, 'all versions' is a bit excessive given the number of different things that can be done with Cat5, up to and including separating out & untwisting the conductors.
Either way, I think we may be forgetting something here. It's wire. There is more to life. You want to worry about something, how about the spread of Ebola?
Re 'soundstage smearing', since it has no technical meaning, chalk up another pass. Assuming any effects exist, it could be any one of a number of factors. Incidentally, 'all versions' is a bit excessive given the number of different things that can be done with Cat5, up to and including separating out & untwisting the conductors.
Either way, I think we may be forgetting something here. It's wire. There is more to life. You want to worry about something, how about the spread of Ebola?
Last edited:
Actually, the graph I provided is a direct consequence of T line theory. I made it using the simple approximation of a T-line with no conductor resistance nor insulation conductance (R and G in the tel..equation).
What is of note with using just L and C, is that disregarding R results in a graph which settles FASTER than reality. So reality is actually a bit worse than my graph would indicate.
BTW, this graph models very closely to the simple lumped element model, but to me it provides a much clearer understanding of the issue.
I understand the meaning of "soundstage smearing" in technical terms. To me, it indicates that the parameterics my brain uses to define the location of a virtual image is being compromised in a time varying fashion as a consequence of some other system response...
One example I would use is program content with a vocal image centered, where a bass note on one channel causes large cone excursion on that channel only. If the sibilance (hf content)is altered in that channel as a consequence of the voice coil velocity changing the hf inductance and series resistance, the resultant single channel modulation may be heard by a listener poised in the sweet spot as the sibilance either drifting off center (disembodied as it were), or simply becoming less location defined.
But I do agree that the terminology in this regard is pretty lacking.
jn
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.