what is the best speaker cable to bring low THD to the voice-coil?
By building a low THD Amplifier and having low THD speakers this question came to my mind.
I do not want to start some voodoo threat - I want to know what makes sense from the technical site.
There is a lot of high frequency around us and i thought a shielding might be a good thing to prevent the high frequency creeping into the amplifier and disturbing the feedback loop with the result of a bad sound. So i bought some shielded 2x0,75mm² cable, connected the 2x0,75mm² as usual and the shielding to the amplifiers ground (a: directly; b:via resistor). It showed up, that in both cases the sound gets worser than before. A look at the oscilloscopes screen showed the reason: the shielding worked as an antenna and did the opposite of what it was intended to do.
Looking around in several forum showed only subjectivists views, none of them showed any technical description...therefore some questions:
- is shielding necessary / useful? If yes - how should it be done? If no - why not?
- is a ferrit clipped on the cable helpful? (as it is dampening common mode signals that reached the cable)
- does shielding help, even when it is nowhere connected? (as it is good practice in KNX home automation cabling)
- what is the optimal diameter for the cable? (with respect to the skin-effect)
- is a twisted pair better than a parallel cable? (as signals that reach the cable are inducted in opposite direction and therefore cancel)
- is the star quad configuration better than twisted pair? (as it promises a better immunity to magnetic fields)
- is coaxial-cable the solution? ( as it is shielded and, referring to cable-theory, all the inductive and capacitive fields from the current that is running through the cable happen inside, and the shielding works as such when connected to ground)
- what is the best way to test and find out?
By building a low THD Amplifier and having low THD speakers this question came to my mind.
I do not want to start some voodoo threat - I want to know what makes sense from the technical site.
There is a lot of high frequency around us and i thought a shielding might be a good thing to prevent the high frequency creeping into the amplifier and disturbing the feedback loop with the result of a bad sound. So i bought some shielded 2x0,75mm² cable, connected the 2x0,75mm² as usual and the shielding to the amplifiers ground (a: directly; b:via resistor). It showed up, that in both cases the sound gets worser than before. A look at the oscilloscopes screen showed the reason: the shielding worked as an antenna and did the opposite of what it was intended to do.
Looking around in several forum showed only subjectivists views, none of them showed any technical description...therefore some questions:
- is shielding necessary / useful? If yes - how should it be done? If no - why not?
- is a ferrit clipped on the cable helpful? (as it is dampening common mode signals that reached the cable)
- does shielding help, even when it is nowhere connected? (as it is good practice in KNX home automation cabling)
- what is the optimal diameter for the cable? (with respect to the skin-effect)
- is a twisted pair better than a parallel cable? (as signals that reach the cable are inducted in opposite direction and therefore cancel)
- is the star quad configuration better than twisted pair? (as it promises a better immunity to magnetic fields)
- is coaxial-cable the solution? ( as it is shielded and, referring to cable-theory, all the inductive and capacitive fields from the current that is running through the cable happen inside, and the shielding works as such when connected to ground)
- what is the best way to test and find out?
Well probably your amp is oscillating because of the high additional capacitance of the shielding (100pF/m) not from external induction.
Also you have to connect the inner conductors together and have the return current through the shield only to function properly, otherwise there is no shielding effect over a normal dual conductor.
BTW your idea of the creeping electromagnetic fields is not correct.
Just hold your cellphone next to the cable, you will not find any stronger HF than that. Well nothing is happening; right?
Also you have to connect the inner conductors together and have the return current through the shield only to function properly, otherwise there is no shielding effect over a normal dual conductor.
BTW your idea of the creeping electromagnetic fields is not correct.
Just hold your cellphone next to the cable, you will not find any stronger HF than that. Well nothing is happening; right?
I conducted some speaker cable FFT investigations between amplifier terminals and speaker terminals with John Dunlavy (Duntech, then Dunlavy Audio Labs) in the 1980s with a few different speaker cable constructions. I wouldn't say that what we found was the everything of cables, but there is a good reason to use star-quad speaker cables IMHO.
Coaxial cables in theory have zero inductance and can drain EMI/RFI to ground in a well designed amplifier, but the only coaxial speaker cables I have tried (Ocos, EuroCable and RG-11) somehow fall short in sound quality compared to star-quad configurations. I do use Eurocable 02N25C coax for Bang & Olufsen gear when warranted because it has 2.5mm² of copper per conductor within a 6mm diameter cable that will fit into a 2-pin DIN plug! Ocos was used by Dynaudio some years ago, but because Ocos was available in only one size, to get low series R needed several cables in parallel. Although it has some fans RG-11 is far from ideal as a speaker cable IMHO, not the least being flexibility and sound.
Mechanical integrity is an important property for a speaker cable, because movement of the conductors induced by the magnetic forces between them from signal currents change the sound, just as microphonics in interconnects affect low level signals.
Star-quad's construction advantage in terms of low EMI/RFI emissions and susceptibility do not depend on the amplifier configuration, i.e. single ended, bridged and balanced/transformer coupled all benefit from star-quad, unlike coaxial which only has a benefit for single ended amplifiers with one side grounded.
My preferences for well constructed cables are Canare 4S8 or 4S11 for modest cable lengths or EuroCable 04N40 for longer cables or larger systems.
Coaxial cables in theory have zero inductance and can drain EMI/RFI to ground in a well designed amplifier, but the only coaxial speaker cables I have tried (Ocos, EuroCable and RG-11) somehow fall short in sound quality compared to star-quad configurations. I do use Eurocable 02N25C coax for Bang & Olufsen gear when warranted because it has 2.5mm² of copper per conductor within a 6mm diameter cable that will fit into a 2-pin DIN plug! Ocos was used by Dynaudio some years ago, but because Ocos was available in only one size, to get low series R needed several cables in parallel. Although it has some fans RG-11 is far from ideal as a speaker cable IMHO, not the least being flexibility and sound.
Mechanical integrity is an important property for a speaker cable, because movement of the conductors induced by the magnetic forces between them from signal currents change the sound, just as microphonics in interconnects affect low level signals.
Star-quad's construction advantage in terms of low EMI/RFI emissions and susceptibility do not depend on the amplifier configuration, i.e. single ended, bridged and balanced/transformer coupled all benefit from star-quad, unlike coaxial which only has a benefit for single ended amplifiers with one side grounded.
My preferences for well constructed cables are Canare 4S8 or 4S11 for modest cable lengths or EuroCable 04N40 for longer cables or larger systems.
Last edited:
Thread moved.If you want to avoid voodoo, all you need is a good quality copper wire of sufficient gauge. EMI nonsense isn’t the cable’s fault - the levels are low enough so that induced signals are not going to affect anything audibly. In the cases where it does, the problem is usually in the amplifier if you’ve built the cable correctly and it should be attacked there.
If you do want to minimize the distortion produced by the cables themselves, pay more attention to the terminations than the cable itself. Any place where two dissimilar metals contact is a potential problem - if it is partially rectifying, there is distortion. But these problems can always be fixed with proper attention to detail - proper soldering or crimping must be used (and crimps can range from better than solder to worse than a loose wire nut). This is also why gold (or tin, or phosphor bronze) connections are used on pressure contacts - so that the work function does not support rectification, and they tend not to oxidize. And any oxidation at the point if contact produces rectification, even if the metals themselves were fine. If you do have a rectifying contact somewhere, you can pick up RF, mix it down to baseband and have audible effects. But it takes a STRONG signal, like living in the shadow of a TV transmitter or server farm in the next room to do it, and misbehaving amplifiers can do it with far lower levels. Amplifiers designed for “very low distortion” often have either extremely high open loop gain or a lot of bandwidth. This makes RF even more of a problem.
Dunno. I always have used RG/58U for speaker cables, one for each polarity. Why? Because they're fat, well insulated, braided and highly durable. When hand counter-twisted and heat-shrink sleeved, they naturally counter induced currents, and are even more durable. Visually presentable, too.
And inexpensive. Except for Belden brand. Which isn't inexpensive. Ever.
Just sayin',
GoatGuy
And inexpensive. Except for Belden brand. Which isn't inexpensive. Ever.
Just sayin',
GoatGuy
First comment: THD has been shown to be a pretty much useless number (all the important information is lost whem you collapse the spectrum) in amplifiers and Geddes, at least, has shown it to be pretty seless for loudspeakers as well.
2: Iif you want to take the speaker cable out of the equation build a currrent amplifier (althou now your speaker impedace could be a problem).
dave
2: Iif you want to take the speaker cable out of the equation build a currrent amplifier (althou now your speaker impedace could be a problem).
dave
Get real. Ampere's force law - Wikipedia
100watts into 8ohms with cable that has 5mm conductor spacing = .0006 Newtons/meter. The equivalent of .06 grams of force per meter will not move your conductors, and even if it did slightly changing the characteristic impedance (it will not change resistance) wouldnt matter anyway.
And cable microphonics is more BS. Show me the measurements, or just tap the cable with your finger (1000 times the force of sound), hear anything? Didnt think so.
100watts into 8ohms with cable that has 5mm conductor spacing = .0006 Newtons/meter. The equivalent of .06 grams of force per meter will not move your conductors, and even if it did slightly changing the characteristic impedance (it will not change resistance) wouldnt matter anyway.
And cable microphonics is more BS. Show me the measurements, or just tap the cable with your finger (1000 times the force of sound), hear anything? Didnt think so.
The Sound Lab electrostat panels drop to as low as 2-ohms and are not terribly efficient. The Aragon 8008 is 200w/channel and runs at pretty how power for normal listening levels. There can be more current than one might estimate. Also, the elctrostat panels are more revealing than box speakers. Small differences in speaker cable construction can be audible.
The inductance of a coaxial speaker cable is several orders of magnitude lower than the inductance of an equivalent length of a pair of conductors. The few nanohenries of inductance in a few meters of coaxial cable used between an amplifier and a loudspeaker in a hifi system is of no practical consequence and effectively zero.
This is totally wrong, where did you get that? Belden RG11 is 100nH per foot and 16.5 pF per foot.
You can compute cable characteristics here. Impedance Calculators - Mantaro Networks Inc. Product Development Services
You can compute cable characteristics here. Impedance Calculators - Mantaro Networks Inc. Product Development Services