T
Terry Given
Guest
Phil Allison wrote:
inductance for spacing s and conductor diameter d is:
L = 0.4uH*acosh(s/d)
my 2.5mm^2 speaker cable has s=4mm, D = 1.8mm
so L = 0.575uH per metre
I had 10m of the stuff, so 6uH. Thats 0.4 Ohms at 10kHz, which is quite
a bit compared with my 4 Ohm speakers, so I stuck them in series and
ignored it.
when I first connected the speakers in series, in a fit of inspired
stupidity I stuck each speaker as far away as possible from its series
partner. The resultant loop in the cable was huge, and much more than
doubled the inductance.
right down. the IDC connections can be troublesome, but a little adaptor
pcb would be trivial
A simple, cheap approach would be a copper stripline. magnetics vendors
often stock copper strip of various thicknesses, and sometimes wrapped
with quite good insulators.
two long strips of 1mm copper, perhaps w = 25mm wide, wrapped in
heatshrink and glued together so separation s = 1mm.
inductance L = u0*s/w = 0.4*pi*1/25 = *50nH* per metre
the DC resistance is about 2*20n[Ohm-m]/25u[m^2] = 1.6mOhm/m
the strip is one skin depth thick at 1mm = 66mm/sqrt(f) so f = 4.4kHz.
For 5 skin depths, 0.2mm = 66mm/sqrt(f) so f = 108kHz. At 20kHz its 2.1
skin depths thick. So the AC resistance Fr is close to 1 over the entire
audio band.
10mm strip will be about 125nH and 4mOhm per metre.
the problem with cutting the strip by hand is jagged edges. and it wont
bend so nice.
Cheers
Terry
for a 2-wire parallel transmission line, the exact formula for external"Trevor Wilson"
"robert casey"
**Really? Where did you study electrical engineering? I hope you didn't,
because you can be shown to be wrong, very easily.
** Mr Casey is an expert on work creation - following him around usenet
correcting all his asinine postings is almost a full time job.
Oh, if you had runs several kilometers
long you'd start to care about inductance and capacitance,
but nobody's house is that big.
**Several km? Are you certain about that? So-called 'zip' cable has an
inductance of approximately 0.75uH/Metre. At 1kM, with 8 Ohm speakers, the
attenuation will be around 3dB at around 1.5kHz. At 100Metres, the
attenuation (with 'normal' 8 Ohm speakers) will occur within the audible
range. When dealing with VERY difficult speakers:
inductance for spacing s and conductor diameter d is:
L = 0.4uH*acosh(s/d)
my 2.5mm^2 speaker cable has s=4mm, D = 1.8mm
so L = 0.575uH per metre
I had 10m of the stuff, so 6uH. Thats 0.4 Ohms at 10kHz, which is quite
a bit compared with my 4 Ohm speakers, so I stuck them in series and
ignored it.
absolutely.** Even a 5 metre length of twin wire ( gauge is irrelevant) has enough
linear inductance to become quite audible when the load impedance drops to 2
ohms near 17 or 18 kHz - as it does with the Quad ESL57. With "stacked"
ESL57s, impedance drops to 1 ohm making even as little as 3 metres of heavy
gauge twin lead audible.
BTW I am not referring to some subtle effect that takes practice to
ear - but a very audible loss of signal that amounts to 1 to 3 dB @
17 - 18 kHz !!!
when I first connected the speakers in series, in a fit of inspired
stupidity I stuck each speaker as far away as possible from its series
partner. The resultant loop in the cable was huge, and much more than
doubled the inductance.
Bob Pease suggested 64-way ribbon cable, interdigitated. That gets the LAlso, when viewed on a scope, there is serious ringing of a square wave at
audible frequencies ( 10 to 25 % amplitude - depending on cable length )
at the speaker end of the cable that it NOT there at the amp input end.
The ONLY solution I know to cure this is to use a low inductance,
inter-woven cable like Tocord.
right down. the IDC connections can be troublesome, but a little adaptor
pcb would be trivial
A simple, cheap approach would be a copper stripline. magnetics vendors
often stock copper strip of various thicknesses, and sometimes wrapped
with quite good insulators.
two long strips of 1mm copper, perhaps w = 25mm wide, wrapped in
heatshrink and glued together so separation s = 1mm.
inductance L = u0*s/w = 0.4*pi*1/25 = *50nH* per metre
the DC resistance is about 2*20n[Ohm-m]/25u[m^2] = 1.6mOhm/m
the strip is one skin depth thick at 1mm = 66mm/sqrt(f) so f = 4.4kHz.
For 5 skin depths, 0.2mm = 66mm/sqrt(f) so f = 108kHz. At 20kHz its 2.1
skin depths thick. So the AC resistance Fr is close to 1 over the entire
audio band.
10mm strip will be about 125nH and 4mOhm per metre.
the problem with cutting the strip by hand is jagged edges. and it wont
bend so nice.
Using 5 metre long Tocord leads, with single ESL57s, high frequency loss is
reduced to under 0.5 dB ( resistance loss alone) and ringing all but
disappears.
........... Phil
Cheers
Terry