What we were taught in school about t-lines and reflections did indeed discount the possibility of reflections in the audio band for short cables. Unfortunately, that was based on an approximation, one used to simplify the engineer's job. Like the skin effect approximation of the exponential equation, where the depth calculation is good enough as long as you remain within the limits where the approximation is accurate enough. T-lines are the same.
The actual effect short line reflections will have on 1Khz signals for example, is extremely small. 5, 10, even 20 uSec delays on a 1Khz sine is so small that it is ignored in standard work. If I were running a few kilowatts at 1Khz into some load and worrying about delays and losses, I would also ignore ten uSec as well.
Unfortunately, humans have this absurdly powerful capability to discern direction of a sound source at the 1.5 uSec and up level. This level of interchannel time discernment is where the standard engineering techniques begin to fall apart.
The complexity of virtual image localization is sufficiently high, that I always recommend any technical person who wished to consider or argue cables first learn a bit about what humans can hear, as that is really the end result wished..what is audible.
For a low impedance amp feeding a 150 ohm speaker wire to a pure resistive load, the actual current delay at the load will be a function of the line to load match. If load = line, the delay will be exactly the propagation delay, measured in nanoseconds. If the load is very low or very high with respect to the line, it will take a large number of reflections and transits until the load current has settled to 90-95% of the expected value.
Since speakers can vary wildly across the audio band, I would recommend trying to get near the center of the range. It's reasonable to run wires at 25 ohms for example, by using perhaps 5 or 6 independently twisted #18 or #20 awg zips. No specific braiding of rancy stuff, just twiste them independently but make sure polarity is correct at each end.
If you try to go at or below 4 or 8 ohms, you will really have a lot of capacitance in the cable makeup. It has to be noted that capacitance is NOT an issue for any amplifier on the planet as long as the load at the far end matches the cable's characteristic impedance.. But if the load impedance climbs up with frequency, then the amplifier will see significant capacitive storage and may oscillate if the unloading occurs below the open loop unity gain point of the amplifier. The use of a zobel at the far end can prevent the unloading which is the problem.. If your amplifier is marginally stable with a low z cable and high z load, a cable made with a built in zobel will indeed be highly directional.
I do not recommend making the cable lower than the load especially if the load is 4 or 8. My recommendation is to make the cable somewhere in the middle of the load impedance min/max.
ps. are all posts on this site moderator approved, or is this a trial period for bad eggs like me?
jn
The actual effect short line reflections will have on 1Khz signals for example, is extremely small. 5, 10, even 20 uSec delays on a 1Khz sine is so small that it is ignored in standard work. If I were running a few kilowatts at 1Khz into some load and worrying about delays and losses, I would also ignore ten uSec as well.
Unfortunately, humans have this absurdly powerful capability to discern direction of a sound source at the 1.5 uSec and up level. This level of interchannel time discernment is where the standard engineering techniques begin to fall apart.
The complexity of virtual image localization is sufficiently high, that I always recommend any technical person who wished to consider or argue cables first learn a bit about what humans can hear, as that is really the end result wished..what is audible.
For a low impedance amp feeding a 150 ohm speaker wire to a pure resistive load, the actual current delay at the load will be a function of the line to load match. If load = line, the delay will be exactly the propagation delay, measured in nanoseconds. If the load is very low or very high with respect to the line, it will take a large number of reflections and transits until the load current has settled to 90-95% of the expected value.
Since speakers can vary wildly across the audio band, I would recommend trying to get near the center of the range. It's reasonable to run wires at 25 ohms for example, by using perhaps 5 or 6 independently twisted #18 or #20 awg zips. No specific braiding of rancy stuff, just twiste them independently but make sure polarity is correct at each end.
If you try to go at or below 4 or 8 ohms, you will really have a lot of capacitance in the cable makeup. It has to be noted that capacitance is NOT an issue for any amplifier on the planet as long as the load at the far end matches the cable's characteristic impedance.. But if the load impedance climbs up with frequency, then the amplifier will see significant capacitive storage and may oscillate if the unloading occurs below the open loop unity gain point of the amplifier. The use of a zobel at the far end can prevent the unloading which is the problem.. If your amplifier is marginally stable with a low z cable and high z load, a cable made with a built in zobel will indeed be highly directional.
I do not recommend making the cable lower than the load especially if the load is 4 or 8. My recommendation is to make the cable somewhere in the middle of the load impedance min/max.
ps. are all posts on this site moderator approved, or is this a trial period for bad eggs like me?
jn