You are correct on all counts.
However, as I mentioned many times before, it is NOT frequency response I speak of. It is timing, or better said, delay.
We all know that an inductor in series will cause a phase shift at the load. We also know that changing the load will also affect the phase shift.
A speaker not only has an impedance variation with frequency, it has an impedance that varies with acceleration. In other words, when a driver is actively pushing say 50 hz hard, it's impedance at 1khz will be modulated.
That cannot be seen using a swept sine impedance test. I've struggled with how to actually measure that despite all the confounders.
If the cable/load is matched, incremental impedance variations will only shift timing at the level of the cable prop delay, roughly 5 nSec per foot. Recall, when the cable RFZ matches the load, energy storage in the cable is a minima.
The timing when the cable stores lots of energy, delays can rise to the ITD sensitivity levels. My 4 pair/25 ohm RFZ recommendation puts the cable in the middle of a typical speaker, so should remove ITD concerns.
As I said, the motion control industry is just now beginning to catch the hints that a magnetic drive be it motor or voice coil, changes it's frequency response based on the forces against the magnetic field. One seminar speaker I just zoomed thought it was the PWM amplifier that was causing the bode plot shift, eventually he will realize it's the magnetics. The best part is he worked out the software to measure the dynamic frequency response.
In my work I control devices that have magnetic forces in the tens of thousands of lbs force variation. When that force changes, the bode plot changes wildly. At present, we only tune the PID so that the lowest force point has sufficient phase margin, we live with the position error at high force locations. Eventually, it will be possible to adaptively change PID based on position.
Unfortunately, when we fixed the motion issues such that errors are below the 1 micron requirement (currently 5 nanometers static and 25 nanometers dynamic), they won't let us play with the devices anymore.
Jn
Last edited:
I too found it an interesting read. I was ignoring speaker cables as I felt that the impact is fairly small and I had good enough diy cables. This article prompted me to swap in some other cables I had made that would more closely emulate the impedance matching the paper advocates. In my case it made a significant difference in clarity, imaging and in bass extension. The bass extension was easy to hear and to measure BTW, so please no double blind flames.
In my case I should have known my cables had been causing problems. I'm the first to admit that if I had done some blind structured listening sessions with multiple people I would have realized the issue. But the system had sounded pretty good. I always got rave reviews from visitors. And I was convinced that the speaker cable at most makes a small difference. I was wrong.
While the article is really a white paper designed to advocate the value proposition of the commercial cable, the info was useful and certainly prompted me to make a significant improvement.
The issue with my original diy cables was the conductors were separated by an inch. I had built kind of a ribbon cable with separated conductors based on success with that approach on interconnects. The premise of the article is that he had measured that cable architecture as having a high impedance and he measured significant ringing on transients as a result. His cable is kind of a capacitor with two foil conductors separated by a thin dielectric. He measures a low impedance and no ringing on a standard 8ohm speaker. Some day I'll get an impedance measuring device but don't have one on hand.
On the Bode-100, the Reference Channel is Channel 1, the Measurement Channel is 2. The HP3577a has "Source", "Ref", "A" and "B". If you are going to use the "Gain" function it's just like 4-wire resistance measurement. The Reference channel is picking up the signal at the load.
standardize one time, even use the usual 2 or 3 meter RG59/U cables and run the calibration routine, "open, short, load". Then replace the measurement channel cable with the DUT.
The cable adverts describe the decreased susceptibility to EMI and RFI. I have a good mind to put one of my wife's hair dryers next to a cable and see how it performs.
Most standard texts have this equation, speak for yourself on who has forgotten it. All the TL simulations I have seen (including the ones we posted here) include R and G. At tiny fractions of a wavelength this has little or no relevance to the answer.
Here we go:
Hence my question above:
Which, of course, will never get a direct answer, since it would be way too damaging to the cloud of dust around this topic. Did I mention using the cable itself to compensate for any weird load properties/nonlinearities is a dumb idea?
Hence my question above:
Which, of course, will never get a direct answer, since it would be way too damaging to the cloud of dust around this topic. Did I mention using the cable itself to compensate for any weird load properties/nonlinearities is a dumb idea?
How many amp destruction cases do you know of from using the standard zip-cord for speaker cables in home audio setting?
What audible effects did that problem create?
What is blind structured listening?
From your previous description of the geometry, would not Cat 5/6 cable closely match your desciption using 24 guage wire?
dave
Yes, and because the 4 pair are twisted differently, no need to even remove the jacket. But the gauge is too small.
jn
Here is a link to a video where the designer of a cable explains cable design. If you don't want to watch the whole video, look at 6:40 into it, and again at 19:05 where he takes questions. The guy asking the first question of why we can't measure something is a friend of mine. This video is from the Rocky Mountain Audio Fest.
RMAF15: The Cable Polygraph - YouTube
RMAF15: The Cable Polygraph - YouTube
Did you not say that #24 was sufficient for home?
As to too small i typically use just a pair of strands of CAT5.
dave
Like your ear/brain combo? Far more analytical then measurement devices according to Toole.
You have a very good memory.
When I had a small stereo in my living room, as well as the 5.1 system up in the attic, I did use the bog standard #24 zip that came with it.
But if my intent were to critically listen to a good system, I would want to go higher gauge for the bass. Since my living room has shelves with vases and pictures and stuff, I could not play a system the way I would wish. Too many things to fall to the floor.
If you can read newsprint while listening to music, you are lacking bass.
jn
The first minute or so of that video is all I need to hear! He has me convinced, though I admit I've never heard the difference between two feet and 50 feet of 12 gauge zip cord. For 100 feet I think I heard a slight drop in volume, but it was a sighted test and I couldn't be sure.
I found it amusing that at least one high-end speaker maker mentions the brand name (not the name of an industrial wire manufacturer such as Belden, but some "hifi" speaker cable brand name) of the wire used between the crossover and drivers.
But yes, once again, this is a "hobby" and customers want to mix-and-match components (and interconnects are "components") to build what they believe is the best system they can afford.
Indeed, this is the usual case for pro audio where powered speakers are common.
I found it amusing that at least one high-end speaker maker mentions the brand name (not the name of an industrial wire manufacturer such as Belden, but some "hifi" speaker cable brand name) of the wire used between the crossover and drivers.
But yes, once again, this is a "hobby" and customers want to mix-and-match components (and interconnects are "components") to build what they believe is the best system they can afford.
- Status
- Not open for further replies.