I don't believe cables make a difference, any input?

[Passion]

Originally Posted by SY

Could you please explain what that chart is? What does it measure and what was the basic test setup?

The chart shows the magnitude of losses in interconnect cables, relative to the input signal, in dB.

The instrument consists of a "silver black box" where the cable under test is connected, fed by the desired signal (chirp, pink noise, music, whatever). At the output of the black box you have the losses for the cable under test, which you can listen to, or measure in the time or frequency domain.

In the attached picture you can see the 2nd prototype, where I was measuring a low-cost short cable and a long expensive one (it's a 2-channel instrument, so two different cables can be easily compared).

I hope I answered your questions...

R

[SY]

R, many thanks. Can you tell me the driving impedance and terminating impedance? Voltage level for the test?

[cbdb]

For more info on this, and a critique of the Hawksford article, see this issue of The Audio Critic (PDF) (somewhat humorously referred to as "The Essex Yecch-o")

Great stuff. Particularly this statement; "Similarly, anything that takes place inside the conductor such as boundries, grains, and all that junk are irrelevant, since any energy that enters the surface of the conductor is lost as heat anyway. ... it never again contributes to the information travelling along the axis of the conductor"

This will get a lot of use in this thread!

Thaks again Andy, the voice of reason and storehouse of information. (how do you pull out these references and articles?)

If you ever come to Vancouver I want to buy you a beer!

[Passion]

Originally Posted by cbdb

Exactly. I think your experiment is exageratig the effects. And then misrepresenting them. You make it sound like one cable is down 20DB at 20k compared to another, which is rubish. What your saying is people can hear the diference in less than .1 db of attenuation between 2 cable while te F response is almost identical. Im having trouble buying that.

The experiment is showing the actual magnitude of losses for all cables, and there is no exaggeration of any effects. Believe or not, losses in a high-quality cable can be 20dB lower than those in a low-cost one.

Here I’m showing how different cables can be, which is the point of my post.

Cheers,

R

[SY]

With source and driving impedances, we can back into the DCR and equivalent inductance, then figure out what the actual frequency response variations at the termination will be. It won't be 20dB and I'll be surprised if it's 0.2dB.

[texaslonghorn]

Hi Passion. Thank you very much for your efforts here. I personally believe that ICs/Cabling do make a difference and have for several years based on my own personal listening experience. But I am curious - your chart appears to indicate a 20dB delta between cables across the frequency spectrum??? That seems pretty massive to me and not something that I have personally experienced. Can you help me understand more about how the 20dB could be for real? Again, keep in mind I am NOT a skeptic and I am very supportive of your work in this regard. I just want to make sure I am understanding what the data seems to be showing. Thank you very much.

[andy_c]

Originally Posted by cbdb

Great stuff. Particularly this statement; "Similarly, anything that takes place inside the conductor such as boundries, grains, and all that junk are irrelevant, since any energy that enters the surface of the conductor is lost as heat anyway. ... it never again contributes to the information travelling along the axis of the conductor"

This will get a lot of use in this thread!

Yeah, that's a great explanation. Maxwell's equations only deal with macroscopic stuff. Typical manufacturers' claims about the microscopic things are just obfuscation.

As far as searching, I love the Google advanced search. I just did a search for "Essex" restricted to theaudiocritic.com. I couldn't remember how "yecch-o" was spelled .

[SY]

Can you help me understand more about how the 20dB could be for real?

Remember, dB is always a relative measurement. If wire 1 has twice the series resistance as wire 2, one can say that wire 1 will drop 6dB more than wire 2, even if the actual delivered signal level variation is negligibly small (e.g., 0.01dB versus 0.2dB).

[Passion]

Originally Posted by SY

R, many thanks. Can you tell me the driving impedance and terminating impedance? Voltage level for the test?

The source impedance is the output impedance of a typical opamp, so it's "low". Keep in mind that I'm measuring the cable losses relative to the input voltage, so any frequency response anomalies due to source impedance are eliminated.

Termination impedance is 10K for now, purely resistive, but it can be "anything" (within reasonable practical limits). R//C loads are not a problem.

Typical source voltage is 2V RMS, but can be lower if the noise floor in your test environment is not too high.

Cheers,

R

[cbdb]

If passion re-posts the graphs showing actual voltage levels I think it would be much clearer. (and plot the Vin Vout ) This will show that Vout will be Vin(1 voltish) +/- micro volts? for different cables.

Another way to put it; The relative levels show the resistance of the cable. If 1 is .ooo1 ohms and the other is .01 ohms thers 20 db difference but compared to a 10k input even 1 ohm or 40db will be inaudible.