Real Ohms or make-belive Ohms?
I am positive you don't know how or what are you measuring...
It would be far easier to show us what Soongsc did and a schematic with a description of the method.
Then he could ask "do these measurements tell us anything?"
Then we could make some sense out of the wavy lines in the pics !
Then he could ask "do these measurements tell us anything?"
Then we could make some sense out of the wavy lines in the pics !
I don't have capability to measure to the frequency range the data are published on, so I really can't verify. What do you think RGU data books publish?
Even better. You do some measurements and tell us what it is?
Last edited:
I KNOW that no way a cable in audio range can have 120 ohms IMPEDANCE.
To me looks like you measured the impedance of the input device since it looks like you did those "measurements" with the cables connected to some device. I cannot imagine thou what input device would have 120-32 ohms (except the headphones) that's why I did say that maybe you don't know how to conduct some basic measurements.
To me looks like you measured the impedance of the input device since it looks like you did those "measurements" with the cables connected to some device. I cannot imagine thou what input device would have 120-32 ohms (except the headphones) that's why I did say that maybe you don't know how to conduct some basic measurements.
For a typical cable in the audio range the "characteristic impedance" can be calculated for any particular frequency, but it is so frequency-dependent that for audio purposes it is almost meaningless. Even without this complication, you don't measure characteristic impedance by mismatching a cable at one end, then measuring the impedance seen at the other end. There is more to measurement than hooking things up, and taking a reading. You have to understand what you are doing in order to correctly interpret the raw measurements. This is where many DIYers, and some professionals, come unstuck.
The HF rise you saw is probably caused by the input impedance of the far end, whatever that is.
The HF rise you saw is probably caused by the input impedance of the far end, whatever that is.
That's what I said, connected to a device. But I think you are mixing what I said together and reorganizing it in the wrong order in your mind.
Last edited:
The issue is that once you connect an interconnect to a device, it's part of the load the feeding device sees. So we cannot just look at the characteristic of interconnect alone, but rather consider it as part of the interface. As we get the overall impedance to flatten out, the sonic characteristics changes. But even with similar impedance shape, the sound can be quite different which perhaps is due to other characteristics of the interconnect structure.
Last edited:
This statement confuses me. It is a trivial thing to design and fabricate an audio cable with a characteristic impedance spanning anywhere from half an ohm out to thousands..
Cheers, John
Well, you hit the problem there- soongc hasn't even given us enough information to see if he's measuring (or trying to) characteristic impedance or impedance, much less any of the basic information (like scale!) that anyone would need to move past, "They're wavy lines, all right."
So let's see how long this thread can last before it has to be shut down by moderators.😀 I'll just work on other things until someone either has their own data to compare against mine, or until there are some other ideas that inspire more discussion. Look in the sky, it's a bird, it's a plane, it's...😱
John, how difficult is it to design a cable for audio range with a constant impedance within 5%? without creating too much load on the source?
This has two components.
1. It is trivial to design a cable to remain with a deviation of impedance in the audio band less than 5%. Manufacturing it will be more difficult, as it requires good control of the geometry.
2. The load on the source is a different beast altogether.
If the load at the receiver is the same characteristic impedance as the cable, then the source will see pure resistance at the characteristic impedance.
Any load impedance different from the cable's impedance will create an effective load on the source that will not be pure resistive. A load lower than the cable results in a net load consistent with an inductor, and a load greater than the cable will result in a net load consistent with a capacitor.
It is the mismatch that the source must be able to contend with.
Given the propensity to design input impedances in the 10k range or so, it is generally impossible to match it with a cable. A coax would require too large a dielectric thickness with a really small inner conductor, and a twisted pair would require large spacing which provides gobs of loop area for trapping externally generated flux (hum and noise).
This doesn't even consider ground loop issues...
Cheers, John
Quite difficult. That is because through the audio range the cable shifts from having its characteristic impedance set by conductor resistance and insulator leakage, to inductance and capacitance. An extra complication arises if the screen is grounded at both ends - the signal return current has a choice between the screen and the alternative ground path (RF signals use the screen, because of the strong coupling between inner and outer - not true at audio frequencies); a 'coax cable' might not behave coaxially at audio freqs.
Take a '75 ohm' cable. At audio frequencies it is not 75 ohm, it may not even be resistive.
Take a '75 ohm' cable. At audio frequencies it is not 75 ohm, it may not even be resistive.
At audio frequencies, it's a simple RC as described above. That's one reason I can't make heads nor tails of those wavy lines. Is this a low source impedance and a high load impedance as is standard for nearly every line level interconnection in home audio? Is this even line level? Is this a measured impedance or a derived characteristic impedance, and if the latter, how was it derived?
Actually, no.
In the audio range, it is enough to measure the L per unit foot and the C per unit foot. Only when the cable length is sufficient that dielectric losses or IR effects is it necessary to even consider those elements.
An exceedingly accurate set of sentences... In the world of consumer audio, attention to EMC is scattered if present at all. Even in the pro world, many aspects are neglected or glossed over.
Your "coax cable might not behave coaxially" appears to be inconsistent with what is generally believed with respect to coax operation, but it is indeed very true. Especially single ended operation where code requires chassis safety ground.
I've posted the actual system model with the shield current frequency breakpoints on various forums over the years..
That of course depends on how "75" was specified and measured. If the vendor considered the core path as being a cylindrical shell as a result of skinning at high frequencies, then as penetration occurs at lower frequencies, the internal inductance of the cylider of wire will recover..this of course being the 15 nH per foot number.
Currently I have the high frequency range to my liking, even though it's not flat, the sound and imaging is very close. I've had some inquiries as to whether I can so the same with balanced cables, this is currently being assessed. One thing I'd like to do is raise the impedance at the lower end of the spectrum. I have a really old Japanese made interconnect that actually does have a very fine inner conductor, just haven't explored the properties of it yet.
FYI, the interconnect under test could not be used for mm/mi phono input because the overall impedance is reduced by about 25%, the sound is compressed with that kind of variation.
If I connect something like 50Ohm to the end of the cable, the whole curve will shift to about 50 Ohm. So actually I kind of wonder why OCOS headed in the speaker cable direction. Seems like once the speakers are connected, it makes not so much difference.
You know, you are correct that the thicker the dielectric, the better the low frequency performance in listening tests. I know people that use 47 lab interconnects as reference.
Last edited:
That is exactly correct, the old OCOS report mentioned this, and I actually have verified it.
- Status
- Not open for further replies.