Cables with BNC Connectors/Adapters Are Generating A Superoir SQ

[PeterSt]

Hey Coen,

I'm afraid I need a couple of hints here. I mean, I don't see what you're getting at and the sense I get from it is (was) negative (derived from previous post(s)). From your last post I must derive a positive sense ??
Am I wrong ?

Mind you that the imperfect 75ohm RCAs are still in service (!).

Unimportant I think because not part of the capacitive load. And resistive (if any) ? Also unimportant (it is about what happens in the cable)

what about the shift from a capacitive to a resistive load for the NOS's output?

I am not 100% sure at this moment whether we can say that. But what would be true is that the capacative (varying) load is out of the equation ?
So this is what I assume for now and this is where the positive properties start. But whether you try to calculate the influence of a "filter" or not, it is not about that at all ...
And what it is *is* about in my view, looks to be to good to be true. So I'm probably wrong on that (but it would be the explanation).

I have prepared a few texts for some possible insight with the clear notice that I don't know all about all.
So, positive or negative, Coen ?

Peter

[christoffe]

-Quote-
Characteristic Impedance

Characteristic impedance is a high-frequency phenomenon that is a function of the "distributed" R, L and C (inductance and capacitance per unit length). Characteristic impedance Z0 is calculated as (Lu/Cu)1/2 . Lu and Cu are inductance and capacitance per unit length. From the Z0 equation, you can see that Z0 is independent of the length of the cable, the frequency and the resistance. This behavior holds true only if the wavelength of the frequencies on the cable is much shorter than the transit time of a signal from one end of the cable to the other. When this is true, the characteristic impedance will stay relatively constant regardless of the length of the cable and the cable will exhibit "transmission-line-effects". In general, if an event is so fast that it completes a significant voltage transition before it is noticed at the far end of the cable, this will cause transmission-line-effects to occur. Transit times are very short for typical audio cables. For example, a 20 foot speaker cable has a transit time of around 40 nanoseconds. In order for transmission-line-effects to occur on this cable, signals at frequencies above 25MHz must be present. This is orders of magnitude higher than audio frequencies. Therefore, transmission-line-effects and characteristic impedance are definitely second-order effects on audio cables, however resonance may play more significant part than previously imagined and this relates to termination of the characteristic impedance.

-Unquote-

Original text see here:

http://www.empiricalaudio.com/computer-audio/audio-faqs/should-i-match-impedances

[christoffe]

Impedance Matching

comment from the manuf. NuFOrce see:

https://www.nuforce.com/index.php?option=com_k2&view=item&id=635:impedance-matching

Edit:
For a mechanical guy the above explanation, identical found on other homepages too (ratio source/load = 1:10), is quite sufficient. The (my) combination NOS1 and power amp satisfies this ratio and I do not have to increase the XXH volume.
I hear no distortions in the highs and lows, they are better than before. The “clicks” on the cymbals are better to follow and the bass is very tight (no wobblings and rumble at the subwoofer anymore).


Joachim

[CoenP]

So, positive or negative, Coen ?

What about both? !

I do think that distortion -for a 'voltage source'  output- will increase with current; at least that's what I've seen in most amp/preamp test reports. Is this audible or important? That's another question. You must be able to measure this.

BTW this is the 'impedance mismatch' Joachim pointed out in his last post, like my tube phono output preamp playing into 25 ohm with gross distortion (way to much current needed), loss of gain and HF (high output impedance) and finally loss of LF (too small output caps).

No doubt either that the phase angle of the output load has an influence on the performance of the DAC's output opamps. These must be in some feedback arrangement (Inverted or not) which will be sensitive to phase twists. This corresponds with anecdotal evidence on this forum that lower capacitance cables sound better.
A rule of thumb 1:10 ratio, that would mean a happy NOS1 at a load of 330 ohms?

The point where cable capacitance becomes important in this setting is determined by the load at the end of the cable.
Now we shift the frequency by increasing the load (=lowering the resistance) this point is moved at least three decades upwards, making the load for the DAC resistive for the whole audioband.
This is the good stuff, again: is it measurable and important? I don't know it's audibility is speculative but plausible.

Unlike for the very high frequencies, audio signals encounter the plain parallel cable capacitance and series inductance (which is very small) and not the characteristic impedance. Joachim's snip points this out clearly. This should be taken care of in USB transmission lines for instance (proper termination, short cable?).
Yet there could still be HF energy in the relevant spectrum on the DAC's output that would remain undamped in a normal setup potentially causing unwanted interference.

I see enough good stuff to offset the bad so in the end you can call me positive .

regards, Coen

[christoffe]

Hi,

I was searching the internet with the following keywords:

"impedance matching preamp power amp"

and there are tons of informations, most of them "driving" in the same direction.

In Audio no "impedance matching" is necessary. (in general: source [DAC] impedance shall be very low + load impedance [power amp] shall be much higher, more > 1:10)

The best article for "mechanics" I found is here:

http://www.soundonsound.com/sos/jan03/articles/impedanceworkshop.asp


Joachim

[PeterSt]

Before we all drift off in the wrong direction, I think this was "wrong" :

R u guys kidding us?

Correct me if I'm wrong but 2 adapaters, 50 ohms each that's a load of 25 ohms for the NOS1(a), output and terminating impedance not taken into consideration. You are just listening to the extra distortion of the output stage. I have speakers here with a higher impedance.

You must have had a loss of gain.

It could work but what impedance was the NOS designed for? Irrc most audio cables and RCAs are about 75 ohm, terminate both sides with that and you will eliminate reflections as well.

Please don't think I am sceptic about the idea, it was on my todo list as well but the output stage needs to be designed for it.

regards, Coen

So if my amplifier is 100K, the impedance relation is 33 - 100000 and this is not related to some 75 Ohm or 50 Ohm under way. And if it is, it is about the difference between the normal 75 Ohm and the now 50 Ohm.
Do not relate this to the output impedance of 33 Ohm of the NOS1.

N.b.: Nobody tells that any interlink is 75 Ohm, but the connectors hope they are. (but they will not ever be).
*My* interlink was 75 Ohm alright, but that's coindicence (not of course but I use coax, so I can know - do you ?).

Peter (will try to come up with some more)

[PeterSt]

I do think that distortion -for a 'voltage source'  output- will increase with current; at least that's what I've seen in most amp/preamp test reports. Is this audible or important? That's another question. You must be able to measure this.

BTW this is the 'impedance mismatch' Joachim pointed out in his last post, like my tube phono output preamp playing into 25 ohm with gross distortion (way to much current needed), loss of gain and HF (high output impedance) and finally loss of LF (too small output caps).

Coen, allow me ... You seem to put things upside down. But I know, this can be "language". So, "will increase with current" or "needs more current" ? I'd say the latter, and this is quite the opposite (BUT, it can be literally what you say just the same, but then you needs to examplain a couple of things !).
And "way too much current needed". I'd say "way too few current available" which again is something different.

It feels like some lecture is boiling up, but it is not quite on the top of your head yet. And oh, no lectures in my head anyway - just saying (again).

But I will try a few things more ...

[CoenP]

Ok Peter,

The 33ohm output impedance is the result of a series resistor? That makes my cable capacitance point somewhat less relevant.

Regards, Coen

[PeterSt]

Yes.
And no on the latter remark.

[PeterSt]

BTW this is the 'impedance mismatch' Joachim pointed out in his last post,

Gentlemen,

We must be very cautious not to dig such stupid texts as that from NuForce. This is just "copying from the internet" behaviour
1:10 - nice (but at least now you understand how I said yesterday that I don't bother about 33 Ohm because ALL will "match" that).

No, this is BS.

Because Joachim linked to an Empirical Audio page, I can do that too :
http://www.empiricalaudio.com/computer-audio/audio-faqs/are-cable-resonances-real
Don't try to understand this - just digest and see what kind of things play a role. Not "may play", they just do.
And now envision that we talk about this in this topic.

Btw, "impedance matching" does not exist. Try to get that from this link I gave.

Peter

[PeterSt]

We must think out of the box. Or maybe not, but the box where our lecture is in, is not one. Just look into that link from my last post and if you can get somewhat of it, you can already see that a couple of things are not solvable for the amplifier-speaker relation, while we think we can solve it for a DAC (pre)amplifier relation, while we KNOW it has been solved for e.g. communications networks.
Now what.

Because it was said, think about USB as such "communication network". Must be 90 Ohm (2x 45 Ohm (differential) actually) and all is right. All ? Yes, for the purpose. This is about the frequency, the voltage and the current. All is here in a fairly small "bandwidth" (because all is a given).

If you, for fun, would look what's actually better for our interlinks, 50 Ohm or 75 Ohm, it will be very hard to find an answer because there's too many trees to see the forest (this is Dutch). You, all over the place, will run into smart remarks like "75 Ohm is for video and 50 ohm is 'enough'". This is all bandwidth related and (obviously I hope) video requires a higher bandwidth. Similarly you may run into "can I use a digital interconnect for analogue". Answer will be often Yes, but nothing is worked out further because why to do it anyway "because it will be over-rated".

Let me tell you in between the lines : All I do (and you don't see much of that anyway) is about over-rating. This will remain so as long as we can detect audible differences which can't be measured anyway (or we don't know how to, but that's the same). Partly this is related to how ultra high frequencies can play roles and those frequencies can not be measured (can be GHz range).

Contradiction :
While most you will run into is about "we don't need that bandwidth for audio", we better think the other way around :
When these high frequencies are let loose things might go wrong so we better use (e.g.) a cable with less bandwidth so it nicely filters that out.
Remember : cables filter (see Coen's equation about the 26.5KHz for example, and don't be bothered about the truth of it - it is just there one way or the other).

Where was I ? Thinking out of the box. We HAVE to because there is no box for us, never mind we thought there was one. Ok, there is one, for class 1. That says "use 1:10 and you're good". But now WHY and in what situations does it count ?

[christoffe]

BTW this is the 'impedance mismatch' Joachim pointed out in his last post,

Btw, "impedance matching" does not exist.

Hi Peter,

all this are "keywords"  for a search within BING or Google. They were discussing this theme in the forums of CA + Audiogon too.

Joachim

EDIT:
If we are looking for a problem in the internet with keywords, we get thousands of informations, and 99,9% are not covering our search.
So we have to filter, and we will find in one article a paragraph or sentence for our problem only, and this goes on and on and on and in the end the “puzzle” is completed.

[PeterSt]

For a cable, all is about reflections (well, on PCB traces this counts just the same, but we ususally don't "see" that). The Empirical Audio page names it "Resonance" but it really is the same.

Again saying that I don't know all (hence I too can't go into all the details when the chips are down) you can take it from me that a signal must be damped at the other end (the receiving end) and that (indeed) this is a function of L, C and R (induction, capacitance and resistance). These are properties of a cable and when the whole of the cable complies to the same outcome (which would be Z, the impedance) then the signal will be 100% damped at the other end. Damped = No reflections.

It is here where we can easily get lost in our audio interlinks, because it is not taken into account that such a thing can exist. This, while in the pre-era of UTP Ethernet cabling, all was about coax and of various protocols. Btw, cable TV still is. And always and ever this is about one thing only : proper termination.

For metter imagination, we can just as well try to see that cable TV was first with this all. Improper termination ? then you'd have ghost images. Aha, those we can see ! Now think what we would be able to hear of that; way more difficult, right ? But what could be seen with TV and could be improved by just looking, worked out into ... digital. I am not 100% sure but I think it was ArcNet which applied the very same as cable TV and it just worked. Later of course cable TV went digital over the same cable (hence damping means did not need to change). Later, Ethernet went the same route but with a different impedance and again later UTP was better. And of course telephony started to implement the same (digital) over copper (but with less (twisted pair) wires).
OK, UTP is not related, although we could start thinking about UTP interlinks. But never mind that (for now ).
Point is : all is an evolution but we might wonder who is going to take care about our audio evolution ?

Why do ALL analysers work with BNC connectors ?
Think of this like analysers contain better oscillators than the equipment they need to measure. But of course *that* became ancient history, and only in the last two years or so.
Next step could be the proper "interlinks" for the equipment the analysers measure.

... Which of course still doesn't answer the question why those interlinks of an impedance of exact 50 Ohm are better.
Well, unless we realy understand ...
(now let's see)

[PeterSt]

It would be nice if someone can come up with why any "resistance" at the other end needs to be (rule of the thumb) 10x that of the output "resistance".

I know, the general answer again will be about this "damping" needed, and to some degree it will be related to current vs. voltage and that implied by the output.

But whatever is said, it will be in the context of a capacitive load (I think !). It will be related to frequencies as well. But in the very very end I think it is about one single thing only : have those reflections under control the best.
"The best"; again see that Empirical Audio page for what actually always happens, with a speaker driver as the example, that not being a "constant" load.
I don't see much difference with a capacitive load for our interlinks.

But if it were about reflections and that only, and our nice all over 50 Ohm BNC interlink takes this out of this world, then what ?

[PeterSt]

But if it were about reflections and that only, and our nice all over 50 Ohm BNC interlink takes this out of this world, then what ?

Then what ? I guess then I need to find those boxes with the answer burried somewhere. But too many boxes with too much in them, so now I am going to be speculative, pretentious and possibly dumn :

Then no capacitive load is in order, and thus it is resistive.

Yeah, maybe.
But like I told in my last post from yesterday : I'd hope for that. But now why ...

Please notice or keep in mind : From here on it is going to be quite confusing. How ? Well, because of my speculation at first how reflections can cancel out (and add) and how this could be an explanation of the better (way more poserful) bass. I am not so sure any more whether this "explanation" is for real, and whether it can be replaced with what's following. Do notice though that whatever is true from the following, this cancellation etc. WILL also be in order with improperly terminated cables. And the relation with digital again : while digital will keep on working until appropriate voltage levels (alsways the same - High or Low) collapse, everything of the same at microscopic level in analogue, just changes sound right away.
So with our new cable we have two changed phenomena in order : The cancellation etc. because of reflections and :

More power because of current and voltage being in phase.
Never forget for now : if I am right that we can look at a resistive load now (and it wasn't even my idea - it was Coen's suggestion).

Side note : A resistive load can be seen as one which is always equal. It is equal when more current is pushed through (up and down going level of the music (loud/soft play)) and it is equal for all frequencies.
WOW, wouldn't we like THAT !
Yeah yeah, that is why I hope for it.
BUT to keep in mind : All I am doing is finding an explanation for the ridiculously better sound.

More side note : A capacitive load can be seen as a battery which needs to be charged first before at the other end the voltage comes out of it. Of course this is the working of a capacitor. But our cable is doing the same (it is a capacitor), normally.

Hopefully digging in the right box, what happens with a capacitive load is that first current is needed to create the voltage (and the voltage is what makes our speaker move (amplifier ahead of it)). The voltage lags (comes behind) the current. Now, still looking for the right box ) what happens next is that the both (current and voltage) react and the sum of the both is the "net power" we feed to the load.

With a capacitive load (or inductive load the same thing) the real power we feed to the other end is always less than with a resistive load. So, with a resistive load the power is optimal (current and voltage is 100% in phase).

Both feet back on the ground : I won't believe that we suddenly are able to "work with" a pure resistive load. Still, it can have become more of that. This means that we feed more power to the other end than we did before. And this is audible in everyting. Also remember what I told yesterday : "All is more of it".
So it seems logical that we are in this direction.

The other part of it - less distortion, all really "better" - should be because of the frequencies having less influence. And THIS would be the real deal. So, this is not about frequencies being attenuated (because of cable length or reflections or etc.) but about the reactance of the other end (which reactance is not there with our all 50 Ohm cable) not changing the phase of the power (of current + voltage) so to speak, while the reactances (now not there) vary per frequency.

If this is not true then the least what I learned from this is that we must aim for it. So it is still just a start.

Peter

PS: Nice reference from The Power Box :
http://www.aspowertechnologies.com/resources/pdf/True%20vs.%20Apparent%20Power.pdf
Of course this is not about audio, but it is AC and audio is AC.