Cables with BNC Connectors/Adapters Are Generating A Superoir SQ

[christoffe]

Hi,

attached a file from a Dartzeel manual with their arguments for the choice for  a coaxial BNC cable.

Keywords are:  “external disturbance immunity”, “virtual absence of losses” and “without any sonic alterations”

This seems to be the explanation for the superior SQ with the BNC cable, but their “impedance matching” does not cover our case.

Joachim

[PeterSt]

If you have read the PDF I linked to in my last post, you will better understand these texts (which I trust) I copied from some forum :
(notice that "driver" here means a device that creates current like in "we need to be able to drive that length of cable) :

For a back-of-the-envelope calculation, capacitances in parallel add together. So, for example, if you had 2 loads of 15pf, the total would be 30pf. Drivers have a characteristic output impedance. Transmission lines (board traces) have a characteristic impedance. Loads (receivers) have a characteristic impedance. Impedance is defined as
the ohmic resistance plus the net reactance, which can be either capacity or inductive, or can equal zero. In the case where the net reactance is zero, you would have a purely resistive load.

In situations where the source, line and load impedances are not equal, there will be a reflection of the signal. In other words, when the source sends a signal to the load, the signal will bounce back to the source and distort the signal (i.e. make it less square). The exception to this rule is when the length of the transmission line is small compared to the wave length of the signal.

and a somewhat more difficult one :

A resistive load passes current in proportion to the instantaneous voltage. If there is voltage across the load, it is passing current, regardless of time.

A capacitive load pases current when the voltage across it changes, and in proportion to how fast it is changing. Any steady voltage drives no current through a capacitive load.

So resistive loads draw current throughout the time when the source applies voltage to it. The capacitive load draws a spike of current when the source tries to step the voltage, and slows the rise time of the step, if the source has series resistance.

If the load is a resistance and capacitance in parallel, then the source must supply both currents simultaneously.

If the load is a series combination of resistance and capacitance, the capacitor blocks any DC current, and the resistor limits the current that can be passed during any fast rate of rise or fall, but extending the time the current passes, after the step in voltage has stopped changing. There will be and exponential decay of current as the resistor drop shrinks, and the capacitor sees the fill applied voltage gradually,  onstead of as the source changes it.

I like to highlight this from the last text :

A capacitive load pases current when the voltage across it changes, and in proportion to how fast it is changing. Any steady voltage drives no current through a capacitive load.

Notice that with "steady voltage" DC is meant. This also implies that the slower the AC change (hence the lower the frequency), the less noticable the result will be when a capacitive load is in order.
Now you also know why we remove input capacitors.
And what the aim is wihh the cable ...

Peter

[PeterSt]

This seems to be the explanation for the superior SQ with the BNC cable, but their “impedance matching” does not cover our case.

Hi Joachim,

You understood this wrongly I think. You must have gotten it from this in that PDF :

After a lot of research, we concluded that the one and only means for transmitting an electrical musical signal with no alteration or losses over a long distance is impedance  matched lines, from end to end.

This is the very same as what I said a couple of time like : "All over 50 Ohm cable".
Or "all through" if you like.

So for maybe better explanation : A cable can be "impedance matched" within itself. And when this is so, it means that not only the cable is matching e.g. 50 Ohms impedance, but also the connectors and their connection.

If you'd observe 50 Ohm BNC connectors and 75 Ohm BNC connectors, you'll see that both are different. This is not because it is another standard or so, but because of the dielectricum causes the "impedance" (think inductance now) and the dielectricum can not be the same for 50 Ohm and 75 Ohm. The cables are also different and this is not because they have a resistance of 50 vs 75 Ohm but because of the LCR properties which togther create the impedance. For a cable this can be done in all sorts of means because it is a relation between the copper, the shield and the dielectricum (but the diameter of the copper is the same for all e.g. 50 Ohm cables because otherwise it won't fit the STANDARD connectors (for 50 Ohm)).

Look at the construction of the connector, which is not for "grip" ...
Now compare an RCA connector and see that nothing CAN be done about any impedance. There's just no (dielectricum) facility.

Btw, all is so fragile (and important) that (they say) not even 75 Ohm or 93 Ohm can perfectly be impedance matched. Only 50 Ohm can (somehow).

Side note : Dozens of such connector means exist with maybe the most familiar one the F connectors (for satellite cable). All come along with their matching cables.
So are we done with this ?

And to get what I mean :
All I need to do for better matching "throughout" is replace the 33 Ohm resistor in your NOS1 with a 50 Ohm and ...

And nothing (yet) because BNC is not the best connector at all. TNC for example is way better. It comes with/for what cable ?
But before that we must first cut out the BS of the additional adaptors, right ?

How many times did I say it by now ? It is only a start.
Regards,
Peter

[PeterSt]

Something else about that DartZeel PDF :

What they say in there I did not DARE to say this morning :
Why the h*ck should the cable length be so limited. So I thought of the cable TV example. Hence from the source to your house is, how long ?

DartZeel talks about 1Km "and not to make fun". Well, I right away agree just because I don't see why not.
But take out the reflections !




It will need some current though.

[PeterSt]

There is something else I like to mention :

From the mere hip-hop scene I have this album (CD) MC-Hammer - Too legit to quit. I have played this a 100 times from my TEAC P1 whatever CD transport. And well, right from the start with computer audio, this did not want to work any more.

Like with so many complaints with computer audio (sure when I started with it) it is too "light-weight". Well, something like that. It is a very dynamical CD of good quality also (IMO) and the dynamics of it became too profound to nicely play it.
And thus I regularly test it (like I do with all which won't play well) and yesterday, finally ... it worked. And HOW !

In other words, no matter the speakers I used since "digital" or no matter the number of woofers and/or sub woofers, and just was too light. Too snappy. Harsh because of that. No fundament.
Now ? oh man. First there's the similarity with the old sound, but next is what all has been added for SQ over the years.

It is (now) also easy to hear what it requires. Something like 3x 380V or so. And that is what we mean (I include Joachim); you can just feel that all was squeezed before. Now it's unlimited flow of current, so to speak.

Peter

[CoenP]

Can't touch this!

Peter, the link provides excellent information on the -loudspeaker-cable resonances. Allthough I think the resonance picture with the simulated Proac is a little misleading since a computer model will assume an ever rising impedance for the tweeter induction (rising of the impedance on the right side). This will not be the case in practice.
How did they solve it? I guess by placing notch filters parallel to the cable.

Nevertheless it is a real phenomenon and the resonance energy can creep into our precious audio signals. Also there is a likely source of UHF because of the NOS character of the DAC specifically designed for very wide bandwidth square waves. Not to forget USB noise creeping into our cable.

Hmmm, I wonder if this is how the Bybee products work: just eating -excess- HF energy (do they work on digital too?)?

There's more ways to skin this cat, yet the simple termination solution sounds excellent to me.

The discriminator between professional telecom cables and boutique audio stuff is the tight tolerance for the characteristic impedance. These cable's intended use is for much longer distances then in our home environment and you cannot have your impedance vary much otherwise it won't work and you invested a lot of money for nothing.

The adapter has a resistance of 50Ohms too.

So the NOS sees 33 + 50/2 = 58 ohms for LF and 33 ohms for HF    and seems to be comfortable with it.

from the article:

If the amplifier is not designed to drive very low-impedance loads, say 1-2 ohms, it may saturate, causing distortion. Even if it does not saturate, the circuitry that eliminates crossover-distortion in most amplifiers may not work optimally at high loads. In fact, the crossover distortion in most amplifiers increases with load.

This is what I was referring to. This is true for any class AB amplifier, including opamps. Seems to be a non-issue.

Now let's have a look what's in my cable box tonight.

regards, Coen

[manisandher]

DartZeel talks about 1Km "and not to make fun". Well, I right away agree just because I don't see why not.
But take out the reflections !

Edit II: Note to Mani - this BNC cable, details and photo later today, might help to place the NOS1 + PC outside of the listening room without a degeneration of the SQ.

Hey Peter, so there's a distinct possibility that I'll be able to use 10m cables no problem? That would really be great.

Will you now be offering the NOS1a with an option of 50 Ohm BNC output connectors... and the Orelos with an option of 50 Ohm BNC input connectors? Or does the use of 50 Ohm BNC-to-RCA adapters negate this?

Mani.

[CoenP]

10m?

For killing reflections/resonances no problem, but cable capacitance will become more significant (this is relevant for audio frequencies) so ymmv...

regards, Coen.

[manisandher]

Thanks Coen.

The 50 Ohm Huber + Suhner cables (the ones I've just ordered and the ones that Joachim is using) are around 95pF/m. So a 10m run would still be under 1nF. Is there a straightforward way of calculating how this would affect audio frequencies?

EDIT: According to the Huber + Suhner data sheet (the one Joachim linked earlier), at 6GHz, these cables attenuate at 1.5dB/m. From the graphs, it's hard to get a meaningful figure for the Mhz range, let alone the KHz range... but it's obviously going to be very low.

Mani.

[CoenP]

I just looked at it:

Source impedance is with 33 ohm series resistor and 50 ohm the first adapter: 19 ohm. This is low enough for a 1 nF cable: -3 dB on 830kHz. The second adapter containing also 50 ohm even improves on this.

Regards, Coen

Ps. audio signals have very long waves (>10km for 20 kHz) so the hf characteristic impedance does not apply. We must take the capacitance  into consideration.

[manisandher]

Thanks Coen.

Cheers, Mani.

[christoffe]

The 50 Ohm Huber + Suhner cables (the ones I've just ordered and the ones that Joachim is using)

 

[christoffe]

Calculation of (normal audio) cable attenuation

see:

http://www.sengpielaudio.com/calculator-cable.htm

[manisandher]

Thanks Joachim.

So Coen, were you calculating with 1nF/m? This comes to -3dB at ~830KHz. Putting 95pF/m comes to -3dB at 8.8MHz.

???

Mani.

[CoenP]

Hi Mani,

You're right. I have misscounted the number of digits on my unscientific calculator. 1nF was for the entire cable.
Anyway in the mhz range other stuff becomes important as wire induction, and of course the hf transmission characteristics.

So thanks to the low source impedance cable capacitance still doesn't matter at 10m. This was actially my first point  .

I didn't make it to the box tonight and i wonder if I did not throw away all my scr*p token ring coax. if so I could start with a simple video cable and do the 75 ohms in the connectors. (Did anybody try this allready?).

Regards, Coen