What makes a Digital Interconnect

Thanks Al,

For me, choosing an audio vendor is 50% quality and 50% service. I'm demoing a few cables by this company and I can say without a doubt that for 4-5 years there has been no better service oriented company.

If you don't like it, they take it back. In one case, I didn't like the cable and let the service rep -- the owner -- know. He sent me a new cable not yet released to the public and 3x the cost of the one I originally purchased. All he asked was that I give him feedback.

A good company is hard to find. These guys really strive to make you happy.

I'll give cable Z a try.

Best,
K

My curiosity piqued, I did a few minutes of detective work and figured out which cables are being referred to.

I notice that their higher-end digital cable (let's call it cable Z), which costs a bit more than four times as much as cable Y, has much higher capacitance, with slightly lower inductance (Ls = 1.5uH center pin; Cp = 270pf). That calculates to almost exactly 75 ohms!

It appears that cables X and Y, and possibly cable Z as well, use an unusual construction technique of having the conductive path for the center pin being a deposited layer surrounding a heavy gauge non-conductive core. The claimed advantages of that approach appear to essentially be mechanical in nature (durability and perhaps easier and more consistent manufacturing).

But I suspect that approach is the underlying reason for the relatively high inductance, which results in the cables having very high characteristic impedance unless the high inductance is offset by high capacitance (which is the case with cable Z).

The bottom line, imo, is that I would not recommend cables X and Y for use as digital interconnects.

Regards,
-- Al

Well, that's more like it, in terms of the shield inductance being less than the inductance of the center conductor. However, 2.6uH and 65pf computes to a characteristic impedance of 200 ohms. So I'm at a loss to explain how it could be a 75 ohm cable.

Regards,
-- Al

I posed my OP to the manufacturer. Here's the response edited to anonymise, the cable in the OP is X:

"The X is an excellent digital cable, the Y is the digital version, and essentially just has more shielding. Both are 75 Ohm cables, with killer bandwidth and excellent group delay. If you are using X [instead of] Y, I would save your cash, spend it on music instead. "

Here are the specs on the dedicated digital cable "Y":

3.3’(1.0m) [RCA > RCA]
Geometry: coax
Bandwidth: > DC - 1 GHz | Linear Phase
Cp: 65 pF (pin / ground)
Ls: center pin 2.6 μH
ground 0.5 μH
Bend Radius: 3.5” (90mm)
Cable Diameter: 0.35” (4.8mm)
Shielding: Shielding shunted to ground at both ends of cable

Almarg

Al wrote
Columbia (BSEE); RPI (MSEE). However, my engineering career (now retired) was in defense electronics, not audio.

For a moment there I felt quite safe... knowing I was out and you were in.... then I read the retired part.

I'm feeling less secure Now.

Good quote from Charles Lamb, Stan. I'll have to remember that one!

For the op's info, 110 ohm digital cables are typically used with xlr connectors for transmitting balanced aes/ebu digital signals. 75 ohm cables are typically used for unbalanced spdif digital signals, via rca or bnc connectors.

In the case of digital or other high speed signals, it IS important that cable impedance match the output impedance of the transport or other source, and the input impedance of the dac or other load. Otherwise increased jitter or even mis-clocking and data corruption can result.

Good point about 1.5 meters (and not less) being the optimal length. The reasons are explained here: http://www.positive-feedback.com/Issue14/spdif.htm.

The importance of all of this is, of course, highly dependent on the jitter rejection capabilities of the particular dac design. There are several unquantifiable variables that enter into the picture as well, including the risetime and falltime of the output signal of the particular transport; the ambient electrical noise environment; ground offsets between the two components; the value of the logic threshold for the digital receiver chip at the input of the dac; the clock rate of the data (redbook or high res), etc. So Stan is right -- give it a try, and if possible compare with a few other cables in your particular system.

Best regards,
-- Al

Actually, despite having at least 3 ICs specifically designed for digital , I use a standard IC for carrying the data between transport and DAC. I find it does a better job in MY SYSTEM. I see some of the Digital cables are specified at 110 Ohms instead of 75 and some makers say the 1.5 M is the proper length. Al , I will leave the technical details up to you, my attitude toward them is the same as Charles Lamb's toward matters concerning Space and Time: "Nothing puzzles me more than such questions and yet nothing bothers me less as I never think about them". Seriously, I have tried a variety of dedicated digital ICs as well as standard ones and never found 2 who were indistinguishable. Possibly if I had tried freebies or really cheap ones this would have happened. Unless you want to do some serious research don't worry too much about the specs., just hook it up and see what it sounds like.

What engineering school did you attend?

Columbia (BSEE); RPI (MSEE). However, my engineering career (now retired) was in defense electronics, not audio.

Best regards,
-- Al

Al,
Something tells me you have more than a passing fancy with audio gear, lol.

What engineering school did you attend :)

Howard

The 75 ohms refers to what is called "characteristic impedance", which has relevance only to frequencies that are much higher than analog audio signals (such as frequency components that are present in digital signals, video signals, and radio signals). It cannot be measured without special purpose test equipment that can deal with those high frequencies.

To a close approximation it is equal to (the square root of (inductance per unit length divided by capacitance per unit length)).

In the specs you've listed, Cp represents capacitance, the p denoting that it appears as a parallel capacitance between the center pin and ground. Ls denotes the inductance, the s indicating that it appears in series, along the length of the wire.

I have no idea why the series inductance for the ground shield appears to be specified as equal to the series inductance for the center conductor. I would expect the inductance across the shield connection to be far less.

Also, 56pf and either 1.6uH or 3.2uH (which would be the total round-trip series inductance if the center conductor and the ground conductor/shield were each 1.6uH) do not come close to equalling a 75 ohm characteristic impedance. 56pf and 1.6uH compute to a characteristic impedance of 169 ohms, while 56pf and 3.2uH compute to 239 ohms.

Re your question about bandwidth, 1GHz is a vastly higher frequency than a cable terminated with rca connectors would ever be used for, and probably just represents the frequency range over which the cable itself (apart from the connectors) can supposedly be used, without unreasonable signal loss. Although the spec is meaningless without cable length and attenuation for that frequency also being specified.

Regards,
-- Al