ANALOG RESEARCH TECHNOLOGY

 www.ar-t.co

 

Analog Research Technology
Electronics for Serious Audiophiles


We all know that when it comes to SPDIF, since it is a digital signal that "bits is bits". And any ol' cable will work. Regardless of length, impedance or connectors.

Wrong!

There is much more to it than just that. While on the surface, it would seem that any cable that will allow the bits to go from the transport to the DAC section would be fine. It takes more than the ability to pass the bits without errors, if one is after maximum performance.

The SPDIF interface needs to be recognised for what it truly is: an RF interface, that has to maintain a tightly controlled impedance of a wide frequency range. The root of this is in the way the clock (that is transmitted from the TX to RX end) is encoded in the digital data stream. It is unfortunately encoded in such a manner that the recovered clock (at the RX end) has strong amount of data-correlated artefacts. These artefacts, when present in the clock signal used at the DAC, cause serious jitter problems.

Yes, we are aware that the concept of jitter being audible is controversial in some circles. At least to the extent as to the amount of jitter is audible. To date, all of the claims that "x" picoseconds of jitter is (or isn't) audible has been made using jitter that is of a random nature. These claims are most likely correct. However, when the jitter has a strong degree of data-correlation, the levels of jitter which become audible are much lower.

The question then becomes how to get the bits from Point A to Point B, in a manner that will allow for the least amount of data-correlated jitter. There are lots of facets, but most require careful attention to design details. Sadly, this is beyond the scope of the consumer. In fact, it seems that in most cases, it is also beyond the scope of most equipment manufacturers.

This compounds the problem for the consumer. They can not alter their equipment. So, how does one go about finding a cable that is able to work around some of these issues.

Well, that is where the U-byte © cable comes to the rescue.

While it can not obviously fix the internal design problems of your equipment, it was designed with these limitations in mind.

So, how do we do it, you wonder.

Well, in the simplest of terms, it all comes down to controlling reflections.

OK, what is a reflection, you ask.

Any time a signal travels down a cable, a certain amount is absorbed by the load. The amount that is absorbed is a function of the load impedance, relative to the impedance of the cable. The amount that is not absorbed is reflected back to the transmit end. Where a certain amount of it is absorbed. Again, the amount absorbed is a function of the source impedance, relative to the cable impedance.

In an ideal situation, the load impedance perfectly matches the cable impedance, and all of the signal is absorbed, and none of it is reflected back to the source. Of course, there never is an ideal situation, and none of us live in an ideal world. In the real world, part of the signal is always reflected back to the source. Where part of that reflected signal gets reflected back to the load end.

Where it arrives along with some other bit of signal information! Isn't that nice? (No, that is bad news!)

While "bits is just bits" when you have a nice bit arriving at the time when a fraction of a previous bit arrives.......hey, that isn't supposed to happen! Is it? Of course not. But it does. All the time.

The result of the desired bit arriving at the right place and at the right time, but mingling with part of some other bit, arriving at what surely must be both the wrong place and time, leads to problems.

Just how is that, you ask.

Simple.

Suppose that undesired signal (let's call it a reflection, because that is what it is) arrives at the decision point. You know, the point in the data stream where the bit is recognised as either a "1" or a "0". While the reflection is always of a small enough level that it will not cause an error (a "1" will always be "1", and a "0" will always be a "0"), it does something really sneaky.

It changes to point it time that the "1" is recognised as a "1", or the "0" is recognised as a "0". And that, our friends, is what is known as jitter. The digital value is correct, but it is off by just a little bit in the time domain.

I think we all know what jitter means when it comes to digital audio. Especially when it is data-correlated. (For those of you who may be new to this, muddy bass and nasty, rough sounding highs are what you will hear, for starters. You say that you don't want to hear that? Well, keep reading.)

So, just what can be done about it?

Well, to most companies, nothing. Whether they view it as an necessary worry, or additional, bothersome expense, or something that they just plain don't understand, well, take your pick on what they think. They obviously don't think enough about it to do anything about it.

To us, lots can be done.

Through a careful process of selecting a cable that has the right impedance, and impedance tolerance, along with connectors that are also of the same nature, we can start to design a cable. One must also select one with high-quality materials, as that will have an effect on how the cable performs over a very wide range of frequencies. In addition, one must have a knowledge of how signals propagate along a cable. And that is where we have the edge on the competition.

Most competitors probably realise that there is more to it than "bits is just bits". One would hope so. However, they most likely view the solution as they would any other audio solution. But when it comes to RF interfaces, and ones as poorly conceived as SPDIF, it is essential that cable (and equipment ) designers have a thorough knowledge of the subject. Alas, few do. Members of our design team have been working with these matters since the 60s. Let us say that we have more than enough knowledge and experience to design a cable that will perform as well as any cable can in any given situation.

OK, now the bad news:

The cable is around 16' long, and is only available with BNC connectors.

Why so long, you ask.

Well, let us say that a detailed explanation would give too much help to the competition. There is a reason, and that reason is to minimise the effects that reflections. The good news is that you don't have to stretch it out, all over your listening room. It can stay rolled up, in a neat coil, and tucked away.

But, your equipment has those nasty RCA jacks. Once again, this shows how poorly conceived the SPDIF interface is. However, all is not lost. We can supply BNC - RCA adaptors. Will this will have a somewhat deleterious effect, it will still allow you to utilise most of the benefits of our cable.

So, if almost all equipment uses RCA connectors, why does this one only come in BNC connectors?

Remember, this is a high-performance digital cable. It is designed for systems that have enough attention to detail to all phases of its design to allow the maximum performance. We will proudly match our cable up against any cable, regardless of price. Even those costing 10 times as much. We believe in it that much. (The ironic part is: the better the equipment design is, the less difference there is between cables. So why pay over $1000 for a cable when you can have one that is just as good, and costs substantially less?) The good news, for those of you who do not own equipment that cost more than a luxury car, is that you can have the same benefits in your system. Even if the equipment wouldn't pass the muster at our lab or uses RCAs, you will still hear a marked improvement.

Buy this cable, and if you don't hear any improvement, send it back. Full refund, no questions asked. (OK, you have to pay to send it back. We can't do everything for you!)

While it has been 15 years since we stopped production of it, when it was in production (and being used in a great number of CES displays), No one ever sent one back. Period. Never.

(Update: July 2009. Someone just sent one back. Always has to be a first time. Or a wise guy. Take your pick!)

Wonder who they sold the ones that they were using to? We feel sorry for those guys!

Further update: February, 2013. They are making a comeback! Found a different type of BNC. No, we did not listen to it. No, we have no intention of listening to it. And it is a moot point, since it took all this time to find it. We are using it.

(It can't sound that awful, since we are already using a different style, for the cheap "freebie" cable we included, with all of our other fine/crappy/whatever products. And no one has sent any of them back. So, there!)

"Maybe because they are free. And you still go out of your way to annoy us, don't you? And include all of your inside, private jokes, as part of the sales pitch."

Si.

Hey, look at the bright side: it is free entertainment. Besides, what do you expect, for this money, Cousin Brucie?



Specifications:

Impedance: 75 ohms, +/- 1.5 ohms.

Length: Approximately 8 feet.

Construction: Center conductor and shield, 100% copper; foamed dielectric.

Connectors: 75 ohm BNCs.

Special (re)(re)introductory price: $200. (Price inlcudes shipping by Priority Mail. Such a deal.)

"Sounds just like the old price."

It is. So what are you complaining about?

For information, or to order one:



Not for sale to Ireland!




 
 

 

 

ANALOG RESEARCH TECHNOLOGY: Years of design experience in all phases of analog electronics, ranging from DC to microwaves.

 

Typical listener/user comments:

"What a sweet spot!  You don't have to sit with your head in a vise!"

"Wow.  My search is over!"

"Now THAT'S what I wanted all along!" 

"I love 'em!"

"These things SING!"