Digital cable experiment

Tweaker · « **on:** 23 Apr 2005, 12:03 am »

I have had an unexpected result from replacing my digital coax (a Bolder Cables coax) with a very old Vampire Wire multistrand analog interconnect. Don't ask why, I just do stuff like that now and then. What was unexpected was the Vampire Wire sounded WAY better. Not quite as potent in the bass regions but tons more texture coming through. True from top to bottom. An example of what I'm talking about is acoustic bass. The Vampire let me hear more of that smacking sound that the fingers make when plucking the strings. Everything is now produced with more texture and life and more of a sense of rightness. Also more warmth which is welcome by comparison to the coax. I tried the same experiment with my existing analog ic's which are JPS Labs (solid core) and did not get the same results. Anyway, I'm pleasantly surprised but puzzled. I do remember, some years back when outboard DA converters were just coming out and, of coarse, the coax cable that "must" be used with them as well, talking to a salesman who said it was a lot of bs that a special cable was needed, that any old analog ic would work. Maybe he was right. Although my experience with trying the JPS Labs suggests that is not the case.
I would encourage anyone interested to try the same thing and post your results here.

art · « **Reply #1 on:** 23 Apr 2005, 04:11 am »

Using coax cable is not BS. It must be the correct impedance for the system that it is used in. If all the impedances are matched properly (which they almost never are), all digital cables sound about the same.

From what you describe, you have significant impedance mismatches if bog standard twisted pair sounds better.

A certain famous audio guru produced a cable back in the early 90s that was 93 ohms. It sounded best on his system.

It sounded best because he could not figure out how to make his transport 75 ohms. It was over 95 ohms.

Most DACs have inputs that are not 75 ohms, and have a significant reactive component. No telling what will work on stuff like that.

And then you have the impedance problem that arises from using Crystal RX chips. Trust me..........

Pat

Tweaker · « **Reply #2 on:** 23 Apr 2005, 12:23 pm »

The Vampire not only sounds better, it sounds sensational. Which it shouldn't because it's not a special coax designed for digital transmission. Forget about whether the coax I have doesn't sound as good because of a possible impedence mismatch. For years we've all bought into the notion that a special cable is required to use as a digital interface. I've got a cheap analogue cable that is defying that convention.
I am hoping others will try the same thing and see if they have similar results or if I am somehow in possesion of the only analogue ic on the planet that sounds great used in place of a coax.

ctviggen · « **Reply #3 on:** 23 Apr 2005, 01:55 pm »

Impedance mismatch is most likely why your "analog" IC works better than a "digital" IC. Or, put differently, your analog IC happens to have an impedance that matches better with your DAC.

art · « **Reply #4 on:** 23 Apr 2005, 02:11 pm »

Exactly, which implies that something is wrong with the DAC input circuit.

It almost always is..................

Pat

Tweaker · « **Reply #5 on:** 23 Apr 2005, 03:14 pm »

I have a new Behringer Ultracurve that I'm also using as a DA. I will put my ART D/IO back in to the chain and see if I get the same results. If what you're saying is true about impedence mismatches then I suspect there are a whole lot of people out there with the same problem without knowing it. The Behringer also has optical inputs and the sound I get using that and the sound I get with the Bolder coax is similar, which, by the way isn't bad. If I hadn't tried the Vampire I would have been quite happy with the sound.

Tweaker · « **Reply #6 on:** 23 Apr 2005, 09:50 pm »

I just completed A/B'ing the cables with my ART DI/O replacing the Ultracurve and while I still thought the Vampire sounded better it was only by a very,very small margin. With the Ultracurve it's by a very large margin. The impedence matching must be the problem as suggested. The Ultracurve I/O's, as it turns out, are 110 ohms. I don't know the DI/O's spec. So I have a transport with a 75 ohm output mating with a D/A converter with 110 ohm input apparantly making my Bolder cable coax unhappy. Damn. Well as long as the Vampire works there's no problem. However there's a bunch of people out there with Ultracurves in their system who, I'm assuming, aren't getting the full potential out of the unit if they are using a 75 ohm coax.
As usual, the technical expertise from members of this forum has been very helpful and has possibly saved me from going on some kind of anti digital-coax crusade.

I also found this informative:
http://www.josephelectronics.com/html/misc/DigitalStudio.html

art · « **Reply #7 on:** 23 Apr 2005, 11:57 pm »

Consider your self lucky. Either 75 ohms (S/PDIF) or 110 ohms (AES/EBU) as your input impedance is at least something standard. A lot of "high-end designers" make gear with neither.

Twisted pair has an impedance somewhere in the range of 100-150 ohms. Usually.

Pat

Tweaker · « **Reply #8 on:** 24 Apr 2005, 12:04 am »

I'm wondering if a mod to the Ultracurve could be done to change the input impedance to 75 ohms?
Hmmm, I guess it's time to start looking for a transport with AES outputs.

dave_c · « **Reply #9 on:** 24 Apr 2005, 01:24 am »

AES/EBU connections are usually XLR's which isn't the case here. I'm sure you could find someone to mod your Ultracurve to match the impedance. Wayne at Bolder Cables could probably do it for you. Steve at Empirical is also another great resource.

Tweaker · « **Reply #10 on:** 24 Apr 2005, 01:40 am »

The Ultracurve has balanced XLR inputs/outputs.

Wayne1 · « **Reply #11 on:** 24 Apr 2005, 02:03 pm »

The stock ART D/IO input impedance is 100 ohms.

That is one of the many things changed in the smART and Mensa mods.

The UltraCurve input impedance can be changed. You could also pick up a Monarchy DIP which has either optical or coax inputs and coax or AES ouputs.

There are cables made just for AES use with 110 ohm impedance. I recently built one for a special order for one of my customers.

John151 · « **Reply #12 on:** 24 Apr 2005, 03:44 pm »

From a novice, theoretical perspective, either the bits are passed thru 100%, or they are not . The bits can not be "colored". If the cable is faulty, and not all of the bits are making it thru, I would expect the sound would be really, really bad. Now I am wondering if the Ohm problem causes timing problems.

Tweaker · « **Reply #13 on:** 24 Apr 2005, 04:50 pm »

To use a 110 ohm cable with my existing setup I would have to have one end terminated xlr, the other rca. I'm wondering if that would still be ideal as the output of the transport is still 75 ohms.
I ran across this last night researching options and I think I might give it a try before spending more for either a mod, a new transport, or something like the Monarchy DIP. It's good to know that there are options,though!

http://www.behringer.com/SRC2000/index.cfm?lang=ENG

audioengr · « **Reply #14 on:** 24 Apr 2005, 04:59 pm »

Quote from: art

Using coax cable is not BS. It must be the correct impedance for the system that it is used in. If all the impedances are matched properly (which they almost never are), all digital cables sound about the same.

True - impedance match is critical.

However, there are other significant factors that affect jitter, including losses and dielectric absorption. These are not a function of impedance. This makes most 75 ohm cables sound different.

Also, cable length should be 1.5m for unmodded transports. See this white-paper:
http://www.positive-feedback.com/Issue14/spdif.htm

art · « **Reply #15 on:** 24 Apr 2005, 05:06 pm »

It is true. If the impedance is off, you will have excessive reflections, and all that other stuff is meaningless. If the impedance is right, then you can worry about all of that stuff. But first you have to get it right, and most can't. (Maybe you can, but I don't know..........never had to fix any of your designs.)

As for "bits is bits".........means nothing. Impedance mismatches cause reflections, and depending on the amount and point in the waveform that it happens, the timing will be off. Then you have jitter, and rotten sound.

Ah.......I see that you have edited your post............

Anyway, since you brought up the cable length..........

Cables that are 1/2 m and shorter almost always sound awful. Don't believe me? Go try it. I prefer something longer, in the 6 m range.

Pat

audioengr · « **Reply #16 on:** 24 Apr 2005, 05:47 pm »

Quote

Cables that are 1/2 m and shorter almost always sound awful. Don't believe me? Go try it. I prefer something longer, in the 6 m range.

I believe you. I did the analysis and wrote the white-paper.

Steve N.

Tweaker · « **Reply #17 on:** 24 Apr 2005, 05:57 pm »

I have a GW Labs DSP I had completely forgot about as I had removed it from the chain some time ago. It will take an unbalanced input and convert to balanced. So I have my coax running from the transport to the GW Labs, and since I don't have an audiophile XLR cable I'm using a balanced 50ft(!) microphone cable from the DSP to the Ultracurve. Not given it a serious listen yet, but a brief listen sounded quite good. I have both a 1 and 2 meter coax, sounds as though I should be using the longer one. And based on what art is saying the 50ft XLR may not be such a bad thing.

ctviggen · « **Reply #18 on:** 24 Apr 2005, 06:04 pm »

Wayne built the AES IC for me, as my Proceed PMDT states that the AES connection would be the best, followed by BNC, then RCA, then optical. I've compared my new Ack Dack 2.0 with a Magic Digital One RCA cable to the AES IC into my Proceed AVP I. With the AES IC, it's a close call between the two.

art · « **Reply #19 on:** 24 Apr 2005, 06:09 pm »

Baloney.

The XLR connector does measure 110 ohms, but it has too much fringing to be useful for a digital audio stream where you have to recover the clock via a PLL.

They are right about one thing: TOSLINK is the worst. "ATT fiber" could be good if done properly, but alas, just as in the case of coax, it is done wrong. In fact, even "more wrong", if such an expression makes sense.

Pat