The DIY'rs Cookbook

[johnjen]

I suspect that the key to this conundrum is throughput performance: do all the bits get to the destination in time for D/A conversion, or not?

BIG snipperdoodles

Hardly a discourse suggestive of "bits are bits" and "all USB cables are the same"...

Another aspect that I didn't bring up, mainly due to the pre-existing girth of part 7, was that of the transceiver chips at each end of a digital transmission line 'System' and their role as data 'interpreters'.

In analog transmission 'Systems' the bandwidth is fairly low and some what wide which means the range of the length of the standing waves and reflection of the signal is also fairly long in comparison to a digital signal.
There is also the added influence of the magnitude of the harmonic content (in contrast to the actual signal itself) difference between an analog vs. digital signal.

Digital signals being square waves introduce plenty of odd order harmonics and generally they tend to be rather similar to the original signal wave form. This can make the 'job' that the transceiver chip performs more difficult and when coupled with impedance mismatches between the chips brought on by the cable, can 'complicate' the transceivers 'job' of differentiating the signal from the noise.
Which in turn makes the transceiver chip 'work harder' at separating the 'real' signal from the rest of the electrical input it sees.
This 'work harder' aspect may be a significant contribution to the throughput performance in that it may contribute more power supply noise in the digital portion of the circuit.
Granted we are talking about small amounts of added noise but then the effects we are hear are also small.
Well unless these 'noise' influences become big enough to disrupt the flow of information and we get pauses, and hiccups etc.

So the 'quality' of the USB cable in terms of matching its impedance to the design expectations of the transceiver chips so they don't have to 'work harder', in the first place, may be a factor here as well.
This impedance matching applies not only to USB, but ethernet, AES/EBU, and other digital transmission line type of data delivery systems.

IOW the cable plugs into and connects two transceiver chips (one at each end) which both transmit and receive data packets interchangeably. How the chips 'deal with' the various electrical and signal characteristics, in real time, and especially in the USB realm where audio and video present very different requirements (ie. real time streaming) to what its original design intent was, also adds more complexity to this already complex situation.

So it may not be the cable itself but how well it plays with the transceiver chips that 'drive' the cable that interact with each other over the cable we use.

Just more food for thought.

JJ
ps transmission lines and what is involved, is a whole nuther complex subject unto itself.

 

[johnjen]

As I'm reading through these recent posts (and looking around at unused cables) I wondered for the first time what differences are built into interconnects that may share the same connectors but are marketed for different applications.

Are their differences between dual audio RCA interconnects and, say, those packaged in threes for composite video? Are they interchangeable back and forth, video to audio, or not at all?

Video cables are designed to handle the higher frequency range and narrower bandwidth than audio.
This isn't to say they won't work, but to work optimally, as in delivering the audio signal intact and with the least amount of 'change', probably not.

And depending upon the degree of resolution the audio system has, using video cables in an audio system may not make much of a change.
But then the one sure way to find out is to try them and listen to determine if they are satisfactory.
Or not.

JJ

 

[landroni]

Another aspect that I didn't bring up, mainly due to the pre-existing girth of part 7, was that of the transceiver chips at each end of a digital transmission line 'System' and their role as data 'interpreters'.

[...]
So the 'quality' of the USB cable in terms of matching its impedance to the design expectations of the transceiver chips so they don't have to 'work harder', in the first place, may be a factor here as well.
This impedance matching applies not only to USB, but ethernet, AES/EBU, and other digital transmission line type of data delivery systems.

 

 
In practice, how would one attempt to match impedance of transceivers and digital cables, say USB or coaxial?

 
I've never encountered impedance being stated as a spec for USB cables (although coaxial cables are often rated for 75 Ohm), let alone for digital receivers... For instance if I look at the specs page for the USB Input Receiver in Yggy, C-Media CM6632, there don't seem to be any impedance-related specifications.
 
PS On the other hand, Audio GD's Master 7 specs, indicate "0.5 Vp-p ( 75 Ohms, Coaxial )" as input sensitivity. It seems like coaxial cables/receivers are often better matched.

 

[johnjen]

In practice, how would one attempt to match impedance of transceivers and digital cables, say USB or coaxial?

I've never encountered impedance being stated as a spec for USB cables (although coaxial cables are often rated for 75 Ohm), let alone for digital receivers... For instance if I look at the specs page for the USB Input Receiver in Yggy, C-Media CM6632, there don't seem to be any impedance-related specifications.

PS On the other hand, Audio GD's Master 7 specs, indicate "0.5 Vp-p ( 75 Ohms, Coaxial )" as input sensitivity. It seems like coaxial cables/receivers are often better matched.

I have been looking into the impedance issue as it relates to USB and ran across this…
"Signaling (excerpt from the USB 'standards')
USB signals are transmitted using differential signaling on a twisted-pair data cable with 90 Ω ± 15% characteristic impedance."

Basically a USB cable is a differential signal pathway that acts like a 'tuned' load to the transceivers for them to 'properly' drive.
Sorta like matching the speaker impedance taps on a tube amp with output transformers.

My question, and one I doubt has been asked let alone answered is, will making the cables impedance match the 90Ω impedance (±1% not ±15%) make any difference.
I'm guessing that it will, and this spec may be a key as to why some cheap USB cables work 'better' than others.

And true enough I have not seen any cable manufacturers list their rated impedance for their USB cables.
Most likely due to the cost and complexity of dealing with this 'sticky' issue, which is not an issue if it remains 'unknown'.

JJ

 

[landroni]

My question, and one I doubt has been asked let alone answered is, will making the cables impedance match the 90Ω impedance (±1% not ±15%) make any difference.
I'm guessing that it will, and this spec may be a key as to why some cheap USB cables work 'better' than others.
 

Here's what usb.org tells us:
http://www.usb.org/developers/usbfaq#sig6
"What characteristic impedances should I design for in laying out my USB device's traces?
A: 30 ohms to ground, 90 ohms differential between the USB data lines. Be aware that many differential impedance formulas don't take into account the presence of a plane next to the data lines and tend to overestimate the reduction in single-ended impedance due to coupling between the two lines. For typical buried microstrips on spacings FR4 stackups the USB traces should be somewhere around 11 or 12mils wide with roughly a 33mil edge to edge spacing (45mils center to center). This is far enough apart that there's almost no coupling between the lines---the single-ended impedance is around 1.5O lower than what a single trace's impedance would be.
 
 
But the maximum allowable termination is 44 ohms! A 28 ohm termination is also within spec. Why?
A: USB termination requirements are quite lenient. For example, a 0 to 3.6V signal input into a 51.75 ohm line with a 28 ohm source termination will swing between 4.6 and -1.0V-this is where the voltage levels in figure 7-1 of the USB 1.1 spec come from. In practice, most USB devices are underterminated, sometimes incorrectly.
The IF recommends sizing the series termination resistors to create a 44 ohm termination. While not enough to perfectly terminate the device when its connected to an ideal cable, it does bring the terminations well within the standard ±15% impedance tolerance-most modern processes are well controlled, so the internal impedance of USB drivers is quite consistent from wafer to wafer. It is, of course, wise to characterize the drivers' impedance range and perform sampling of production parts."
 
Clearly there is such a thing as an ideal USB cable.

 

[johnjen]

I recently contacted Shunyata Research about this very issue and they kindly sent me some info for further research.
This was the document he referred to…
IEC 62680 USB 2.0 (Universal Serial Bus interfaces for data and power).
This was the text from a section he sent me…

Cabling (excerpt)
The data cables for USB 1.x and USB 2.x use a twisted pair to reduce noise and crosstalk.
The USB 1.1 standard specifies that a standard cable can have a maximum length of 5 meters with devices operating at Full Speed (12 Mbit/s), and a maximum length of 3 meters with devices operating at Low Speed (1.5 Mbit/s).

"USB 2.0 provides for a maximum cable length of 5 meters for devices running at Hi Speed (480 Mbit/s). The primary reason for this limit is the maximum allowed round-trip delay of about 1.5 μs. If USB host commands are unanswered by the USB device within the allowed time, the host considers the command lost. When adding USB device response time, delays from the maximum number of hubs added to the delays from connecting cables, the maximum acceptable delay per cable amounts to 26 ns. The USB 2.0 specification requires that cable delay be less than 5.2 ns per meter (192 000 km/s, which is close to the maximum achievable transmission speed for standard copper wire).The USB 3.0 standard does not directly specify a maximum cable length, requiring only that all cables meet an electrical specification: for copper cabling with AWG 26 wires the maximum practical length is 3 meters (9.8 ft)."

Signaling (excerpt)
USB signals are transmitted using differential signaling on a twisted-pair data cable with 90 Ω ± 15% characteristic impedance.

I have yet to search for the source document, it is on my list of things to do.

JJ

 

[gefski]

PS On the other hand, Audio GD's Master 7 specs, indicate "0.5 Vp-p ( 75 Ohms, Coaxial )" as input sensitivity. It seems like coaxial cables/receivers are often better matched.

My understanding is that cable-wise, it's real easy (& cheap) to make a coax with "perfect" 75 ohm impedance end to end, with BNCs on each end (like the one you had at the meet, JJ). But as soon as one uses an RCA adaptor or even a direct connection RCA, it gets "fouled up".

 

[landroni]

My understanding is that cable-wise, it's real easy (& cheap) to make a coax with "perfect" 75 ohm impedance end to end, with BNCs on each end (like the one you had at the meet, JJ). But as soon as one uses an RCA adaptor or even a direct connection RCA, it gets "fouled up".

Interesting. But what kind of source gear uses BNC? I've never seen BNC on a computer, DAP, TV...... Is this only the stuff of recording studios?

 

[gefski]

Interesting. But what kind of source gear uses BNC? I've never seen BNC on a computer, DAP, TV...... Is this only the stuff of recording studios?

Not many dacs do, though lots of external usb boxes (not usb "cleaners") have bnc out. I've got a Musical Fidelity vLink and a bel canto mLink that both do. Yggy and Gungnir have bnc in. Had a Metrum Octave here and it did also. But yes, if one is shopping dacs, you have to look at the back.

 

[landroni]

Not many dacs do, though lots of external usb boxes (not usb "cleaners") have bnc out. I've got a Musical Fidelity vLink and a bel canto mLink that both do. Yggy and Gungnir have bnc in. Had a Metrum Octave here and it did also. But yes, if one is shopping dacs, you have to look at the back.

Hmm, but what would you use to feed Yggy BNC? (In other words, what devices out there have BNC out?)

 

[bimmer100]

 
Hmm, but what would you use to feed Yggy BNC? (In other words, what devices out there have BNC out?)

if you must have BNC out than the Hydra Z would be an excellent choice to feed a BNC input dac
 
http://audiobyte.net/products/hydra-z

 

[gefski]

Hmm, but what would you use to feed Yggy BNC? (In other words, what devices out there have BNC out?)

In my case, I used the bel canto mLink that I liked with other dacs, but ended up liking the Yggy built in Gen 3 usb better, so the bel canto is now retired. Sometimes the dac's built in usb is better, sometimes an outboard usb is. But I don't know of much widely available gear that uses bnc. Didn't mean to sidetrack the discussion, but you guys were talking about coax & 75 ohm impedance, and I'm familiar with outboard usb. Yggy and most dacs have lots of different inputs.

 

[BIG POPPA]

@gefski , hey Glen, I have a BNC spdif converter you need to hear, hook it up to the yggi and enjoy. It is the JKeny. Just let me know and we can meet at Duck Island.

 

[gefski]

@gefski , hey Glen, I have a BNC spdif converter you need to hear, hook it up to the yggi and enjoy. It is the JKeny. Just let me know and we can meet at Duck Island.

I wanted to hear the JK stuff at the meet but it was wandering around the room. But I'm so tired of Windows driver hassles that I'm not trying any new usb interfaces or dacs on my PC, and am trying out a MAC that my wife handed me down (no drivers needed). So it will be a while before I do so.

Duck Island shouldn't be put off though. I like mid to late afternoon arrival so I can get one of the car seats. PM me anytime.

 

[BIG POPPA]

Okay noob's listen closely, we are having a secret meet at Duck Island, you need to pm me. After the New Year. Glen, Duck Island sounds mighty fine. Who knows who will show up and what we will bring. Portable Gear recommended.... We will not plug in there...... Tap in absolutely.........