Jocko was/is insistent that Spdif coax cables need to be specific length for clean transfer of Spdif data.
I forget the details but IIRC it is all to do with cable Vp and avoiding reflections coinciding with transmissions.
Dan.
Okay. But, that would appear to only apply when there are significant reflections. If the cable is properly terminated at the receiving end, there shouldn't be significant reflections. Is there some reason we should expect the termination quality to be poor? Or, is there something else I'm missing?
Mark, I'm not sure about the digital thing. This was some time ago, and I've been using 6ft cables since. I've seen mentions of jitter in association of SPDIF cables under 1.5m, where reflections are additive, and on longer cables not. But I don't specifically recall that being the original reason.
My example with the Ncores is just that sometimes short's innate qualities are not best, as sometimes a little R/L stops resonations/oscillations. In the case of tweeters and active crossovers there might be some protection depending the amp. But really unless you got finicky gear the sound difference in cable lengths with a few feet, ten foot? Uhuh...
My example with the Ncores is just that sometimes short's innate qualities are not best, as sometimes a little R/L stops resonations/oscillations. In the case of tweeters and active crossovers there might be some protection depending the amp. But really unless you got finicky gear the sound difference in cable lengths with a few feet, ten foot? Uhuh...
I can't come this exact second but if you reply to my email, then it certainly could happen. It would be nice to bring some toys to play with... do you have other speakers than just M2's? I know you don't have a TT anymore....
RCA is 'standard' connector for Spdif and is undefined impedence, plugs and sockets both.
Canare does make a '75 ohm' RCA plug but I am not aware of matching sockets.
Dan.
Jocko is a smart guy, and his advice on SPDIF was technically sound, but jitter and SPDIF issues are way overblown with the latest generation of SPDIF receivers like the WM8804/5 and especially ASRCs. Now I would just use optical/Toslink for isolation and run it through an ASRC if I had to use SPDIF. I am sure he is rolling his eyes at this post if he still reads this thread, but I just don't think it's an audible issue. It might have been a more easily measured problem back in the days of the crap Yamaha, Philips, Cirrus, etc. receivers when you were clocking something more sensitive to jitter as well.
Last edited:
Sure, in some cases it would make sense. Depending on the circuitry, that is. But, for interconnection standards supposedly designed to work with a range of cable lengths, one would expect it to work as specified. Otherwise, something is broken. And that can happen, but I'm not sure that all kinds of things that in principle ought to work okay, are in reality broken when it comes to audio. Some might be, but some of the claims of problems and solutions kind of seem broken too.
One that I always have trouble with has to do with AC cables. In some cases, when a lot of equipment is plugged in the same power strip, maybe. But, when the AC cable plugs into 75' of Romex going back to a breaker panel, then the contribution of the AC cable relative to the Romex seems rather small.
It's lower than 75 ohms, I have seen 30-40 ohms mentioned on audio forums before, but never with a measurement. Even if the connectors are generally 30 something ohms, it's not a controlled impedance. If you are actually concerned you should just use a BNC or other RF connector and appropriate coax.
How much ghosting do you see with one of those 25 cent video cables that still comes with every DVD player and TV?
Several other points he electrical length of the connector is also important (it's picoseconds) most likely buried in the rise time of the signal in this case so the connector is a TINY L or C discontinuity. If either end is carefully terminated the length can't matter. This is not 40gig backplane stuff.
Understood. But, I am trying to have a conversation and hopefully see where some of the claims came from. I'm not trying to teach anything now. Maybe some mutual agreement can be reached without any need to go there, I don't know. If there were a need and/or a desire to develop more theoretical intuition, probably much easier to do in person with some pencil and paper, rather than though forum posts.
Otherwise, it can be very hard to reach agreement on something when one side pulls rank and basically starts saying, "I know way more advanced math and theory than you do, and you are wrong and I am right, end of story." Things like that can make people stop listening and start thinking the other person is not trustworthy. And, obviously, things can deteriorate further from there.
Very easy for misunderstandings of other's people intentions to occur in internet forums. Just the way it is. You can't see facial expression or hear tone of voice. And words by themselves can be interpreted in many different ways, especially as people try to read between the lines to discern perceived intent of the writer.
Last edited:
I remember weakly having calculated a specific connector while studying an it was around 50 Ohms if i recollect it correctly.
WBT, having a patented design with their Nextgen series realized 75 Ohms (plug and socket), once wrote that the range of dimensions in rca connectors (and the diversivity of materials i´d say) allows a range from 16 - 45 Ohms. (given that i was "slightly" off with my calculations back then
) .
I would hope more folks than you think come to this forum to actually learn some electronics. This is EE101 stuff it is not relative or subject to opinion. What happens when you have a 1/4" impedance discontinuity in a transmission line is easily estimated. If you say the say the signal is reflected just as if the line is mis-terminated you are wrong.
If you choose to paraphrase a simple statement of basic engineering that way it says more about you than me.
Scott, Yes, I know. But apparently there were some threads on this very forum where a case was made that reflections could cause errors with SPDIF, and the suggested remedy was to control cable length in order to control reflection timing relative to edge rise time at the receiver. Without reading through all of that, I don't know what the story was. In the meantime, I am allowing that there may have been some problem at one time, I just don't know. That's what I thinking about, not whether EE 101 calculations are trustworthy.
In addition, there are some engineers who barely graduate and only do electronics as a day job. Couldn't care about it less off the job. Some technicians have seen such engineers jump to conclusions and pull rank, and then turn out to be wrong. People who have seen that enough times might not know that Wucer never makes a mistake or jumps to a conclusion.
In addition, there are some engineers who barely graduate and only do electronics as a day job. Couldn't care about it less off the job. Some technicians have seen such engineers jump to conclusions and pull rank, and then turn out to be wrong. People who have seen that enough times might not know that Wucer never makes a mistake or jumps to a conclusion.
Yet again, a mix of reality, and deep misunderstanding of basic engineering principles.
A transmission cable, that is (impedance) unmatched at one or both of the ends, does indeed have amplitude maxima and minima along it's length; for an ideal cable (no losses) and total mismatch the minima are theoretically zero and the maxima 2*Vin.
The spatial distance between two maxima or two minima is Lambda/2 and the distance between a maximum and a minimum is Lambda/4, where Lambda is the wavelength of the propagating signal in the cable. If the unmatched cable has a length exactly a multiple of Lambda/2, then theoretically you could get along with an unmatched cable.
However, this is true only for a pure sine signal. Now, if the signal is not a simple sine, but say a digital pulse, with lots of harmonics, what is going to happen is the waveform at the end of the unmatched transmission line is going to be distorted. This is going to happen even if the cable length is exactly a multiple of Lambda(fundamental frequency)/2.
But all these are happening in the absurd case of a coax that is not matched. In all practical cases where we use coax (cable TV, CD, DVD, etc...), the cable is matched at both ends. When a transmission line is matched at both ends, the signal amplitude is the same all along the length, so the cable length (barring the cable losses) doesn't matter.
If the signal quality depends (say, "audible") on the cable length, then either the transmitter or the receiver, or both are defective, or the cable itself is defective.
Note1: Lambda above is the wavelength in the cable, which is different from free air/space. Lambda=v/f where v is the propagation speed in the cable material, approximatively v=c/SQRT(Epsilon) where c is the speed of light, and Epsilon is the relative permittivity of the cable dielectric.
Note2: The matching condition requires the source, load and characteristic cable impedances to be equal. Choosing a 75ohm characteristic impedance cable and trying to match it with 50ohm source/load impedance will result in maxima and minima, so a signal amplitude variation across the cable length. This is a common mistake, using an improper coax cable for a certain source/load impedance (example: using 50ohm coax for TV signal distribution (75 ohm) will create ghost images in analog TV and image noise in digital TV).
- Status
- Not open for further replies.