Jocko insisted that cable length is critically important...think reflections, cable signal velocity.
This could explain subjective differences.
Dan.
Sure, both, but it is illustative pictures, that cables are different.
Effect similar to what is pictured on Acoustic revive web:
Digital Cable(DSIX-1.0PA/DSIX-1.0BPA)
I've decided to look on oscilloscope after I've found how big difference made in sound with just swap of cables.
All are equally (un)important. What I mean is that for perfect transmission they all have to be right. Fortunately, SPDIF is sufficiently robust that perfection is not needed. The sort of length of cable used in domestic circumstances is too short to create any problem other than a tiny amount of HF jitter which is then removed by the PLL in the receiver.Sergelisses said:
All this seems to be straining at gnats and swallowing camels. How many people, when swapping an RCA for a BNC, ensure that 75R impedance is maintained throughout the connection? In fact, how many people who do this swap know how to do this or how to measure it afterwards? I say this because most people who would know how to do this (i.e. they understand how transmission lines work) also know that for SPDIF it is not necessary. If the original manufacturer wired the RCA in a way to partly compensate for its impedance not being 75R, then simply replacing it with a 75R BNC would make things worse unless the compensation was recognised and removed.
Nope, shorter is always better, reflections again should be sorted out if detrimental by some sort of termination or by reducing the drive strength of the signal. In fact with a properly matched cable length should not make any difference, within certain boundaries.
The difference is very minor and would have NO effect whats so ever on the resultant signals. It shows some minor slowing of the rising edges, no great deal and not enough to change things for this sort of interface.
Last edited:
I have said numerous times that most PCB's will not be laid out for 75 ohms nor will the PCB's be bought as controlled impedance PCBs, so at the very best expect +/- 20% and probably more deviation from the 75Ohms required, but is doesn't really matter for what is effectively slow Manchester encoded signals.
Master clock jitter. Poor 75R line driver in source creating cable jitter. Grounding issues. Placebo effect. Expectation bias. Possibly but unlikely, timing issues in error correction logic if no FIFO output buffer used. Any pre-DAC digital filtering or upsampling in source.Sergelisses said:
Definitely there will be the difference. All is jitter-related, as cheap cables do not provide stable 75Ohm and it is very important for wave transmission - it is not as with analog audio. If you can go fo BNC - it's even better, as BNCs are more precise connectors in terms of 75Ohms.
I've compared D-75 vs cheap cables - TechLink and Hama with SPDIF out. On oscilloscope you can see that D-75 have more square front vs cheap cables (they rounded). I've posted those pictures here http://www.diyaudio.com/forums/parts/220756-looking-silver-digital-coax-3.html As usually "cable thread" splitted into believers and skeptics
Don't worry xjr100. I'm used to the wire naysayers mocking and criticizing most everything I say and believe about wires. However unlike most wire aficionados I have openly posted here in diyAudio.com that I'm willing to submit to having someone test my ability to detect different analogue ICs.
Sy has stated he'd be open to running such a test with my following stipulations:
- I MUST be intimately familiar with the audio system that we are to use.
- All wire changes MUST be made manually by one person physically replacing one set of ICs with another set of ICs by hand. By switching ICs manually we will be keep the system as simple as possible and it will remain the system that I am intimately familiar with sonically. Thus the only possible sonic changes I would hear would have to be a result of the wires being used.
- I get to choose the music and the volume used for testing.
- Someone I know well & trust will stay with Sy at all times during the test. While someone Sy knows well & trusts will stay with me at all times. This is just to keep things honest.
- If there's anything else I or Sy would require be added or if there's anything I or Sy require that one of us takes issue with. We would need to discuss and agree upon any changes before they'd be implemented and the test is run.
- The final results would be published whether I pass or fail the testing.
So now it comes down to the not so simple matter of synchronizing a time when Sy and my schedule are both open! Being disabled I'm available most of the time. However Sy must first find the time available to take from his regular day-to-day job! And then Sy must find a time he can take from his family activities. Perhaps people are now starting to understand what a nightmare getting this testing started really is.
Yet despite all these issues. Sy and I are determined to run this test and prove once and for all whether or not I can detect differences in analogue ICs and eventually we will do just that! And the best part of all this testing is the results whether they be in favor of the wire naysayers or in favor of the wire believers, will be published here by Sy!
Thetubeguy1954 (Tom Scata)
Hello Max! I've read somewhere that 1.5M is supposedly the ideal length for a digital cable precisely for the reasons you stated above!
Thetubeguy1954 (Tom Scata)
Marce,
it seems that double sided PCB can give ~100ohms for a trace pair.
What could be done to vary this?
and in particular to achieve 75r & 50r for the trace pair?
If a twisted pair were brought to the edge of a 100r pair of traces, does one simply attach one core to top and one core to bottom to maintain a reflection free transfer onto the PCB?
What would be required to bring a 75r or 50r wire pair, or coax, onto a pcb?
The word was to use TDR to determine exact ideal cable length....makes sense.
Dan.
Do you have any citations to back that up marce?
I actually believed the very same thing you do, until I read an article about why 1.5M is actually the ideal length when using S/PDIF coax. The article is by Steve Nugent an engineer with numerous patents. That's posted in Positive Feedback. It's called: "PFO Audio Discourse: Why longer is generally better for an S/PDIF Digital Cable."
You can read the article here: spdif In the article Steve provides the technical reasons why one should use S/PDIF cables of 1.5M. Of course there are other issues that Steve gets into in the article, but he explains why 1.5M is the length to use!
Thetubeguy1954 (Tom Scata)
Interface Jitter Reduction...
The trick is to time reflections transition timings away from valid data transition timings.....makes sense.
Dan.
His assertion IIRC was that typical systems have lousy impedance loading/matching, hence typically multiple signal reflections are well evident.
The trick is to time reflections transition timings away from valid data transition timings.....makes sense.
Dan.
Yes, I spend all day trying to get digital signals from one place to another, I do high speed layout as well as RF and all other sorts of layout. I work on numerous high end systems including aerospace, military and medical, I regularly go on site to help and advise companies on layout, signal and power integrity. Recently I spent 4 years of my life being responsible for the layout and SIV a 2.4 billion military project, so I work with this every day, I see the problems of all aspects of signal transmission every day (analogue, digital, RF) mainly for high end products and I am responsible for the layout and signal integrity (as well as power integrity) so I know what it takes to get a signal from A to B.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.