I have been looking at the waveshape of various 75R cables terminated with correct BNC plugs from the cd digital output of the Guido Tent XO3, treating it as a TDR.
First, all my my high end cables of 1m length or slightly less give very good square waves into my scope without any soecial termination. The differences were confined to the corners of the square waves and the 'slight distortions therein'. Terminated with a 75R BNC adaptors, the waveforms were essentially perfect, with slight corner overshoot and very small ripples on the top hat. The cables were Essemble DIGIFUX, Aural Symphonics, Mark Leveinson Silver of many years ago, and several others.
Changing to Belden 1550A RG59 5m and 1.5m lengths, I have a bad step about 3/4 way up to the top hat, and significant ringing on the top hat sevtion. A Canare 1.5 m factory terminated cable gave a less severe discontinuity at the rising and falling edges. This is without the 75R termination. With the 75R terminator, I am getting more ringing than the shorter 'hifi' cables but otherwise satifactory waveforms.
Listentening to the cables, neither the Belden or the Canare have that special something of the branded 1m or 1.5m cables.
Can someone clever please tell me why I am finding these waveshapes? The 5M Belden should, as I understand it, give me the much better wave shape and less serious reflections
First, all my my high end cables of 1m length or slightly less give very good square waves into my scope without any soecial termination. The differences were confined to the corners of the square waves and the 'slight distortions therein'. Terminated with a 75R BNC adaptors, the waveforms were essentially perfect, with slight corner overshoot and very small ripples on the top hat. The cables were Essemble DIGIFUX, Aural Symphonics, Mark Leveinson Silver of many years ago, and several others.
Changing to Belden 1550A RG59 5m and 1.5m lengths, I have a bad step about 3/4 way up to the top hat, and significant ringing on the top hat sevtion. A Canare 1.5 m factory terminated cable gave a less severe discontinuity at the rising and falling edges. This is without the 75R termination. With the 75R terminator, I am getting more ringing than the shorter 'hifi' cables but otherwise satifactory waveforms.
Listentening to the cables, neither the Belden or the Canare have that special something of the branded 1m or 1.5m cables.
Can someone clever please tell me why I am finding these waveshapes? The 5M Belden should, as I understand it, give me the much better wave shape and less serious reflections
Without actually seeing the wavefroms, I can not tell you how to interpret them.
Usually on a TDR, the close-in reflections are a mismatch at the interface. One reason for using a longer cable would be to accurately determine the Z of the line, and observe the anamolies at the termination end; as opposed to the source end where they can be easily obscured.
Jocko
Usually on a TDR, the close-in reflections are a mismatch at the interface. One reason for using a longer cable would be to accurately determine the Z of the line, and observe the anamolies at the termination end; as opposed to the source end where they can be easily obscured.
Jocko
RG59 and RG59A are no no for S/PDIF. These are good for analog standard video signals up to 5 MHz but have too much loss at higher frequencies. If you intend to use a low cost 75 ohm cable use 75 ohm CATV or Satellite TV cable with foamed isolation and solid core, preferable double screened. I myself have good results with Belden H121.
😎
😎
HUH????
And just how much loss is there at "the higher frequencies"?
The problem may not be one of loss................
Jocko
And just how much loss is there at "the higher frequencies"?
The problem may not be one of loss................
Jocko
Hmm, you are right. Just looked up the 1505A at belden.com I am not familiar with that particular cable. Regular standard RG59 is usually not suitable for HF.
But this Belden 1505A looks pretty good. Right Jocko, there must be something else wrong in fmak’s test set-up. Maybe a loose connector?
But this Belden 1505A looks pretty good. Right Jocko, there must be something else wrong in fmak’s test set-up. Maybe a loose connector?
You may want to try the Canare LV-61 cable with this set-up also. It's rated for 75 ohm use, and is excellent for digital connections.
Cheers,
Zach
Cheers,
Zach
I doubt that his setup is the problem. But I would not use a transformer in a TDR type arrangement.
Having spent more time measuring, testing, and listening to cables more than anyone here.....except Phred.....I can safely say the following:
I have found no reliable way to correlate how a cable will sound as to how it measures. Part of this is due to wide variances in how close to 75R the actual Z of the transmit and receive unit are. Each cable will react differently depending how close all three are to actually being 75R.....and pure resistive.
Once you get both ends to be 75R without a reactive component, all cables start to sound alike. Again, ones that aren't will sound not so hot, but exactly how they will sound is not predictable.
Jocko
Having spent more time measuring, testing, and listening to cables more than anyone here.....except Phred.....I can safely say the following:
I have found no reliable way to correlate how a cable will sound as to how it measures. Part of this is due to wide variances in how close to 75R the actual Z of the transmit and receive unit are. Each cable will react differently depending how close all three are to actually being 75R.....and pure resistive.
Once you get both ends to be 75R without a reactive component, all cables start to sound alike. Again, ones that aren't will sound not so hot, but exactly how they will sound is not predictable.
Jocko
TRANSMITTING TO EARTH....
Hi,
Neither have I, be that for analogue or digital connections.
Even when respecting all possible Xmissionline parameters I can't help but still hearing cable differences...granted, you only hear it on high resolution systems but I'm positive I don't imagine this.
On the bench I've not found anything I can accuse of being the culprit yet...maybe FFT analysis might help?
Truly puzzling....
Cheers,😉
Hi,
Neither have I, be that for analogue or digital connections.
Even when respecting all possible Xmissionline parameters I can't help but still hearing cable differences...granted, you only hear it on high resolution systems but I'm positive I don't imagine this.
On the bench I've not found anything I can accuse of being the culprit yet...maybe FFT analysis might help?
Truly puzzling....
Cheers,😉
Frank:
(The one in London.......)
Since your source (Guido's clock) seems so sensitive to load, you may want to put a "L-pad" in series with it and the "T". Try 6 dB......10 dB, whatever 2 resistor values that you can come up with to make a 75R source.
Most TDRs work with only a volt or so, you should still have enough level to work with after padding the clock.
Make sure that the pad goes between the source and the Tee. You will get bogus results if the pad goes between the Tee and the load.
Jocko
(The one in London.......)
Since your source (Guido's clock) seems so sensitive to load, you may want to put a "L-pad" in series with it and the "T". Try 6 dB......10 dB, whatever 2 resistor values that you can come up with to make a 75R source.
Most TDRs work with only a volt or so, you should still have enough level to work with after padding the clock.
Make sure that the pad goes between the source and the Tee. You will get bogus results if the pad goes between the Tee and the load.
Jocko
Digital cables sound different?
"Even when respecting all possible Xmissionline parameters I can't help but still hearing cable differences...granted, you only hear it on high resolution systems but I'm positive I don't imagine this."
That's same reaction I got from about three or four hundred people that bought my digital cables....... Mechanical damping of the cable is a big part of sonic differences as well as the dielectric, attention to termination, direction, and brand of RCA connector (with an impedance compensation network I designed) I never heard a Teflon cable that I liked, but several foam dielectric cables sounded great. One of the most absolutely fatiguing, but detailed, cables was some laboratory grade coax from Gore. It was $180 a foot as I recall and the was the actual price and not some audiophile mark up. Cable with losses at very high frequencies also often sound better than extremely high bandwidth. The upper bandwidth for the logic driving most digital interfaces is in the 30 to 100 MHz range. Stuff outside this range is noise and causes problems. The best soundind digital interfaces I have heard are in this region and the ones with pulse transformers were at the low end of this range and sounded the best. The low frequency end of the interface doesn't need to much lower than around 50 kHz as I recall. I wouldn't make cables any shorter than a meter and a half. Half meter cable usually sound awful.
"Even when respecting all possible Xmissionline parameters I can't help but still hearing cable differences...granted, you only hear it on high resolution systems but I'm positive I don't imagine this."
That's same reaction I got from about three or four hundred people that bought my digital cables....... Mechanical damping of the cable is a big part of sonic differences as well as the dielectric, attention to termination, direction, and brand of RCA connector (with an impedance compensation network I designed) I never heard a Teflon cable that I liked, but several foam dielectric cables sounded great. One of the most absolutely fatiguing, but detailed, cables was some laboratory grade coax from Gore. It was $180 a foot as I recall and the was the actual price and not some audiophile mark up. Cable with losses at very high frequencies also often sound better than extremely high bandwidth. The upper bandwidth for the logic driving most digital interfaces is in the 30 to 100 MHz range. Stuff outside this range is noise and causes problems. The best soundind digital interfaces I have heard are in this region and the ones with pulse transformers were at the low end of this range and sounded the best. The low frequency end of the interface doesn't need to much lower than around 50 kHz as I recall. I wouldn't make cables any shorter than a meter and a half. Half meter cable usually sound awful.
Re: Digital cables sound different?
Hi
This to me is the reason that I do not use higher bandwidths at the driving side either.
I once informed at Crystal about the input bandwidth of their 8412 receiver. Never got an answer. I then asked for gm of the input FETs used (one can recalculate BW and noise from that). No answer as well.......
regards
Fred Dieckmann said:
Hi
This to me is the reason that I do not use higher bandwidths at the driving side either.
I once informed at Crystal about the input bandwidth of their 8412 receiver. Never got an answer. I then asked for gm of the input FETs used (one can recalculate BW and noise from that). No answer as well.......
regards
Don't make me draw a picture......
It will look horrible!
Anyway......put the L-pad on the source. Run a short cable to your 'scope. Put the Tee there, and connect the cable you want to look at on that point.
Is that better now?
Jocko
It will look horrible!
Anyway......put the L-pad on the source. Run a short cable to your 'scope. Put the Tee there, and connect the cable you want to look at on that point.
Is that better now?
Jocko
Re: Don't make me draw a picture......
Correct, in fact you don’t need a correct termination resistor at the transmitting side of a coax cable, only at the receiving end when it comes to reflections.
But I must admit this is audio with lots of sonic magic.
😉
Jocko Homo said:
Correct, in fact you don’t need a correct termination resistor at the transmitting side of a coax cable, only at the receiving end when it comes to reflections.
But I must admit this is audio with lots of sonic magic.
😉
Hi Folks
This is an honest question, because I really want to know, not an attempt to diss this discussion.
What difference does the wave shape actually make? Surely as long as it is sufficiently above the noise floor to be recognised as a bit, or positive going waveform, the actual shape is not related to the final output?
This is an honest question, because I really want to know, not an attempt to diss this discussion.
What difference does the wave shape actually make? Surely as long as it is sufficiently above the noise floor to be recognised as a bit, or positive going waveform, the actual shape is not related to the final output?
I am somewhat surprised about the results, but nevertehless will redo some measurements at my side
------------------------------------------------------------------
Guido et al
Thanks for the interest. I seem to have partially resolved the problem. Placed a variable resistor at the output of the XO3 in place of the 150R series resistor and adjusted for best waveform with 5m Belden. This time, I can move the discontinuity up the leading and trailing edges and minimise the 'distortion' to make the square wave look very respectable. This turns out to be 91R. Ran the cable into a 75R termination at the scope and got an even better one. My 1m cables still look very good except for a little overshoot. The voltage with the 75R termination is now almost 0.5 and not 0.42 before.
There is , however, one problem. My CS8414 inout receiver will not now lock with the 5m cable, but will lock with others. The fact that the connectors are sound in the scope measurements may mean there is something else and I shall investigate.
Guido, any ideas on why. The cable has about 300-350 pf of capacitance. With the 91R series resistor, I am now getting about 40% of the rail voltage, not 50% before.
------------------------------------------------------------------
Guido et al
Thanks for the interest. I seem to have partially resolved the problem. Placed a variable resistor at the output of the XO3 in place of the 150R series resistor and adjusted for best waveform with 5m Belden. This time, I can move the discontinuity up the leading and trailing edges and minimise the 'distortion' to make the square wave look very respectable. This turns out to be 91R. Ran the cable into a 75R termination at the scope and got an even better one. My 1m cables still look very good except for a little overshoot. The voltage with the 75R termination is now almost 0.5 and not 0.42 before.
There is , however, one problem. My CS8414 inout receiver will not now lock with the 5m cable, but will lock with others. The fact that the connectors are sound in the scope measurements may mean there is something else and I shall investigate.
Guido, any ideas on why. The cable has about 300-350 pf of capacitance. With the 91R series resistor, I am now getting about 40% of the rail voltage, not 50% before.
- Status
- Not open for further replies.