[jneutron]

Pull the other one. The actual rise time is irrelevant. What is relevant is that despite your assertion to the contrary, the pictures you referenced in support of your 'inductance' example show no sign of an exponential decay.

w

I think your a tad confused.

What are you talking about?  I've posted Pat's first pic, which has been shown to be inductive.  And I've posted Joe's pic, which is active node.

John

While your edits to try to denigrate are humourous, it is exactly that attitude that gets you in trouble.  Try civility please.

[wakibaki]

I've posted Joe's pic, which is active node.

Yes. And depending on the Q of the circuit, the recovery will be exponential, or in this case the linearity indicates that the stray inductance is trivial.

And when I said this before, you said:

While you seem to understand t-lines, you've obviously never worked with 200 pSec risetime power circuitry.

But we're not talking about 200 pSec risetime power circuitry. Are we?

And why all this matters is because you said:

With this in mind, consider what the output will look like if R2 is inductive.


Where will that second spike be in relation to the initial waveform?

Inquiring minds want to know...

And the fact is that it makes no difference whatsoever to the arrival time of the reflection.

Now you go on to say:

The spikes pass through without delay.  A returned reflection will do exactly that.  Remember your calc of a 10 nSec double transit with a 3 to 4 rise??   Will the active circuit have settled in 10 nSec?

And this is a good point, and it gets art off the hook as regards the 1 metre cable, if I'm generous.

But it has no impact whatsoever on the main thrust of my argument with regard to longer cables. Which is that returning reflections @ 44k1 have 177nS in which to land, that they are not themselves normally of a magnitude to be mistaken by the receiver as edges (which would make itself known as something more profound than jitter), that even if the duration of the reflection is 75nS (50 looks a lot fairer to me, given that the reflections in art's scope pictures have only reflected off the DAC), and that they in all probability would be reduced in magnitude if not eliminated entirely by encountering the transmit termination, this still allows for a cable length of at least 10 metres before the reflections start to interfere with the next edge.

And in fact if you repeat the estimates at 96k a cable length of 3 metres can still be accommodated without incurring interference.

There is a further issue, concerning the other effects of attenuation.

I have produced documentation (in this thread) which quantifies the increase in jitter due to thermal noise with slew rate reduction. Now, it turns out that the actual increase in jitter as finally calculated by JosephK is pretty minor. Femtoseconds. I was, however, attacked and vilified for even suggesting that attenuators might increase jitter in this way. At no time did either of the protagonists (art and JosephK) ever investigate such a possibility of their own volition.

The increase in jitter with thermal noise documented in the app. note refers to comparatively normal operating conditions. When the signal is reduced to marginal levels just above those where the DAC fails to lock up, other mechanisms affecting the level of jitter may come into play. No effort has been put into investigating this possibility.

Given that most SPDIF chip manufacturers reccomend a simple resistive termination which is pretty hard to get wrong, while recognising that it is possible to create a termination which measures 75R resistive at DC but not at AC, I still nevertheless recommend that, given the known fallibilities of human judgement in sighted auditions, anyone discovering 75R with a DMM at the input to their DAC, and not being posessed of skills in electronics, or anyone running 44k1 or 48k in any cable length within the specification, or 96k in a cable under 3 metres should find some outlet for their creative energies other than experimenting with attenuators.

w

[wakibaki]

Waki it seems to me that you are the one expending an awful lot of energy. Have you ever tried these attenuators? I presume if you have an inquiring mind you might just want to see if all you say holds true in practice or is it just all text book knowledge? You are so sure of yourself & busy trying to save others from themselves that maybe a little cheap experiment in the real world would not go astray? 

Waki it seems to me that you are the one expending an awful lot of energy.

Yes. Is that wrong? I enjoy doing so. It's my energy to expend. I may not have much time left. This is true for all of us, but the imminence of death focussed my mind. Perhaps you will feel the same if (God forbid) you are ever in the same position. I feel this is a good way to employ myself. God knows I can't get a job, I've had 2 interviews since I turned 50, and none since I was taken ill.

More answer than you wanted perhaps?

Have you ever tried these attenuators?

No. Why would I? If I had a system that I thought was suffering from reflections, I have recourse to more satisfactory techniques. You mean, make a degraded system and then patch it up with attenuators to get it to work? To see if I can hear the difference? No.

I'd be happy to listen to a system and see if I can hear a difference, but make an effort myself? No. I wouldn't want to feel under any pressure either. You know what I think of all this stuff though; a sop for people who can't play.

I'm not the one who thinks attenuators are a good idea. They're not a good idea. A correct termination is a good idea. And easy to achieve if you don't **** around with video amplifiers. Maybe I will design a DAC when I've finished these touch-sense attenuators.

I presume if you have an inquiring mind you might just want to see if all you say holds true in practice

You are correct in that I have an inquiring mind, that pretty much delimits what you know of me. You presume in error, and your behaviour is by definition, presumptious.  What I say is based on my experience. I'm not the kind of person who has to keep going back to check if I turned off the gas-taps.

just all text book knowledge?

Just text book knowledge? I don't know what you mean.

You are so sure of yourself

Is that something you aspire to? Perhaps if you got enough education to participate meaningfully in this discussion you would be more self-confident.

& busy trying to save others from themselves

Evidently you don't understand why a person would feel like that. That's your loss.

maybe a little cheap experiment in the real world would not go astray?   

I can think of an experiment. Put increasing attenuations in an SPDIF cable. See if the increase in jitter in the recovered clock in a range of SPDIF receivers remains linear when the signal level becomes marginal.

You know, the experiment that should have been done before this whole business of attenuators was trotted out in public. I shan't be doing it though, I'll confine myself to pointing out that it hasn't been done. I have  to conserve my energy you know.

w

[Joseph K]

You have posted the manufacturer recommended input circuits. Fig. 17, AK4113.

This is the reference response from my TDR, a ~2nsec long 75ohm cable, bias tee (75ohm) + 75ohm termination:


Vertical is ~200mrho, horizontal is ~5nsec. That little <1nsec blink is the 75ohm! bias tee. (Everything is ~ because of the (small) conversion issues while registering / plotting this >30 years beautiful technology)
There is a glitch at 5div height, that is the original 50ohm incidental signal.
The top is the reflection on the 75ohm cable. As is visible, the cable fits very well the termination impedance. The termination impedance measures 74.9ohm on a DMM. 

Now, here it is that .47nF, 470pF shunt capacitance (fig 17)added to the bias tee:



The generated reflection is 25nsec wide..
So much for the "easily satisfied 75ohm input impedance measured by a DMM.


Now, another nice circuit, which You had recently cobbled together:

http://www.diyaudio.com/forums/digital-source/183536-spdif-splitter-2.html#post2479522

Is showing ~94ohm input impedance.



Calculate Yourself: Full incident edge is 390mv, the upper trace here is the reflection with power OFF, that is, 100 ohm input impedance, the lower trace is with the Eliot receiver powered ON. Height is 44.5 mV.
Equals to 94ohm input impedance, reflection caused is 11.4 %.
Yes, you are right, it is not 15% initially suggested by me. My wrong.

Another 75ohm input design, this time by yourself?
By the way, that Eliot receiver does not work. I had to modify to make it work.
I knew it already before, but anyway had put together strictly as You had suggested, just to see how bad is your suggestion.

Maybe You could tell why is it so? 


 

[jkeny]

Waki,
I do feel for your situation in life, the little bit I know of it. I can understand some of the relentless anger & striking out is as a result of this situation. Unfortunately, it makes it difficult to communicate with you with the emotional load that you bring to this communication. I won't react to you as I have done in the past.

You have proven that you are a learned guy now prove that you are a smart guy. (EQ is just as important as IQ, if not more so). Try to excercise some restraint & perhaps listen a bit more as a number of people here have suggested - I'm sure it's not too late to learn some new knowledge! 

[wakibaki]

Now, here it is that .47nF, 470pF shunt capacitance (fig 17)added to the bias tee:

As is visible, the cable fits very well the termination impedance. The termination impedance measures 74.9ohm on a DMM.

w

[Joseph K]

As respect to that note with DMM, I just could not leave it out..

AKM: "In this case, it is possible to lower the coupling level by adding..."
And raise the reflections in that line by, uhm, 1000%? 

[Joseph K]

About that DMM-note, again:

Here it is a second "75ohm" cable with a second "75ohm" termination.


I have the tools to tell them apart. How about You & the rest?
Please note: it's an "RG59U silver plated cable" and on the termination is written: 75ohm.

(Hint, not for You, but all you others: attenuators..)

[Joseph K]

And we did not start about those transformer-coupled transmitters..

[wakibaki]

Maybe You could tell why is it so?

10n should be 100n.

Yes, you are right, it is not 15% initially suggested by me. My wrong.

30% is what you suggested. Do not overegg the pudding.

Yes, I copied the Elliott circuit without thinking.

As is visible, the cable fits very well the termination impedance. The termination impedance measures 74.9ohm on a DMM

You prove my point, the 470pF is only fitted in unusual circumstances as is shown in the note.

Yes, misterminations occur. What is at issue is their significance, and whether it is appropriate to recommend attenuators as a general solution.

And raise the reflections in that line by, uhm, 1000%?

I have to agree. Recommendation?

Please explain the significance of the last pic.

And we did not start about those transformer-coupled transmitters..?

Regardless, the reflections will not coincide with the edges @ 44k1 in a cable under 10 metres.

w

[Joseph K]

100N - I have tried, not enough. And is not the right direction.

Edit: with the right correction, it works with the 10nF - it is what is in the scope shots.

[Joseph K]

30% is what you suggested. Do not overegg the pudding.

Here it is the thread, with your answer, so You can not even say that you had not seen it:

http://www.diyaudio.com/forums/digital-source/183536-spdif-splitter-6.html#post2481357

[wakibaki]

Yes, you modified it to 15%, I had forgotten.

w

[Joseph K]

I was, however, attacked and vilified for even suggesting that attenuators might increase jitter in this way. At no time did either of the protagonists (art and JosephK) ever investigate such a possibility of their own volition.

Strange..

From the Diyhifi thread, ~9month ago, and at the second post of mine ever talking about this:
"Ok, to make it clear: I have only brought up it now, because the Hiface is the special case, where it's really relevant.
Because of the high output level. Padding in the case of standard levels, 500mVpp, only makes it worse, because it makes life even more difficult for the already very critical input receiver sections.

"http://www.diyhifi.org/forums/viewtopic.php?p=43407#p43407


Then, from this present thread:

"Also here you had picked up only a part of  the original message. All this debate started by saying that attenuators could do good in a system with one special unit, the Hiface.
Where the signal level was already way above the standard. Placing an attenuator in the line just made all things turn back to the normal situation. "

Again from some posts ago:

"I think I understand his point of view: he says  that the noise inherent to the receiver side remains the same, and so the ratio deteriorates. This is why I had shown that even taking account of this, we are still better with the attenuators. Not talking about the fact that he is cheating, and likes to forget that we were only attenuating back to normal."

Shall I go on and show what Pat had to say about it, ~ five years ago? And then again, recently?

Is it a tactical thing, or hearing problem or a memory problem?

[Joseph K]

And, You had never reacted to this:

Joseph K
So, when we insert an attenuator - and so we turn back to normal - as a side-effect, we also gain a reduction in the level of reflections.
Reflections are causing data-correlated jitter.
So - in general, through application of attenuators, one is 'exchanging' possible data-dependent jitter for a slightly raised level of random noise at the input. Also, while the relative noise is raised proportionally with the attenuation, the reflections are reduced by twice as much.

[wakibaki]

Is it not the case, JosephK, that you only calculated the increase in jitter on the basis of information provided by me, and that you made no effort yourself to obtain such information?

Is it not true that nowhere in this thread was there any mention of the Hiface until I started to intervene?

Is it not true that anyone reading this thread would have got the impression that the suggestions were intended to apply to ANY SPDIF interface.

And is it not true that even in the case of the Hiface the reduction in the slew rate affects the jitter, regardless that this is small, and that you did not know the dimensions of the increase until I provoked you into calculating it?

And was this not remiss of you not to determine the dimensions of this increase?

w

[Joseph K]

My god.. I had sit down to calculate precisely, so as to try & convince You ! about the real weight of it..
And using the "evidence" brought up by yourself.

I have always said that it is there, and small.

[wakibaki]

Anyway, this is trivial.

There are the issues I require you to respond to

... it has no impact whatsoever on the main thrust of my argument with regard to longer cables. Which is that returning reflections @ 44k1 have 177nS in which to land, that they are not themselves normally of a magnitude to be mistaken by the receiver as edges (which would make itself known as something more profound than jitter), that even if the duration of the reflection is 75nS (50 looks a lot fairer to me, given that the reflections in art's scope pictures have only reflected off the DAC), and that they in all probability would be reduced in magnitude if not eliminated entirely by encountering the transmit termination, this still allows for a cable length of at least 10 metres before the reflections start to interfere with the next edge.

And in fact if you repeat the estimates at 96k a cable length of 3 metres can still be accommodated without incurring interference.

The increase in jitter with thermal noise documented in the app. note refers to comparatively normal operating conditions. When the signal is reduced to marginal levels just above those where the DAC fails to lock up, other mechanisms affecting the level of jitter may come into play. No effort has been put into investigating this possibility.

Given that most SPDIF chip manufacturers reccomend a simple resistive termination which is pretty hard to get wrong, while recognising that it is possible to create a termination which measures 75R resistive at DC but not at AC, I still nevertheless recommend that, given the known fallibilities of human judgement in sighted auditions, anyone discovering 75R with a DMM at the input to their DAC, and not being posessed of skills in electronics, or anyone running 44k1 or 48k in any cable length within the specification, or 96k in a cable under 3 metres should find some outlet for their creative energies other than experimenting with attenuators.

w

[Joseph K]

As I see, You are just re-quoting and repeating yourself, even if I just some posts ago had shown differently, and had shown that YOURSELF keep on advertising a circuit with wrong input impedance..

I can not do different, and just repeat what I had shown 9months ago:



Here I'm using my home SPDIF cable (2.3m)- real situation -, Hiface as driver, and the scope is in place of the DAC receiver input. This is exactly what a dac would see, if it has very moderate reflection coefficient. The 470pF case shown here is far worse! The input ferrite case shown by Pat is again far worse! The 94ohm Elliot receiver, pushed by You, is still worse!

Count the nanoseconds, this is a 192kHz signal, the primary reflection is smack in the middle, but there is still life going on at 88nsec as well!
Even in the case of moderate reflections, You still have to take into account the 2nd round trip (4th trip) - the noise what You see at 88nsec. This is the bit time of 96kHz.

From one side, You kept on fighting about 14 -28uV noise level increase at the edges because of the attenuators - here the vertical range is 200mV, even the smallest noise is tens of millivolts level - that is 1000 times higher than the "noise increase" You cried about up to now!!

Here it is the same, with a 10dB attenuator:

[jneutron]

Yes. And depending on the Q of the circuit, the recovery will be exponential, or in this case the linearity indicates that the stray inductance is trivial.

And this is where you are completely incorrect.

The input node has to be forced to zero by the output.  The lag caused by the parasitic inductance of the circuit from output to input means the amp input differential will be too large during high slew and can saturate the amp stage(s).  That is a non-linear response.  And it is slew rate dependent.  This is what generally occurs when active circuit slew capability is exceeded. Edit:   This is exactly the same problem encountered in the digital world with respect to sampling rate, time delay, and closing a control loop.  In the audio world, I believe it's consistent with "TIM" distortion (someone correct me if that's incorrect.)

And the fact is that it makes no difference whatsoever to the arrival time of the reflection.

IFF you ignore the reality of high speed circuit design.  Granted, many designers do indeed neglect this aspect.  Digital links should NOT be alterable by an attenuator, but poor design prevents ideality..

But it has no impact whatsoever on the main thrust of my argument with regard to longer cables. Which is that returning reflections @ 44k1 have 177nS in which to land, that they are not themselves normally of a magnitude to be mistaken by the receiver as edges (which would make itself known as something more profound than jitter), that even if the duration of the reflection is 75nS (50 looks a lot fairer to me, given that the reflections in art's scope pictures have only reflected off the DAC), and that they in all probability would be reduced in magnitude if not eliminated entirely by encountering the transmit termination, this still allows for a cable length of at least 10 metres before the reflections start to interfere with the next edge.

Which is why I detailed two very specific cases.  Totally linear response inductive input, and a termination which is active node.

You are still trying to call everything passive linear.  The real world of audio doesn't listen to you.  Not saying that's good or bad, just is.

Cheers, John

ps..based on your demeanor here, the way you react to others, I would consider you unemployable regardless of your personal situation.  If you could temper your emotional issues (I understand the difficulties given the situation), perhaps you could hang a shingle as a consultant?  You seem to know your way around a scope and t-lines. (I apologize for the test earlier, but honestly, all you had posted to date did not distinguish you from a high school student copying wiki facts...your pickup on the scope risetime tells me you are more than a wannabe.. best of luck with your situation.

pps..it's great to see the return of actual technical talk...