[jneutron]

The only reason that you see that hump in JosephK's picture is because it's done on a TDR. TDRs have a very fast rise time, that's how they achieve their resolution. The TDR's rise time exceeds even that of the video amplifier, which has a respectable slew rate. Consequently the voltage at the input node rises faster than the amplifier can compensate by feeding back through R2.

If, however, the video amplifier were hit by the ~6nS typical rise time of an HCMOS chip, this hump would be barely visible, if at all, because the amplifier would be able to maintain the voltage at the inverting input close to 0V. N'est ce pas?

Trivially, the stray inductance in the feedback loop is insignificant

w

Oui, c'est vrai.

Do us all a favor.  Go back to Pats VERY first post.  Read where he stated the rise time of his spdif is less than 1 nanosecond.  and, that this fact was very important.

You need to slow down and read everything again..

And, as an added bonus, you'd probably see that Joseph's setup actually caused a 1 nSec total risetime.  You can see this by the text reference to  "5 nSec per box" coupled to the fact that the rise is one set of hash marks wide..

And....as an added added bonus, look at the first shot with the higher amplitude.  Note that it takes almost 10 nanoseconds for the input node to settle.  But, with some attenuation, that is dropped to 6 nanoseconds..

Now...think...what happens to the output as the node is being driven back down.  Answer...it ain't sittin still.

You are working so hard and fast at trying to prove others wrong, that you are missing all the details..

Cheers, John

[jneutron]

Trivially, the stray inductance in the feedback loop is insignificant

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

It's fun stuff.

Cheers, John

[wakibaki]

Please keep the topic technical

[jneutron]

Then answer the questions after you have read the posts.

You are confusing t-line input parametrics with active circuitry response.  You called an inverter with a parasitic inductance a differential inverter with a dirac output..

You are making many many sweeping technical generalizations without actually looking at the content you are trying to trash.

Discuss technical.  If you don't understand what I'm talking about, please ask questions.  There is no shame in having no experience with sub nanosecond power circuits..not many have.

But mis-application of "very accurate" t-line understandings is wasting bandwidth.  As I said, you seem to understand t-lines, but you need more.  You can't just toss that square peg into every round hole you find..


Cheers, John

[jneutron]

Joseph K.

It would be wonderful if at some point, you could get some snapshots of the vid output as well as the vid input for various drives and input slews.  It would be very educational for all if we could see how the output settling time is affected by the "failure to zero" the input node.

It may require you wrap a bare wire around the probe tip ground and hit the ground plane at the point of measure so that you keep the tip loop down.  I would venture a guess that you don't have one of those tip adapter widgee's.  Oh, and it'll be fun listening to you gripe about how difficult it was for you to find a suitable ground point that doesn't bounce the daylights out of the waveform..Hey, why should I have all the fun?

Again, pat and josephk...thank you for the wonderful scope photos, keep up the good work.

Oh, and ps...It'd also be great if somebody would explain what an SPDIF receiver uses to lock it's clock to.  What, with all these here reflections bouncin back and forth, I plum forgot where in the signal I'm supposed to be referencing my eeeenternal clock to so's I can reconstruct my bit stream....

Cheers, John

[wakibaki]

You need to slow down and read everything again..

And, as an added bonus, you'd probably see that Joseph's setup actually caused a 1 nSec total risetime.

The TDR sends a ~ 90psec rise time, 400mV magnitude step signal

You need to slow down and read everything again..

w

[wakibaki]

the ... dialog is providing many of us with useful learning material.

Rather than mislead readers it is only fair to point out that I have made a mistake here...

There is no question of this being an 'active' reflection, it's shape is entirely characteristic of a passive inductive termination, if it were active one would expect to see a change in level on the RHS when the output has settled.

This is NOT in fact the case. There is nothing to distinguish this reflection from an active one. My apologies for any confusion caused.

w

[jneutron]

The TDR specification may as well be 90 femptoseconds. (yes, it's possible...it's done here at work).  That is the rating of the TDR, not what it will actually do once it's connected to a middle to low quality coax.

If you really want 90 pico risetimes, you have to contend with the coax dielectric, the surface finishes of both the inner conductor and the outer conductor, and outer braids certainly cannot support 90 pico risetimes.  In fact, we use copper tubing, sma's, and because it's a high radiation environment, we gotta go with SiO2 dielectric.

So, no...the DAC is certainly not seeing 90 pico risetimes..  Have you ever actually used this equipment, or are you just using book or google information?

If you are going to argue theory with someone who does it for a living at levels 3 orders of magnitude beyond any experience you could possibly have, you need to get your ducks in a row.   

Cheers, John

note...I changed the 40's to 90 for consistency.  The actual numbers are moot of course.  We have an experiment here where they have a race condition between a class 4 laser beam and a particle beam...they use a mirror on a moving stage to alter the laser travel distance in increments of 100 millionths of an inch..how long does it take for light to travel .0001 inch?..

[jneutron]

Rather than mislead readers it is only fair to point out that I have made a mistake here...

I'm not sure exactly what you are saying here.  However, the text you have quoted was not aimed at you.  It is my belief that this dialogue with everybody, yourself included, is educational and well worth everybody's time.  The fact that you, me, or anybody may make mistakes in the dialogue is actually inconsequential..  It is what it is.  My concern is that everybody be civil and not lose the content for the attitude.  I personally am very happy that you continue to engage.

My apologies for any confusion caused.
w

A worthy comment which is appreciated..thank you.

Cheers, John

[wakibaki]

how long does it take for light to travel .0001 inch?..

Can anyone explain to me how this is germane.

Twice now I've pointed out to you the errors you have made in your own words. Twice you have attempted to to ignore it and engaged in attempts to browbeat me and wow the audience. You have clearly not read the material, while insisting that I should read it again.

My engagement with you is at an end.

In the meantime, here are a few extracts from SPDIF receiver datasheets from Cirrus, AKM, Wolfson, and Texas (SRC4392).

They show the manufacturers recommended SPDIF termination.









Not a video amplifier in sight.

w

[jneutron]

Can anyone explain to me how this is germane.

Easy.  You have claimed that the vid amp was being driven at 90 ps rise times.  I pointed out the reality that no matter what the TDR specification is, the rise time at the receiver will NOT be 90 pSec.  If you actually looked at the scope pic, it's 1 nSec.

AND, that 1 nSec is the EXACT same signal slew that Pat CLEARLY stated.

You have looked only at the spec numbers and assumed...   An experienced individual who deals with this knows better.  That is why I asked you if your assertions and understandings were coming from a book, or google..  edit:  you need to understand that there are some who work with 50 femptosecond resolution systems...and at those levels, we have to do silly things like keep the room temp, the wires, and even the fiber optic cables within a tenth of a degree C stable..  For some, these speeds are normalcy and nanosecond speeds require a calender..

Twice now I've pointed out to you the errors you have made in your own words.

Actually, you've done no such thing.  Stating such doesn't make it so.

Twice you have attempted to to ignore it and engaged in attempts to browbeat me and wow the audience.

What I'm trying to do is get you to learn the topic.  Part of your learning would be to read more slowly and carefully.  Another part is for you to ask questions.  So far, you've done no such thing..

My engagement with you is at an end.

Quite honestly, I'd prefer you continue.  But as part of your continued dialogue, I'd hope you slow down and learn something.

Telling you to slow down, read accurately, ask questions is not browbeating dude.  Saying so doesn't make it so.  Diversionary statements are unbecoming, please stop.

In the meantime, here are a few extracts from SPDIF receiver datasheets from Cirrus, AKM, Wolfson, and Texas (SRC4392).

Thank you for taking the time to post this. 

BTW, what is the note associated with the Rx line of the AK4113?  Is that an attempt at limiting slew rate, and if so, is it application specific?

Cheers, John

[wakibaki]

AND, that 1 nSec is the EXACT same signal slew that Pat CLEARLY stated.

This is with a very fast source. It has a rise time, around 0.8 pSec. The problem is...........

Now, we know he means 0.8nS, but will you PLEASE read the material.

w

[wakibaki]

If you are going to argue theory with someone who does it for a living at levels 3 orders of magnitude beyond any experience you could possibly have, you need to get your ducks in a row.

What's this. if it's not browbeating and trying to wow the audience?

w

[wakibaki]

If you actually looked at the scope pic, it's 1 nSec.

Scopes have rise times too.

w

[sts9fan]

"read what I wrote"
"No, you read what I wrote"
"No, you read what I wrote"
"No, you read what I wrote"
"No, you read what I wrote"
"No, you read what I wrote"
"No, you read what I wrote"

I think you guys should call this a draw.

[jneutron]

Scopes have rise times too.

w

Excellent.  It took you a while.  I was praying that everybody else would wait for you.  As it were, you should have pointed out my intentional error immediately.  You pass, but not with an "A".

What wakibaki has correctly pointed out, is that in general, toy scopes have a limited rise time.  Typically, they run about 700 pSec.  Also, they do tend to load the tested point to the tune of about 10 pf (for the probe tips).

For my work back in '81, measuring 250 picosecond risetimes accurately on a 1Khz rep rate square wave required using the scope in a dark room and waiting for my eyes to adjust to the darkness.  In '83, sampling scopes were getting good enough.

However, waki...the simplistic phrase "scopes have rise times too" doesn't eliminate the fact that a 90 pico rise will NOT survive the transit through a cheap cable nor an RCA nor even a wirebonded leaded device.

There is much you can learn here.  Just stop attacking everybody.

Cheers, John

[jneutron]

What's this. if it's not browbeating and trying to wow the audience?
w

Umm..it's a statement of fact. And, quite a few of the "audience" who would read this dry stuff, already know what I do.

Actually, here is the most important thing you must learn.

Your initial premise of "they can't be right because I know......" is certainly not a bad one.  Many times we see some crazy assertions that cannot be correct.  Many times, that is exactly my take.

There will be times when a crazy assertion actually is correct, even if the author of that assertion makes up some half crazed physics hooey..

In this particular case, the assertion of audibility may not be a bad one.  Discussion of the root cause as related to the proposed solution is always worthwhile.

In this particular case, you simply confused a discussion of a complete system with that of only the t-line front end.  You were accurate enough with that aspect, but you were way off the mark once active nodes were involved..

In summary:  You and I are closer in thinking than you care to believe.  The only differences between us?  My experience (which is not so important), and my demeanor (which is the whole ball o wax here).

You need to lighten up, slow down, and approach the discussions with more civility.

Cheers, John

[jneutron]

"read what I wrote"
"No, you read what I wrote"...

To this, I say to you...

"They only perform twice on a Saturday"

Now..who said that?

Cheers, John

[jneutron]

hmmm..I was hoping to have it pop up as a pic, not a link..oh well.

wakibaki.

If you really wanna flex your thinking about t-lines, try an analysis of a speaker cable driving a speaker load.  Here's a graph.  If you're interested, we could start a new thread.
http://www.audiocircle.com/image.php?id=38430

edit:  as it is, I've been considering a collaboration with some guys on actual hardware and testing to validate this..
Cheers, John

[wakibaki]

my intentional error

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

Sorry, that should read 'despite your attempt to divert the argument into an irrelevance'.