[tanchiro58]

Pat offers an alternate viewpoint, which should be tested by those who have the curious nature to test. 

Yep, that is what I was intending to do and screwed up two of my SB3 but I have learned the excellent lessons of how improving my system to the point of enjoying my musics. But still I will try some more experiments.

 

[art]

I've been ranting on this subject for at least 15 years now. Where have you guys been? I get tired of 'splaining this.

(I just got through with this a week or two ago on some other forum. I posted some graphs there of different cables.)

As for who is right.........I am.

My background is RF and microwaves, not marketing.

The sad fact is.....as one of you has noticed........SPDIF is a kludge, and while we can futz all day long, it will still sound like doo-doo.

All we can hope for is less of a stench.

As for test.......you should be able to tell by listening. A TDR would tell you more. And convince you why some transformers don't have a chance.

Pat

[crooner]

Thanks Pat for your informative post.

Yeah, I guess I hadn't been paying too much attention to SPDIF until the Squeezebox came along. Before this I used good quality CD players and never felt the need to use external DACs.

BTW, I've just finished listening to my SC transformer equipped SB and DAC tonight. They sure sound very nice. Joni Mitchell sounded like she was in the room with me! 

Like I said before, I don't feel the urge to "upgrade" to a better transformer just yet. Perhaps next year. Call me lazy...

What's your take on word clock technology? Do you think this is the wave of the future?
The rudibium oscillators are pretty interesting...

[jhm731]

I've been ranting on this subject for at least 15 years now. Where have you guys been? I get tired of 'splaining this.

(I just got through with this a week or two ago on some other forum. I posted some graphs there of different cables.)

As for who is right.........I am.

My background is RF and microwaves, not marketing.

The sad fact is.....as one of you has noticed........SPDIF is a kludge, and while we can futz all day long, it will still sound like doo-doo.

All we can hope for is less of a stench.

As for test.......you should be able to tell by listening. A TDR would tell you more. And convince you why some transformers don't have a chance.

Pat

Have you tested the Newavas against the SCs?

[art]

I don't need to test the Scientific Conversions parts. A waste of my time.

Pat

[rajacat]

I don't need to test the Scientific Conversions parts. A waste of my time.

Pat

You would be more credible if you were not so arrogant.

Raj

[crooner]

Agreed!

I don't need to test the Scientific Conversions parts. A waste of my time.

Pat

You would be more credible it you were not so arrogant.

Raj

[jhm731]

I don't need to test the Scientific Conversions parts. A waste of my time.

Pat

Pat-

Thanks for your reply.

Now I don't need to waste my time trying the Newava parts you recommended.

[art]

I'm an RF engineer who designs audio gear.

And you are an expert in.............?

SC claims they have the lowest coupling capacitance around. I see no reason to waste my time to see how low it is.

Pat

[tanchiro58]

Hi Guys,

Please give us a little peace here. We are here to discuss about improving our audio components not to fight about which products are the best. Everyone has his own right to say something that you like but please give your knowledge and experience to someone who needs your help (I mean real big help) to improve his or one music system in a good and less spending money ways. 

[art]

I really don't care if anyone takes my suggestions or not. I don't sell transformers.

The point is that over......I dunno.....the last 15 years or so...........a lot of manufacturers have sent me equipment to fix the problems with the SPDIF in/out. Some were companies that had products that needed to be fixed; other were ones that wanted the gear used in their show exhibits to work better.

I don't have to keep a running list of every crappy transformer that I have ripped out to know whether or not I have to compare Brand A to Brand B.

Here is what I do know:

Transformers with low coupling capacitance are not worth diddly.

Transformers with turns ratios other than 1:1 are not worth diddly. Even worse are the ones that have stupid things like 75:75/110 ohm windings. (What do they think a hunk of inductance hanging on the end is going to look like?)

Transformers with shields are not worth diddly. Besides lower capacitance, it adds a big chunk of capacitance to ground. (What did they think that a big hunk of capacitance to ground was going to do?)

The bottom line is:

I have designed pulse transformers. I don't have to look at every POC  transformer around to know if it is good or not. If someone had asked "How does the cheap Pulse Engineering one look like, compared to Newava or Schott?", that would have been a good question. I already know what transformers like anything SC makes is not worth diddly.

Sure, they have a web site, with lots of hype. You can believe it if you want to. I am not here selling transformers, or my services. You can do what you like with my advice.

Except make snide remarks. If you guys want people from industry to donate their time, then you are going to have to show a certain amount of respect. "We" receive enough grief on all of the other forums. No one here needs to add any more.

Now, unless anyone here has a TDR, there is no way to tell which ones are even worth looking at. I suppose that more than half of the crappy transformers that I have ripped out were made by SC. Honestly, I don't know. I suspect that they were. (I know that if Schott had made them, I most likely would not have taken them out. Unless they had a shield. Schott may have made some of those.) The only way to design this stuff is with a TDR, and very few companies have one. Or the knowledge of how to use one. (I know of one company that used to make SPDIF accessories that had one. Usually they borrowed mine. Mind you, they wound their own transformers, because they knew how bad most of the ones out their are.)

There are a lot of other little tricks that you can do to fix up the impedance match. But if I am going to get grief, then I don't see why I should contribute.

Your choice.

Pat

[jhm731]

I suppose that more than half of the crappy transformers that I have ripped out were made by SC. Honestly, I don't know. I suspect that they were.
Pat

Dear Mr. "I'm an RF engineer who designs audio gear" :

How can you make a statement like this, attacking a company's product without any proof?

[crooner]

Rather than respecting a fellow peer's approach, even if he disagrees with it.

He's really offering no solid proof, just vague references...

[JoshK]

Guys, guys....  No good is going to come from this.  Take it as an alternate viewpoint, arguably a credible one, that you can try out or not. I agree that his approach is lacking in tact and comes off as arrogant.  That is blinding you of his assertion. 

You can dismiss him if you think his approach negates his credibility, but the best thing is to try it yourself or move on.

Let's not degenerate into a pissing match.

[art]

I'll give you guys so more proof that you can handle. Of course, this will take time.......my time.......not yours......to assemble it. In the meantime, tune out if you intend to give me any more grief. I don't need it.

OK.....for all of you that think that little things don't count, let's start with something easy like measuring a cable on a TDR.

Here is a Suhner 75 ohm cable, labeled as +/-1.5 ohms. IOW, a decent cable. Here is what it looks like on a TDR, with a 75 ohm load, from my HP cal kit:



The vertical gain is set to 0.02 rho/div. On this cable, you can see (on the right side) where the BNC barrel is that connects to the load resistor. The rho is one division (or so), and that means the rho is 0.02. Or a 34 dB return loss. Or.......the impedance dips to around 73 ohms for a very short period of time.

OK, no big deal, right?

Well, probably.

When you start getting down to -34 dB or so, you have a real hard time hearing changes. (Let's not argue how low it actually has to be.........-34 dB is a good starting point. Trust me.)

But what happens when I use a bog standard 75 ohm load......you know.......the one that you may find laying around some network patch panel........well, here is what happens:



Hmmm.......that's funny.......there is this big doo-hickey jumping up out of nowhere.......and its rho is.................0.07! And the impedance of the load is also up.......up 0.01 rho.

Translation: The connector.....while a decent looking 75 ohm BNC.........jumps up to around 86 ohms for a very short period of time. And the load measures around 76 ohms.

So, if little things like changing a termination show up this much, and are audible (believe me, they are!), what do other things measure like?

Well, hold on and you will see. It will be fairly easy for me to measure some other cables. Digging out my stash of mystery transformers............hooking them up and all.........that will take time. I have a life outside of audio, so you will just have to sit tight until I find time to dredge them all up and measure them.

Pat

[JoshK]

Pat, thanks for taking the time to educate us.  I for one appreciate it.  Can you define rho for me?  I am not privvy to RF.

[art]

A big fancy term for reflection coefficient. IOW.......you send a signal, any signal, down a transmission line, and part of it bounces back. Just like radar.

From that, you can take it, and do this:

20 * log₁₀(rho) = return loss.

"VSWR" to you ham radio types.

Or:

(1 + rho)/(1-rho) * Load Z = Actual Z

Or, if it is less, you just flip the equation upside down.

OK, back to the show.......

Here is a bog standard RG-59 cable on the TDR. Using the good termination:



Well, I guess not bad for something I found on the floor. Around 76.5 ohms. Within the limits for a good cable. The termination leaves something to be desired. Dips to 66 ohms for a short period of time.

It will be important to remember the key word "short" once I get to transformers. You won't see it mentioned there. Even for the ones that I like.

So....what do other cables look like?

Here is one we used to sell. ("Used" to means no longer, and none are available, except the few that are still in the lab. Don't ask to buy any.)

Its claim to fame is that it is 6 m long. I will save that story for another day. Since it is longer, the horizontal scale is different. So, one can say that it is in theory harder to see any glitches.

Yes, but they are still low, using the previous scale. I am showing the longer scale here for demonstation purposes. (This will become clear later.)



Anyway.....hey, it isn't too bad. Impedance seems to be one the money, and the glitch at the termination is about the same as the good Suhner. (Amazing, since I had to modify each connector to fit the cable. Which was made by Belden.)

Now, let's look at a real piece of crap. Oh, wait.......this is one of "our" pieces of crap. Yep, this is the cheap cable we included in our D/A convertor (the one no longer made.......). I am not sure as to the exact length of this example, as it is a cable I use around the lab. The stock one was also 6 m long. I had to change the horizontal scale again, as its velocity of propagation is a bit on the slow side.



And as you can see......there is a distinct slope to the trace. This is what you get when you have a lossy cable. (You will also see how wiggly the line is......not exactly a very tightly controlled impedance. Around 5 ohms too low.)

So, you may conclude, erroneously I might add, that this cable sounds like doo-doo. Nope.........we compared it to lots of cables in the <$200 range. Sounded better than any of them. The key to using a cheap piece of junk, and making it sound good, was to not make the same mistake that most everyone else did. And that was to make it 0.5 m or so.

By making it longer, the reflections that do occur arrived at a point in time after the decision point in the waveform. That, and the fact that it is a bit lossy, actually made for a good sounding POC.

And yes, since we made it, and I designed it, I can call it a POC if I want to.

What do you expect for a pair of $8 BNC connectors (Trompeter "wrench crimp" series) and less than $4 of Belden (9221?) coax?

Maybe that is why we called it the U-Byte 1.

Pat

[jhm731]

Dear Mr. "I'm an RF engineer who designs audio gear":

The information on cables/terminations doesn't answer my question:

How can you make a statement like this:

"I suppose that more than half of the crappy transformers that I have ripped out were made by SC. Honestly, I don't know. I suspect that they were."

attacking a company's product without any proof?

The SC transformers combine high CMRR, low coupling capacitance, low pulse abberation and wide freq response.

Please see the AES 1998, 2001 and 2003 papers on:

www.scientificonversion.com

I believe the Newava parts you recommended are similar to the type "X" units tested in the AES 2003 paper(see page 30), which the SC transformers out performed on all parameters.

Mr.Paul told me he would be happy to test the Newava parts when he returns to the US.

[art]

I've read their papers, and I don't buy a word of it.

Their assertion is that noise cause problems in SPDIF.

Well, yeah, but.........

Reflections cause a bigger problem.

Low coupling capacitance is one of the biggest causes of reflections. And I have tests to prove it.

SC may have a series of tests showing how the jitter is lowered by sticking their parts in some unit. We do not see how they are put in.

There are lots of ways to terminate a transformer that make it look worse than it is. I am not going to give away the stuff that I get paid (by other manufacturers) to do that. Frankly, whether you believe that or not......I don't give a damn. But I am not going to give away proprietary knowledge to a bunch of hobbyists, just to make my point. My "clients" pay for that privilege, and they won't be willing to pay any longer if they know I give it away gratis on some hobbyist forum.

SC does not provide details to satisfy me. I know nothing of the details.......the ones that really count...... to believe what they are selling.

My "clients" don't buy it, either. Which is why I can still make $ in this crappy business.

For those of you who are actually interested in learing, and not trying to argue with me......I found some transformers to measure. The first will be a Newava. Terminated in a 75 ohm load. All transformers will be done that way.

If I can find any SCs, they may be 2:1, which will make a honest test tough.

In any case, I will show a Newava terminated by one of my trick methods at the end. Anyone who isn't blind will see the difference.

Whether you want to believe it or not.............that is your problem.

Pat

[art]

Here is one, terminated by 75 ohms on the secondary. Set-up is Suhner cable-Trompeter (isolated) BNC jack-transformer-BNC jack-HP 75 ohm termination. All parts are "space wired", but the leakage inductance of all of the transformers will swamp that out. (You need not be concerned about the inductance of the wires until you see the "final" version of the Newava.)

OK..........briefly:



The vertical scale is 0.1 rho/div. So, this has a rho of around 0.25, which means that it has an inductive "kick" up to around 125 ohms.....and since the horizontal scale is 2 nSec/div, it lasts for about 2 nSec. (From this, one can calculate leakage inductance.)

Now....one thing I should 'splain, since I am sure that none of you have a TDR sitting around on your bench.........

Inductive reactance is a "kick" or spike upwards. Likewise, a capacitor will show a spike downwards. A resistance will show up a straight line. If it is the Z of the transmission line, it will be at that level. Higher......the line will be higher on the trace. If it is less, well I think you guys can grasp how all this works by now. That is TDR 101, abbreviated.

TDR 102......knowing the amount of the spike, and the time it takes to return to the reference level, there is some formula that allows you to determine its actual value. But since we don't really have to know that actual amount, relative values will do.

Plus, most of us did enough math in school, right? But anyway, you can now see how we can come up with a number for the leakage reactance.

And since leakage reactance is..........inversely proportional.........to the..........coupling.....capacitance.. ..............I think some of you can see where this is heading.

Pat