[sandab]

 
It also shows the operating transmission rate for the Pulse at 1-7 Mbps -way over kill for 192K (funny Murata doesn't report theirs).
 

It is?  I would think it's way low, unless they mean data transmission using a 2x clock (dual phase).  AES3/SPDIF runs at 3Mbps for baseline 48k/24 with two samples per frame.  But it uses a biphase clock (Manchester coding) so it actually clocks the wire at 6MHz.  96k is twice this, 192k four times, and 384/24 8x; so you're looking at a 48MHz clock for 384/24.  This is why there's a 48MHz crystal on the board.
 
Edit: 6MHz plus change due to framing overhead (preambles and such).

 

[rb2013]

It is?  I would think it's way low, unless they mean data transmission using a 2x clock (dual phase).  AES3/SPDIF runs at 3Mbps for baseline 48k/24 with two samples per frame.  But it uses a biphase clock (Manchester coding) so it actually clocks the wire at 6MHz.  96k is twice this, 192k four times, and 384/24 8x; so you're looking at a 48MHz clock for 384/24.  This is why there's a 48MHz crystal on the board.

Edit: 6MHz plus change due to framing overhead (preambles and such).

My Musiland 3.0 USB ran perfectly at 384/24 using the Pulse transformer. So it must be passing the 192/24 bit stream easily.

What's the operating transmission rate on the Murata?

 

[rb2013]

Come to think of it. The USB 2.0 throughput rate is 480Mbps, even with bus limitations its 240Mbps. Even lowly USB 1.x Full Bandwidth is 12Mbps. We know that to pass a 192/24 audio data stream without issue we needed USB 2.0. I used to use 1394b before the advent of USB 2.0, especially Async. So we must be talking apples and oranges. No way 384/24 would make it through.

It makes no sense to design a sota USB audio inteface, designed for 384/24, and have it bottleneck on a $3 transformer.


http://en.m.wikipedia.org/wiki/USB#USB_2.0

 

[Alexsander]

"funny" Murata is installed in very expensive devices:
Burmester 089 cd player: €14500
http://www.i-fidelity.net/en/Printview.html?showUid=1015
and M2Tech Vaughan DAC Price: $7,999.00 
http://wizard-highend.blogspot.com/2012/12/m2tech-vaughan-inside.html
http://soundnews.ro/wp-content/uploads/M2Tech-Vaughan-1-of-1-11.jpg
http://soundnews.ro/2013/03/18/m2tech-vaughan-review/
 
....here question not of the price at 1.57..... qualities and its application!

 

[rb2013]

"funny" Murata is installed in very expensive devices:
Burmester 089 cd player: [COLOR=545454]€14500[/COLOR]
http://www.i-fidelity.net/en/Printview.html?showUid=1015
and M2Tech Vaughan DAC Price: $7,999.00 
http://wizard-highend.blogspot.com/2012/12/m2tech-vaughan-inside.html
http://soundnews.ro/wp-content/uploads/M2Tech-Vaughan-1-of-1-11.jpg
http://soundnews.ro/2013/03/18/m2tech-vaughan-review/

....here question not of the price at 1.57..... qualities and its application!

The Vaughan is a DAC. I had the TOTL M2tech PC interface, the Evo. It sells for $500, I had it powered by a Acopian Yellow box linear ps that was $300. The U12 is way better. And isn't the Evo limited to 192k.

The Burmister stuff is insanely overpriced audio jewerly, and it's a cd layer/dac. Not a dedicated pc interface.

Still no one can answer my question: what the 'operating transmission rate' of the Murata. We know it's worse then the Pulse on common mode rejection.

You can't compare the prices of dacs and cd players to a pc interface. Again - Apples and Oranges. Proves nothing.

Edit Again from an audio engineering pov, why use a part that's twice as expensive if it's inferior? Makes no sense.

 

[rb2013]

I've always been a big fan of Auralic's stuff. They're truly cutting edge, sota, designers. Not cheap, but really great dacs for the money.

They use the Pulse in their latest sota $3,300 dac - the Auralic Vega.

http://www.digitalaudioreview.net/2013/04/auralic-vega-review-part-2-pcm/


In the $1999 Ark MX+ they also use the Pulse. 'Nuff said

http://www.6moons.com/audioreviews/auralic/1.html

 

[sandab]

My Musiland 3.0 USB ran perfectly at 384/24 using the Pulse transformer. So it must be passing the 192/24 bit stream easily.

Their suggested data rate has little to do with it - the PE-65612 has a 55MHz bandwidth @ 3dB (50% amplitude loss).  This tells you how much it rounds the corners of a square-wave pulse train.  It also tells you at which point - 55MHz - a square wave has become a sinusoidal.   A 48MHz pulse train will no longer be square and will look closer to a sinusoidal.  For AES3 this is pretty poor, as the leading-edge timing of each pulse matters (it's used to recover the source clock) and a sinusoidal will produce a crap-ton of jitter.  A little bit of rail noise or offset voltage will cause the trigger voltage to change and the trigger point on the rising side of the sinusoidal to shift temporally.  This in turn will make the DAC sensitive to any noise whatsoever - no matter where it comes from as its notion of the source clock based on its receiver PLL will wander, causing the sample clock to wander.

 

[sandab]

And looking at their lego-block "Application Circuit", it's utterly laughable.  It essentially connects the two devices' grounds together separated only by the cable shield!!!  That's totally, positively, patently, bat-crap insane.  There is absolutely no use case in the life of the universe one would ever want to do that, especially since their diagram uses differential signaling, whose sole purpose of existence is isolation.
 
http://media.digikey.com/pdf/Data%20Sheets/Pulse%20PDFs/PE-65x12_Rev2006.pdf

 

[rb2013]

Their suggested data rate has little to do with it - the PE-65612 has a 55MHz bandwidth @ 3dB (50% amplitude loss).  This tells you how much it rounds the corners of a square-wave pulse train.  It also tells you at which point - 55MHz - a square wave has become a sinusoidal.   A 48MHz pulse train will no longer be square and will look closer to a sinusoidal.  For AES3 this is pretty poor, as the leading-edge timing of each pulse matters (it's used to recover the source clock) and a sinusoidal will produce a crap-ton of jitter.  A little bit of rail noise or offset voltage will cause the trigger voltage to change and the trigger point on the rising side of the sinusoidal to shift temporally.  This in turn will make the DAC sensitive to any noise whatsoever - no matter where it comes from as its notion of the source clock based on its receiver PLL will wander, causing the sample clock to wander.

I guess you're a smarter audio engineer then the fellows at AURALiC. You should have your own audio company. If it's good enough for them, it's good enough for me.

Been eyeing that Vega, nice unit. But I'm sure you have many suggestions on how they went wrong.

 

[rb2013]

And looking at their lego-block "Application Circuit", it's utterly laughable.  It essentially connects the two devices' grounds together separated only by the cable shield!!!  That's totally, positively, patently, bat-crap insane.  There is absolutely no use case in the life of the universe one would ever want to do that, especially since their diagram uses differential signaling, whose sole purpose of existence is isolation.

http://media.digikey.com/pdf/Data%20Sheets/Pulse%20PDFs/PE-65x12_Rev2006.pdf

Laughable, bats**t insane, absolutely no use, - yeah ok. When's your sota stuff hitting the market.

 

[sandab]

Laughable, bats**t insane, absolutely no use, - yeah ok. When's your sota stuff hitting the market.

That's a comment on the data sheet, not the product.  You need to learn to read and pay attention to what's said; not what you think was said.  Not interested in your fanboi pissing contest.  After 30 years as an electrical engineer, working on everything from broadcast video codecs to consumer electronics I've seen technical writers come up with some really ridiculous documentation.  This is far from the worst, but it's pretty bad.  The Vega doesn't connect the AES3 shield to the isolator ground, guaranteed.  Because it would sound like crap if they did, they'd wonder what's wrong, and then they'd tell the junior freshly minted junior engineer responsible to not do that.

 

[sandab]

So I did check the 5V spectrum under 110ohm load on various cables with and without the Wyrd.  Nominally this is 5V 45mA.  I only had 1/4W resistors on hand, and didn't want to make smoke...
Computer is a late 2013 MacBook Pro 15" retina.
There's no data traffic on the cable (I don't have the breakouts needed for this).
The ports are identical, and it makes no difference which is used.
 
There's no significant difference between the cheapo 3' clear shielded cable and the 6' Ghent Audio U01, with the Ghent being slightly better shielded.
The 10' Monoprice cable is an effective radio antenna and picks up tons HF spikes up to 30MHz.
All the cables pick up noise, but the cheapo clear and Ghent U01 show much less than the Monoprice.
The RF noise shows as singular spikes across a wide band, approximately 1-30MHz.
 
The main difference is under 300kHz, and for this the cable really didn't make much difference, maybe a few dB at most.
The Wyrd reduced it greatly, by roughly 25dB around 150kHz.
 
It made no difference what I used the computer for, even playing music on the other port, or copying files on the internal SSD, or both.  The output
spectrum looked the same.
 
The noise is additive, on top of the 5V reference.
 
Clearly up to 300kHz is computer sourced while above this there's RF interference.
The Wyrd did nothing for the RF.
RF is a hugely affected by the cable used.
 
MacBook Pro... This is an overlay, allowing the spectrum to build up.  I placed a cursor at 300kHz 20dB for reference.  Each vertical divider is 10dB.  Ghent U01 cable.
 

 
Same but with the 10' Monoprice cable.  This shows the cable affects computer-sourced noise as well.
 

 
 
MacBook pro via cheapo clear cable to Wyrd, to Ghent U01.
 

 
Here's the Wyrd on its own, disconnected from the MBP.  Not much difference, which shows good isolation.
 

 
The absolute value at 20dB is about 20mV above the 5V reference.
 
There's a peak around 150kHz without the Wyrd, showing 25dB spikes (= over 10V) when fed straight from the MBP.  There are more of these with the MP cable (curiously) than the Ghent... 
 
Clearly the Ghent is a superior cable to the Monoprice, and clearly the Wyrd cleans up the USB 5V, and not by a small amount.

 

[rb2013]

That's a comment on the data sheet, not the product.  You need to learn to read and pay attention to what's said; not what you think was said.  Not interested in your fanboi pissing contest.  After 30 years as an electrical engineer, working on everything from broadcast video codecs to consumer electronics I've seen technical writers come up with some really ridiculous documentation.  This is far from the worst, but it's pretty bad.  The Vega doesn't connect the AES3 shield to the isolator ground, guaranteed.  Because it would sound like crap if they did, they'd wonder what's wrong, and then they'd tell the junior freshly minted junior engineer responsible to not do that.

I'm sure the engineers at AURALiC will be holding their breathe awaiting your further critisms. I guess 6moons and Darko are just like little children in your immense shadow. What else do I need to learn from your greatness? Besides how this Pulse transformer works perfectly despite your deep technical gibberish about how that's impossible.

 

[sandab]

I'm sure the engineers at AURALiC will be holding their breathe awaiting your further critisms. 

What is this BS?  I haven't criticized their products in the slightest.  You're the only one being criticized here, but you're clearly too stupid to recognize this.  Blocked.

 

[rb2013]

What is this BS?  I haven't criticized their products in the slightest.  You're the only one being criticized here, but you're clearly too stupid to recognize this.  Blocked.

Now some advice for you - try reading your own posts. It's hard to have an intelligent debate with someone who doesn't understand what he's posting. You might start a few posts of yours back...some incoherent rambling about their Vega and a junior engineer. Criticizing me? I didn't write a data sheet for the Pulse you called bats**t insane, I think that's what that rant was about - hard to say, I'm trying to keep track of your many tangents. Maybe time to return to the topic of this thread.