DIYHiFi.org

OK, let me tell you my experience with USB I2S Dac.

I have build the exact PCM2706 Dac that was build by John Swensson over at AA (see link in "USB I2S solutions... TAS1020B?" here).

My journey so far was:

1.Modified Samsung DVD player - Good sound for little money

2.Highly regarted DDDac with 8 x TDA1543 driven by a Pioneer 801 as transport - much better very analog sounding.

3. Modified Pioneer 801 with tube output stage ? more detail and seperation of instruments than the DDDac

4.USB I2S with PCM2706 to single TDA1543 1uf Auri coupling cap 7.5V alkaline batteries for Dac and 3V alkaline batteries for PCM2706.1.5k Vref; 2.7k Aout to ground. ? WoW ? Same detail as the moded Pioneer but more air and better location of instruments. Very clear and free from any form of digitalis. I have not done any critical adjusting of Vref and output resistors, neither have I played with different coupling caps. John Svensson had better hights with Elso's clock on the 2706.

This poject was very easy to build for a novice like me and nothing that I have heard so far can top this performance.
It is so much more convenient to rip your CD collection with EAC to the harddrive and not to worry about changing disks anymore. No wear either.
I will never go back to any CD transport again.

Next project will be a fanless mini PC with Linux OS and remote control maybe without a screen.

Klaus

No replies,

What`s wrong with you guys?

Are you not ready yet to use your transports and players as a hamster-cage?

I will not look back.

Foobar or ACXO player with DRC (digital room correction) will be on the list as soon as I can master the technical terms used in the manual.

No telling what can be done if we used dacs like PCM63,1704.....

Here is a pic of my wild bord.

Thanks for sharing, Klaus.

Since it obviously works and looks as simple as one can imagine, I may turn back to life my USB project sooner than I planned...

As I understand, the main idea behind USB DACs was to get rid of S/PDIF?s jitter. But... I have tried around to dig out some real info, i.e. actual jitter measurements of USB DACs with no much of success. I was sure that just a few days before your post I came across some measurements that show relatively high jitter level of USB DAC but now I can not even remember where it was... I am obviously getting senile. The only other graph I can find at the moment is Stereophile?s measurements of 12dAX7 and yes, jitter there is high. Do you have any means to check the situation with the unit you?ve made? Anyone else can point to other references?

And yes, PC fans noise is the issue.

Pedja

Hello Pedja,

Actually I have no chance to mesure anything since I do not own a scope.
I don't even have any clue about electronics. The only thing I can do is copy something that has been done already.
So in reality I have no right to be in this forum but I just wanted to share my finding with you guys, so you can evaluate this route by yourself and maybe refine it.
In this case and some others I greatly benefitted from not being able to apply some clue logic that would ceep me from trying something that someone raves about.
The only thing that keeps me going the DIY (copy) route is that I am addicted to good sound.
I have build so far an active XO, AKSA 55N, EL84 tube amp, DDDac, Tube pre, modified two players and build this PCM USB dac. I have greatly benefitted from guys like You, Thorsten Loesch, Carlos FM, Elso Kwak, Hugh Dean ..........
It is a little embarrassing that I still can not maintain an intelligent conversation about electronics with you guys, but I am in it just for the results.
All I can say is that my system at the moment sounds better than the one I had years ago (Accuphase E303, expensive Thorens table, speakers with $1300 for the parts alone.
The analog like sound, ease of presentation, seperation of instruments, cleaness of sound this little Dac puts out has to be heard to be believed.

Cheers,
Klaus

Hi Klaus,

Excellent post. I think there are a lot of us out there. I've built lots of amps will very few problems and I understand electronics enough to get by, but I couldn't hold a technical discussion with a ten year old because I'm not knowledgeable enough to use the terminology correctly without embarassing myself.

I also would like to thank all the gurus and very one that's puts the effort in to help.

Keep building, (although I don't think you need to worry about a SS amp as the AKSA is the best I've found).

_________________
Greg Erskine

Sure you do. If everyone here spent 30+ years doing this stuff, it might get boring. A lot of us are more than willing to help. We don't always have enough time to spend lending our expertise, so hang in there.

How do you think that I got started???? Building stuff that I copied from ham radio magazines. Eventually, I progressed.

Some will say that I have now de-evolved! [joke]

Jocko

Hi Pedja.

Well, I also can't see what the fuss is all about on the USB interface for digital audio.
You can have bit-perfect copies on the hard disk, but how to get it out of there without (not possible, but at least minimal) jitter?
Besides miserable PSU and huge amounts of noise inside a PC, the bits have to be converted, and I can't see how that conversion can be better than a direct I2S connection of even a well implemented SPDIF interface.
This is Jocko's field.

Take a look here:
http://www.stereophile.com/accessoryrev ... ndex3.html

Quote:
"Finally, examining the Total BitHead's rejection of word-clock jitter when fed USB audio data was somewhat problematic, as the measured amount of jitter was surprisingly dependent to some extent on the setting of the volume control. With the Miller Jitter Analyzer, the worst case was with the volume control full up, with a measured jitter level of 1.12 nanoseconds peak-peak."

What's "acceptable" for the common PC has yet to prove to be hi-fi.
In a world of MP3, everything is acceptable.

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

Thanks for the link, Carlos.

I actually just don?t have a problem to believe Klaus the thing works nicely. The issue I am about here is related to the ?USB solves jitter problems? claims. The things are apparently not at all that simple. Like they never are in reality.

I have dug another Stereophile?s measurement that may be probably more symptomatic, though the unit again may not be any representative. Figure 8 shows it has significantly higher jitter if USB input is used (it is already high with S/PDIF).
http://www.stereophile.com/amplificatio ... ndex8.html

That is the one side of the things. The other would be that I just don?t believe the relation between jitter and subjective performance of digital sources (I/V, supply, layout etc aside) is firm and simple.

Pedja

Thanks guys for the encouraging words,

Carlos,
You got me confused.
You are argumenting from so called ?known limitations? of the implementation.
You had long arguments with people over at DiyAudio that took this aproach with your suggenstions: snubber, OpAmp bypassing, Buffer for chipamp......
What I heard you saying was: try it or stay with the second best
Now you take exactly the same approach.
Just order some PCM2706/7 samples and try it yourself.
I have the DDDac and some extra TDA1543 (Euro 0.75 here in Germany) that I played with and I can tell you that the USB I2S approach at least with the PCM2706 gives a major improvement over the Spdif route.
Jidder is a very complex subject.
Maybe there is more about jitter that we don't know, than that what we know?
Jidder could consist of more components than what we know.
Like with speakers: there is k2 and k3 distortions that makes up the total distortion.
We are less susceptible to k3 than to k2.
Actually I don't care if the jidder increased as long as the sound did.
Sometimes I think that messuring equipment keeps you more from making progress than from making mistakes.

Klaus

Klaus, sorry if I was not very clear in my post.
I was not talking about your dac, it can sound very good as you say.
My remarks were related to Stereophile's artice.
I was generically speaking about the USB interface, in response to Pedja's interesting post.
My point was that I have my doubts that USB is a better interface than any other.
This does not mean that a USB dac cannot sound good.
Of course it can, as a dac connected by SPDIF can.

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

The degradation of the sound lies obviously in the spdif transformation process.
Let me explain why.
My TDA1543 Dac is as simple as it gets and it stays the same for either implemetation: one resistor and one cap at the output for each chanel; 7,5V alcaline battery supply.
In both cases the Dac gets an I2S signal.
If the I2S signal comes from the CS4312 with the Pioneer transport it is clearly inferior to the setup with the USB I2S.
To say it again, it is not a subtile difference
The Pioneer stable platter transport is an exellent transport and the implementation of the CS4312 on the DDDac is also very good, as John Svensson says his DDDac outperforms his Scott Nixon Dac.
The conlusion of my test clearly tells me that I want to avoid the spdiff conversion.
If I recall it right Carlos, you seem to have a Dac or a player with the TDA1541. You could easily feed it with I2S from this setup and see for yourself.

Klaus

I have 3 external dacs with direct I2S (two TDA1543 and one TDA1541) and one external dac with AD1862, with SPDIF.
The AD1862 (SPDIF) sounds now on par with the TDA1541A (I2S), since I made some changes on the filter component values on pin 20.
I used 390R+330nf.
You try it.

_________________
Carlos Filipe

"Any intelligent fool can make things bigger, more complex, and more violent. It takes a touch of genius - and a lot of courage - to move in the opposite direction." Albert Einstein

Hi Klaus,

Thanks for starting this thread. I am excited to try it, waiting for parts now.

I have been using a simple NONF TDA1543 for a couple years. I use S/PDIF from a Chaintech AV710 soundcard. This sound card can output "bit perfect" 44.1 khz output. (Flash to Prodigy 7.1 bios)

Have you determined if you can get "bit perfect" output using the PCM2706? A test for "bit perfect" output is to send/play a DTS file to a DTS receiver and see if it lights up. I keep a old DTS receiver around for this test.

Also how about by passing Window's kmixer? Can you use ASIO2 with it and so on. Pros and Cons so far?

I don't mean to imply it has to be "bit perfect" to sound good.

I am considering using LVDS Transmitter/Receivers to by pass S/PDIF and to also allow me to use CAT5 cable to send I2S to multiple DACS, digital xo, DRC and so on as mentioned here (not by me):

http://www.diyaudio.com/forums/showthre ... post658109

The PCM2706 may get me there sooner then the road I am currently taking. Mind you the sound card will do 24 bit 192khz. Currently all my music is Red Book so no rush to 24 bits yet.

Cheers

If you are going to use LVDS to send I2S, why even mess with SPDIF?

Jocko

Hi Jocko,

Sure if LVDS works out for me I will give up SPDIF. I like the idea of using networking cable/technology to shoot I2S around. Have you tried it? Pros and Cons?

If LVDS works out then next I would take a Prodigy 7.1 card and attach/hack LVDS to its 4 I2S outputs and use VST plugins to play with digital xo and the like. I am building 3 way speakers and this would be a nice way bring it all together or at least help me find the xo points.

But first I will try this USB method since it looks so simple. I am sure it will be useful for some situations. Not sure if it can be used for digital xo setup yet.

Once I get it all figured out I will share the details for others to build, speakers, DAC, amps and all. A low cost "PC as source" system for beginners (like me) and those who have outgrown their iPods and MP3 players.

Cheers