The new Nonoz II DAC page !!!

[fedde]

Hi there !

Sorry for my absence I was celebrating my fathers birthday this weekend...
Luckily most question could be answered by other members. Thanks, also for the interesting links. Some reactions:
- I directly used the S/P-DIF input circuit at the mentioned page. I will include it later on my page. I am quite busy these days...
I guessed that most people will want to build it without transformer. And I did not forget it !
- I use the transformer in the DAC because I also use my PC as source. I do not plan to modify my sound card to add one
- I was already planning to calculate the roll-of of my filter. A little bit on the late side, but...
The -3dB point is where the capacitor reaches 1.5 k impedance, this is at 32 kHz. Maybe a tad on the low side. I will try 2,2 nF or so later for a -3dB point around 50 kHz. I think I have some 2,2 nF styroflex around here. And actually I am quite deaf ;-) I can't hear 20 kHz. But it could be my DAC ... ;-)
- You could add a transformer to Scott Nixons DAC in the same way as in my DAC.
- I never said/claimed that the TDA1543 is better than the TDA1541(A). I think the TDA1541A S1 should be better. But it's very easy and cheap to make a DAC with the TDA1543 and it sounds quite nice!!! I think that the TDA1541 has more potential for a top-end DAC. But then you need tubes in the output and a lot more components (triple supply!). The cost will be 5-10x higher than a very good build TDA1543 DAC. So it's more about quality/price. I never heard a good TDA1541 DAC. I use the silver-crown TDA1541A in my Philips CD-960 player. I am still busy with that player. The supply and output stage is not very good yet. So at this moment I will not compare them..

Ciao and good luck building your DAC Peter! Just try it and tell us what you think.

Fedde

[jean-paul]

Thorsten (Kuei Yang Wang) and Lesha will of course prefer the TDA1541AS1. If you ask Fedde or Jean-Paul, they seem to prefer the TDA1543.

I like TDA1541AS1 a lot too, just as I like PCM63K in os DAC's. But TDA1543 is convenient to use and IMO it gives the highest rate in soundquality vs. cost / hassle / complexity. Sound in non os mode really is a feelgood experience and to me it was quite an eyeopener after other experiments with DAC's. Whether it be os or non os topology does not matter that much to me as I switch regularly between os and non os DAC's. Good sound is good sound

Newest star at my TDA firmament is TDA1545A of which I am very curious what the results will be. But I also have some PCM1728 and PCM1738 and ... too little time

[Peter Daniel]

Should I bother with 4 in parallel, as Kusunoki DAC, or one is definitely better? Also, active or passive output stage?

[jean-paul]

Hi Peter, I would use one like Kusonoki did in his 2 last DAC designs. Why do you want an active output stage while you can do it passive ? When fed with 6V nearly 0.8 V comes out. Not too much but nothing to complain about since only one resistor accomplishes that.

Please consider the drawbacks when the DAC is used with lower input impedance amps. Although it looks fine on paper the DAC really likes high impedance loads. Don't ask me why but I noticed this several times. Output cap was a 2.2 uF in all cases so that couldn't be the problem. When lower input impedance amps are used I would advise to add at least a buffer and preferably an amplifying output stage to achive 2 V output voltage. Otherwise don't bother.

[RAYSIMMONS]

Peter,

Which non-oversampling dac design are planning to use?



Ray.

[Peter Daniel]

I was always thinking about building that one http://www.homestead.com/whaan/files/tda1543_dac.jpg

[jean-paul]

That's the first version of the TDA1543 DAC by Kusonoki with which he won a contest of MJ magazine in 1996. Two other designs came later on with some refinements but also with some cuts in cost/parts.

You can check which version the man himself likes most:

http://www.tnt-audio.com/intervis/kusunoki_e.html

[rbroer]

Dear all,

I recently had a listening session with friends at his home with 3 cd players (and a decent record player):

1: cd player with 3 parallel TDA1543's, non oversampling, LCaudio Lclock-c (very old version) clock, passive I/V
2: cd player with single TDA1543, non oversampling, Kwak-CLock6, passive I/V, synchronous reclocking
3: cd player with TDA1541A-S1, non oversampling, Kwak-Clock6, common base (active) output stage, synchronous reclocking, DEM reclocking

To summarize:
2 sounded like 1, but with more detail and slam/weight
1&2 compared to 3 sounded restricted dynamics, a bit choked at times.
3 sounded by far best to our ears, much clearer, more detail, better dynamics, more depth, very natural, clarity.

With some albums the sound of the record player was very good (and preferred by the three of us).

I've been suspecting the "passive" output stage of the TDA1543's for a while.
I've started adding a common base output stage to one of them, but it's on hold for a while since I'm moving.
Will post when I've got some results.
Also coming up; an improved common base with folded cascode, DC coupling for my TDA1541A-S1 player.

IMHO I always have heard a substantial difference between TDA1543 and TDA1541, even in standard cd players with opamps (yuk).
A real improvement was synchronous reclocking for the TDA1543, it got a warmer sound that 'd been lacking before.
A big improvement for the TDA1541A-S1 is DEM reclocking IMHO, more than adding a low jitter clock.

Just remember; there are many permutations possible considering supplies, clocks, output stages etc. synergy with loudspeakers, amplifiers, cables etc.
For example, ELSO prefers the TDA1543.
I've heard his dac before he skipped the DIR. Maybe on his system I would prefer his implementation of the TDA1543...

So....you've got to listen and try for yourself.

Cheers,

[Guido Tent]

Originally posted by king30
Jean-Paul,
Do you mean something like this...Thanks Lee

Hi

Why the caps ?

Guido

[Guido Tent]

Originally posted by Elso Kwak

Hi jean-Paul,
If I remember correctly Fred told us the transformer should be in the transport and the 74HCU04/ 74HC86 at the DAC end of the transmission line preventing the CS8412 spitting back into the cable. Of course my interface with AD8561 would also work.

dear all,

The RF voltage at the RXN / RXP pins of the Crystal (at least the 8412) is due to groundbounce (tens of mV in the time domain). Hence, it is a common mode effect, as both pins are close to each other and refer to the same substrate, locally that can be treated as an equipotential plane.

A solution is the use of a well balanced transformer at the input, like Jean Paul points out. Ofcourse this transformer needs to be well connected to ground, RF wise that is. We use a capacitor from secundary mid tap to ground. The primar ground is directly connected to the cabinet such that common mode currents can interact with other cables without literally crossing the inner electronics of the DAC. All works fine up to 100's of MHz

All abouth the transformer can be found in the link as shown by JP.

Feeding the 8412 single ended is asking for problems to my opinion.

Guido