The well segmented dac – DSD discrete dac

[The Well AudioParticipant]

TWSDAC-DSD is running, driven by the FIFO Lite.

The DAC is under test, it plays up to DSD128 with 5/6 MHz oscillators, up to DSD256 with 11/12 MHz oscillators and maybe DSD512 with 22/24 MHz oscillators (to be tested).
The FIFO Lite also accepts DoP.

After the test there will be a listening section to compare the TWSDAC-DSD with the DAC Lite.

Stay tuned!

Attachments:

You must be logged in to view attached files.

[The Well AudioParticipant]

The new DSD discrete DAC is now available.

There are two version of the DAC:

– with standard thin film resistors

– with tantalum nitride resistors

Attachments:

You must be logged in to view attached files.

[G600Participant]

Dear Andrea,
Apart from going balanced, what is the advantage of this DAC over the PCM one?

Is it desirable to have the FIFO input pcm and output dsd, or have the rpi making the conversion?

And could you describe the addition of the tantalum resistor?

Thanks, and bravo for the designs!

• This reply was modified 1 year, 1 month ago by G600.

[The Well AudioParticipant]

Both R2R and DSD DAC sound very good, maybe the DSD DAC is more precise and detailed while the R2R DAC has better midrange.
Difficult choice, maybe I prefer the R2R DAC for simple acoustic music and the DSD DAC for complex music like orchestra.

Anyway both DAC sound a little better with tantalum resistors, although they already sound very good with standard thin film resistors.
With tantalum resistors the R2R DAC is more detailed, while the DSD DAC has more midrange fullness.

[The Well AudioParticipant]

For those who are interested on the DSD discrete DAC, we have reached the best sonic performance when the DAC is driven by the FIFO Lite with 22/24 MHz DRIXO oscillators and RTZ enabled in DSD256 input format.

[The Well Audio]

The Well Audio Said

For those who are interested on the DSD discrete DAC, we have reached the best sonic performance when the DAC is driven by the FIFO Lite with 22/24 MHz DRIXO oscillators and RTZ enabled in DSD256 input format.

I suspected something like that, or DSD128 with 11/12 clocks.

A very good way to reuse frequency doublers with 5/6 DRIXOs

[The Well AudioParticipant]

https://www.diyaudio.com/community/threads/the-black-hole.349926/post-7306229

It’s truly amazing that some still hate everything we design, despite Andrea Mori and The Well Audio having been banned for a long time.
It is also very incorrect, since we don’t have the possibility to reply.

We, on the other hand, are correct people, so we invite all detractors, even those who hate us, to post on this forum.
Maybe we can have a civilized discussion.

[KazumaParticipant]

If you set out and want to try to convince those people, don’t bother.
If someone is willing to learn, and know, and change and advance, that’s one thing.
But the type of persons that you’re dealing with, they just want to ruin everything, and shit on everything, and make whoever oppose their beliefs miserable.
Especially, if it develops a pattern, repeating over and over, that’s when you know for sure.
These are people that they couldn’t even have it in them, to learn or think.
It doesn’t matter, no level of evidence or truth or whatever will change the ideology.
As a matter of fact evidence means nothing with these people, so why bother.
In other words in a large scale, you do it for yourself, and people that are close to you, and your friends.
Focus on those, who appreciate you, and find your work inspiring. And who are grateful for your gifts and your help in advancing this great quest for music joy.

[The Well AudioParticipant]

Thanks for your kind words.

I know them, they are not engineers, they only are obtuse people who talk only about measurements and then they are not even able to read the measurements they are pointing out.

Someone referred to Auralic products and mocked Mark (and his friend) for his statements.
Well, they are not even able to compare a few basic measurements in order to understand that Auralic devices are nothing special although very expensive.
Let’s pick the LEO GX.1 Reference Master Clock, the Femto seconds master clock as they claim.
From their website:
Equivalent Phase Noise: -110dBc/Hz@1Hz (Ref Frequency: 10MHz)

Well, our 10 MHz DRIXO oscillator measures -120dBc/Hz@1Hz, 10 dB better than the Auralic.
Of course, we have published the phase noise plot, they don’t.

To put it like Mark, DRIXO “destroys” Auralic.

Just the last curious thing: our DRIXO costs 251 Euro (crystal + finished board), while the Auralic Leo costs 9899 Euro!

But perhaps such engineers need to go back to school to figure out how to compare the measurements they are very fond of.
And if they don’t like what I wrote they can always come here and reply, they are welcome.
But they won’t.

• This reply was modified 1 year ago by Moderator.

[The Well Audio]

The Well Audio Said

https://www.diyaudio.com/community/threads/the-black-hole.349926/post-7306229

It’s truly amazing that some still hate everything we design, despite Andrea Mori and The Well Audio having been banned for a long time.
It is also very incorrect, since we don’t have the possibility to reply.

We, on the other hand, are correct people, so we invite all detractors, even those who hate us, to post on this forum.
Maybe we can have a civilized discussion.

I wouldn´t bet on a civilized discussion if they post in this forum as well. Even some “moderators” participated in your banning…it is very obvious that they were not very objective in this matter. This always surprised me….

On the other hand envy is so baddddd….. 😉

[MarcelvdGParticipant]

Andrea, you used to dislike anything sigma-delta. Why did you decide to design a DSD DAC, as DSD is a single-bit sigma-delta modulate?

[G600]

G600 Said:

A very good way to reuse frequency doublers with 5/6 DRIXOs

 

 

If you use a frequency doubler for a DSD DAC, beware of subharmonics mixing with sigma-delta idle tones,

https://www.diyaudio.com/community/threads/the-well-tempered-master-clock-building-a-low-phase-noise-jitter-crystal-oscillator.261651/page-143#post-6373471

Then again, Andrea’s DSD DAC looks like a long FIRDAC. As such, it should be much less sensitive to this issue than the rather basic model I simulated.

[The Well AudioParticipant]

Marcel, you are welcome.

Right observation since I wrote that I would use sigma delta DACs for cooking a pizza.
When I was in Auburn last year Mark and Jam pressed me hard to build a DSD DAC.
Even though we weren’t very convinced, especially Roberto, we finally decided to give it a try.

In the end I have to change my mind, after the last adjustments the DSD DAC sounds very good.
Roberto still prefers the PCM DAC, I prefer both, so much so that in the Sonic Empire version we are working on we will implement both with some improvements.
IMHO the DSD DAC it’s more detailed and the soundstage is better, while the PCM DAC has better midrange.
Depending on the music I listen to I prefer one or the other.

We had a listening session just last tuensday, the differences between the two DACs are getting smaller and smaller.

Both DAC were driven by the FIFO Lite with DRIXO oscillators.
For the DAC Lite we have used a Zanden filter follewed by a diamond buffer.
The DSD DAC was configured as folow:
– Laptop
– JLSounds I2SoverUSB
– Simple DSD converter to DSD256 (FPGA project posted on diyaudio)
– FIFO Lite working in RTZ mode
– 5/6 MHz oscillators followed by a pair of doublers for each sample rate family to reach 22/24 MHz
– differential to single ended buffer

We also have tried single ended output from the DAC, decoupled with 33uF Mundorf EVO oil. Maybe it sounds still better, but the output voltage was too low (divided by 4, RTZ and single ended).

There was a clearly audible difference when we have used the 22/24 MHz DRIXO directly. With 5/6 MHz DRIXO and doublers the DSD DAC sounds better. Which means that subharmonics don’t matter much.
Obviously the FFT wouldn’t detect any difference, but that’s another topic.

• This reply was modified 1 year ago by The Well Audio.

[SebastianParticipant]

Hello Andrea!

 

Given that your last comments on the DSD DAC where back in April, could you give us your impressions after all this time comparing the DSD and the PCM DACs?  Just before the GB closes, it could be interesting to hear what you think.

How would you compare the the SE (/ + TWTSEBX4) and balanced versions of the DSD dac?  What is the Vout of the twsdac-dsd-se alone?

Regarding the new TWSAFB-XMOS, does the older FIFO still have sonic advantages over it (particularly with the DSD DACs, which it seems more aimed at)?  A think the lower freq clocks alone would be better, but are there more advantages to the FIFO?

“For the DAC Lite we have used a Zanden filter follewed by a diamond buffer.”  Is the Zanden filter part of the product line? Could you comment on its use?

 

[The Well AudioParticipant]

DSD and PCM DAC have different sonic signature.
Both DSD DAC are more detailed and precise than the PCM DAC and maybe the soundstage is wider and deeper.
PCM DAC has better midrange and in our opinion is more dynamic than the DSD DAC.
In the end I believe it’s a question of personal taste.
We like both DSD and PCM DAC, that’s the reason we will implement both in the Sonic Empire version of the DAC.

The TWSDAC-DSD-SE DAC is the entry level of the Lite series DAC. We designed it to offer an affordable DAC which sounds anyway very good.
Of course, the balanced version of the DSD DAC driven by the FIFO and 5/6 MHz DRIXO oscillators followed by 4 doublers to reach 22/24 MHz sounds better, but it does cost almost 3 times.
Does it worth?Difficult to say, I would start with the cheaper DSD SE DAC in order to understand if I like the DSD sonic signature.

TWSAFB-XMOS USB receiver and TWSAFB-LT FIFO buffer share almost the same architecture in their clean side of the circuit.
The FIFO buffer is more versatile. It has 4 x I2S selectable inputs, while the XMOS receiver has only one USB input. The FIFO buffer ca drive almost all the DAC (old and modern), while the XMOS board can drive only standard DSD and I2S DAC. For example the XMOS receiver cannot drive the PCM DAC Lite and the balanced DSD DAC.
Although the TWSAFB-XMOS can drive standard DSD and I2S DAC, it was designed specifically to drive the TWSDAC-DSD-SE in order to offer a good DAC at an affordable price.

The Zanden filter is not part of the Lite product line, it will be implemented in the Sonic Empire version of the DAC only.

[Author]

Posts