Atoll DAC200 Signature - Review & Measurements (DAC)

[Interference]

I am not surprised at all: had in my hands an Atoll DAC 100 once, and from a quick measurement with REW and an E-MU 0404 I could see pretty similar distortion patterns. It seems nothing much has changed since then.

 

[K0RnY]

Great review, as always !
I'm asking myself if an Emotiva XDA-2 Dac would do a better job...

 

[JSmith]

Ok, sent a mail with the link

So did you get any kind of reply to your email of query?


JSmith

 

[Milanraf]

Do you think it would be feasible to measure the signal before the analog output stage? Just to understand where the design failure is, at least in terms of THD and IMD (less for HUM).
I suspect the most part of the problems lays in the discrete gain stage, rather than in the actual DAC implementation (which would be even worse, in respect of designers capabilities).

 

[VintageFlanker]

Do you think it would be feasible to measure the signal before the analog output stage?

It would not be possible. A working DAC (product) is a combination of D/A chip, filtering, clocks, output stage and so on...

I suspect the most part of the problems lays in the discrete gain stage, rather than in the actual DAC implementation (which would be even worse, in respect of designers capabilities).

I have no clue, and that is none of my business, honestly. My contribution is to measure an actual product to buy. Not to investigate about why it does or doesn't perform. That should be the manufacturer's job...

 

[boXem]

Do you think it would be feasible to measure the signal before the analog output stage? Just to understand where the design failure is, at least in terms of THD and IMD (less for HUM).
I suspect the most part of the problems lays in the discrete gain stage, rather than in the actual DAC implementation (which would be even worse, in respect of designers capabilities).

What are you calling "DAC implementation"?

 

[Milanraf]

It would not be possible. A working DAC (product) is a combination of D/A chip, filtering, clocks, output stage and so on...

I have no clue, and that is none of my business, honestly. My contribution is to measure an actual product to buy. Not to investigate about why it does or doesn't perform. That should be the manufacturer's job...

Didn't mean to make you work. Your job's over, on a blackbox testing perspective.

I just proposed that for the love of science. In front of such a broken product, the curiosity rises.

 

[Milanraf]

What are you calling "DAC implementation"?

 

[boXem]

View attachment 283445

This diagram explains your question. And also why you think that the "implementation" is the easy part. But it is quite misleading.
In the case of this DAC, the "DAC" on the diagram you attached is an ES9028PRO chip, one or several clocks, power supplies for the digital part of the chip, power supplies for the analog part of the chip, an I/V conversion stage and it's power supplies, a circuit board with proper layout, ...
A lot of reasons to screw up a design before reaching the output stage.

 

[Milanraf]

This diagram explains your question. And also why you think that the "implementation" is the easy part. But it is quite misleading.
In the case of this DAC, the "DAC" on the diagram you attached is an ES9028PRO chip, one or several clocks, power supplies for the digital part of the chip, power supplies for the analog part of the chip, an I/V conversion stage and it's power supplies, a circuit board with proper layout, ...
A lot of reasons to screw up a design before reaching the output stage.

My electronics professor used to say: "almost every newly graduated can design a proper low* frequency digital circuit, there's loads of literature, patterns and working examples to inspire and steal from, but analog is another story. Today a good analog designer worths like gold".

ATOLL decided to take the hard road: design a discrete gain stage.

* sub GHz

 

[Scytales]

Do you think it would be feasible to measure the signal before the analog output stage? Just to understand where the design failure is, at least in terms of THD and IMD (less for HUM).
I suspect the most part of the problems lays in the discrete gain stage, rather than in the actual DAC implementation (which would be even worse, in respect of designers capabilities).

To my mind, the harmonic distortion spectrum of 1 kHz fundamental tone (2 kHz, 3 kHz and so on) are the actual linearity performance of the discrete output stage. Numerous third parties measurements of digital and analogue products built by Atoll show the same type of spectrum. It can be safely stated that this level of performance is obtained by design (for whatever reason).

But the 100 Hz spaced side-bands under and above 1 kHz that form the "grass" on the FFT, it's another matter. As was discovered by jipihorn on another implementation of a previous model of ESS, an ES9018 (see the 2 videos I have posted above, but there are in French), those tones are perhaps the manifestation of some jitter affecting the rise and falling edges of the width modulated square signals at the output of the DAC, said jitter being produced by some residual hum over the 3.3V Vref DC power supply of the DAC. Thus, it can be hypothesized that there is a mistake in the implementation, or failure, of the integrated 3.3V regulated power supply ES9011 Atoll uses for powering Vref.

 

[boXem]

My electronics professor used to say: "almost every newly graduated can design a proper low* frequency digital circuit, there's loads of literature, patterns and working examples to inspire and steal from, but analog is another story. Today a good analog designer worths like gold".

ATOLL decided to take the hard road: design a discrete gain stage.

* sub GHz

Did you actually read my reply?

 

[JSmith]

ATOLL decided to take the hard road

After being caught out being misleading about the spec's of their devices... one would imagine they will be on said hard road for some time.


JSmith

 

[VintageFlanker]

So did you get any kind of reply to your email of query?

I would be really interested yo know as well.

I have zero doubt that Atoll is aware of the review by now. Still waiting for them to chime in and discuss.

 

[Lupin]

I do see in their published Manufacturing Process that they seem to check the quality control by... listening? I have no issue with that, but, would you guys at Atoll care to also measure? Definitely you should.​

They make it sound manually soldering is actually a good thing..

They do QC and "tuning" by listening... I don't know but I like my DAC to be transparent to the source. Any and all colouring of the sound should come from speakers/headphones used and/or EQ. Definitely not from a DAC, that is not the function of a DAC.

 

[renaudrenaud]

So did you get any kind of reply to your email of query?


JSmith

Yes.

But, I thrashed the mail without reading it. I guess it's a waste of time for each side. If they want to expose their position, they can come here.

Few days ago I was looking for some information and started to read this forum https://www.ls3-5a-forum.com/

Where I learn:

• @amirm is sold to the Chinese HiFi
• when any gear is excellent in term of measurements, this is the main reason to avoid it
• etc...

So I understood there is many way to be passionate on the subject. It depends of your church, magic or science, why to believe more in science than in your own ears bla bla bla.

 

[Trell]

But, I thrashed the mail without reading it. I guess it's a waste of time for each side. If they want to expose their position, they can come here.

But why send an email in the first place if you don’t want to read the reply?

 

[K0RnY]

But why send an email in the first place if you don’t want to read the reply?

And eventually share it with the ASR community, that was waiting to know Atoll position about the bad result of the product measured

 

[ChristianL_Mtl]

Thanks @VintageFlanker

I own the DAC100 that I bought a while ago, so I'm glad to finally see some measurements of their products. I assume mine is probably even less performant as it is their entry-grade product.

What I really like with it though - I cannot compare it sound wise to anything - is the multiple inputs. I use one for my Apple TV and the other one to receive audio.

Is there something in Topping and SMSL kind of companies that would do this but with top grade performance and with a reasonable price ?

Best wishes,

 

[VintageFlanker]

Alright folks,

At least something finally changed:

DAC200 Signature - ATOLL ELECTRONIQUE

The DAC200 Signature converter was implemented to offer you the best of your favourite music. Its essence of musicality resides in the line signal processing stages, the power supplies, the digital signal processing before conversion, and the quality of the key components.

www.atoll-electronique.com

Specificated THD+N went down from -120dB to now 0.002% (-93.9dB).

I'm not sure if I am happy to see them getting closer to the truth... without taking the time to explain why they did publish these specs in the first place. Finally, the "subject to change" disclaimer wasn't for nothing.

PS: It appears to be the case since May 2. (according to @jipihorn screenshot), a couple of days after the review was published.
For the record: