Atoll DAC200 Signature - Review & Measurements
View attachment 282201
Hey folks,
Planned for a while, finally comes my first review about an Atoll product, and the first extensive measurements of any Atoll DAC to date. For those who are not familiar with the name,
Atoll Electronique, while not widespread worldwide, is an extremely popular manufacturer here in France. They initially started to design amplifiers by the end of the 90's, then the brand extended their offering with CD Players, DACs and Streamers.
So, why the interest? Being not that old, I hear about Atoll since the beginning of my Hi-Fi journey. Based on my own observations of the market, I personally feel that it remains somewhat a polarizing brand, along with both faithful customers and haters (I am neither). The first category seems to always point their purchase decision with one major argument: it is Made In France. Yes, it is.
Their entire line of products is assembled at Brecey, Normandie. Having Hi-Fi electronics designed and built local, that are respectively priced to remain competitive against big(ger) names like Cambridge, Marantz, NAD or Rotel, is certainly a good thing, and I understand those who want to reward it. Yet, the place of manufacturing is nowhere near a guarantee of good design, nor actual performance. And that is where I feel that objective analysis is needed. If Atoll manages to design and build products locally, and if they also perform great, then it would be a great story to tell !
The
DAC200 Signature, reviewed today, is their second most expensive DAC, priced at
1790€. Then comes the
DAC300 at 2595€. In this day and age, these are far from cheap offerings, and we are right to expect stellar levels of performance from them to be competitive. This particular sample has been bought a few days ago by a friend (
Mikycoud on other forums) and kindly drop-shipped to me for the purpose of this review. The DAC200 integrates an ESS ES9028 PRO D/A chip, and I am glad to see the brand communicating about specs. On the other hand, I feel kind of afraid that they have just taken the specs from the ESS datasheet and call it done:
"-120dB THD+N, 133dB Dynamic, 129dB SNR". Sounds good, doesn't it? I want you to keep these numbers in a corner of your head for what comes next...
The Atoll DAC200 Signature comes in an elegant and quite huge chassis, in fact as big as some integrated amplifiers. It is 440mm wide, as is the rest of Atoll amplifiers and DACs (expecting more compact products). It is relatively heavy for its size (5Kg) and build quality is mostly good. To nitpick a bit, while the front machined aluminium plate is beautifully finished, the rest of the chassis seems to be made of sheet steel and somewhat lacks rigidity.
The buttons also feel a bit loose on touch. You have most of the controls needed for a DAC, from left to right: Input select, filters settings, volume control and standby. You may notice the circuit around the screen area because of the lighting on pictures, but it really appears to be all black under regular light conditions. I like the display, since it offers decent contrast and visibility from good distances. At last, you also have some hardly noticeable headphones 3.5mm output beneath the settings buttons. The DAC200 also comes with a
classic remote control, designed to be paired with other products of the brand.
Do you see that? Yes, it is quite uncommon on any DAC to have nine digital inputs available! There are three coaxial and three toslink inputs, which is quite welcomed when you have several digital sources (TV, BD Player, Streamer, CD Transport, Consoles etc.) to deal with. I would like to see it more often, and I praise Atoll for this. Then, there is some regular USB-B input, AES/EBU, and at last a Bluetooth antenna. It does not say which codec are supported (more on that latter).
The rest is pretty regular: XLR and RCA analog outputs, and both optical and coaxial digital outputs. At last, I have a little remark about the markings, tho. The fact that these are beneath connectors make them very difficult to see when looking from above.
Measurements
Disclaimer: Measurements you are about to see are not intended to be as precise or extensive than what you get from a 30k€ AP. There is obviously both hardware and software limitations here, so not quite apples to apples comparison with Amir's testing. Still, this data is enough to have a pretty good idea if the gear is bad or not, stellar, broken, or sub-par...
- Instruments : RME ADI-2/4 PRO SE. E1DA Cosmos ADC (Grade B), Minimum phase filter. E1DA Cosmos APU 60dB preamp is used for DR measurements. Output voltage is measured separately, using a DMM with a 0dBFS 1kHz tone.
- Software : REW V5.20.14, Multitone Loopback Analyzer 1.0.80 and RMAA 6.4.5 PRO,
- Method : 8 runs for each test, then I choose the closest to the average. Bandwidth and sampling rate to be specified for each test.
The Atoll DAC200 has both a fixed output setting (called "bypass") and a variable one, with 80 volume steps. I will start measuring the first, through XLR balanced output and USB input.
Tests | Results (L & R) | Rating |
Noise Level (REW) | -115.2dBA | Very Good |
Dynamic Range (REW) | 112.6dB | Very Good |
THD (REW) | 0.03890% | Very Poor |
THD+Noise / SINAD (REW) | -68.2dB | Very Poor |
IMD SMPTE (REW) | -55.7dB | Terrible |
Stereo crosstalk (RMAA) | -114.6dB | Very Good |
IMD+Noise @10 kHz (RMAA) | 0.038% | Very Poor |
Multitone 32 TD+N (Multitone) | -63.3dB | Very Poor |
Alright. Seems like we got a lot to talk about here. What the heck is going on exactly? Apart from Crosstalk or raw Dynamic Range/SNR, (which by the way are good but not quite what I would expect from a DAC in this price range), everything else is simply wasted.
I start with my first concern: Why is the company even advertising "-120dB" THD+N, while we get -68dB at the end? Do people at Atoll simply believe that an ES9028 PRO will always perform at its best, regarding of the implementation? Do they even know how to measure THD+N to begin with? I do not feel sorry to ask, when I look at this complete failure. Speaking about it, did you know that bad measuring products are more time-consuming to review than ones that perform right on specs? Why? Because, I have to spend a hard time trying to understand what is wrong here and there. Is my ADC clipping? After trying several levels from both inputs and outputs, it appears to be not (both Cosmos and ADI-2/4 PRO). I see in the manual that Atoll advises to be careful about the phase. Well, I reversed my Schuko plug one side or the other and got the exact same mess over and over. Are we dealing with some ground loops, regardless of the so-called balanced signal? Performance is the same using toslink (more on that latter). Everything led me back to the common denominator: a simply bad measuring product.
Apart from THD+N, the Intermodulation Distortion is acting crazy, with plenty of grass between both fundamental frequencies. The IMD Versus Level test does show the presence of ESS Hump, before growing up way too early. The THD Versus Frequency is a funny one: For no reason, it starts to perform (in some unremarkable way) from 8kHz, before clipping completely at 22kHz. In the same league, the IMD Vs Frequency (swept CCIF) behaves the opposite way it should, while slopping down until 20kHz. I cannot believe, in any shape or form, that this product is performing that way on purpose. Even when trying to understand how it could have been designed that way to please subjective listening, I am not able to find any valid reason. In my book, it is just broken. Period.
This first bunch of tests has been performed with the fixed line out, that is, at 4.6V, slightly higher than regular 4.0-4.2V XLR I am used to measure. What happens when I turn down the variable output by a couple of step?
For some reason, the results are slightly better with reduced volume. Do not get me wrong, it is still acting poorly nonetheless. But this already rises some questions: why is the bypass output, advertised (supposedly) as the best setting as a line out would actually be... worse?
Let's move to S/PDIF. I just wanted to make sure that not only USB-B input (or overall XMOS implementation) would be flawed over Coaxial and Toslink input. I does not seem to be the cas, since S/PDIF performs equally bad:
SINAD did not change much, while being a touch lower. At least, it is cohesive in that regard. Maybe, Jitter could behave differently...
Nope. The good news here is that both Coaxial and Optical are performing the same each other, and virtually very similar to USB. The bad news is that all three offer far from acceptable performance.
I also measured (in digital domain) the Coaxial and Toslink output through my ADI-2/4 PRO, hoping that not all would be messed-up. The Atoll DAC200 cannot miss this one, can it?
Phew! This is the first thing so far that the DAC200 is doing right. What you see here is typical digital performance of what I may capture out of my S/PDIF interface (should ideally be around 141dB SINAD, but it is not related to the Atoll). But, (yes, there is a "but")...
Note: I cannot understand why, nor that if it should be activated somewhere in settings, but both S/PDIF outputs do not work when running USB input ! I have never experienced such a thing with any product with digital outputs I tested so far. Even when using Bluetooth input, digital outputs work normally. Speaking about Bluetooth...
I found somewhat funny that Bluetooth (By the way limited to standard SBC Codec, which is a pity) almost reaches the performance of what you would get wired. In fairness, it is not that this specific performance is particularly "good" by any means (it is indeed
what to expect from digital SBC performance), but more likely that others inputs are so bad that they are just a hair better than the weakest BT Codec around. Quite a shame, indeed.
Well, could things be even worse at this point? Yes, they can. We still need to find out the performance, unbalanced....
Tests | XLR | RCA |
| Output Voltage @1Khz | 4.664Vrms | 2.346Vrms |
Noise level (REW) | -115.2dBA | -97.7dBA |
Dynamic range, dB (REW) | 112.6dB | 82.6dB |
THD (REW) | 0.03890% | 0.11350% |
THD+N (REW) | -68.2dB | -58.9dB |
IMD SMPTE (REW) | -55.7dB | -48.8dB |
Stereo crosstalk (RMAA) | -114.6dB | -91.1dB |
IMD+Noise @10kHz (RMAA) | 0.038% | 0.053% |
Multitone 32 TD+N (Multitone) | -63.3dB | -66.3dB |
You have the whole picture in the tabs. Performance drops quite significantly over single ended outputs in all regards (but Multitone). The first victim is obviously SINAD:
OK... I did not think RCA could be more flawed, yet it is, and by quite a margin. Now, the second harmonic is dominant, and I could not imagine this has been done on purpose whatsoever. Why, then, XLR would not share the same harmonic profile?
You get a lot of hums at 50, 100 and 200Hz. These are so strong that some may think about the DAC200 to be sensitive to ground loops. Well, it would be very unlikely to be the case. First, these tests are done through a battery-powered laptop and with galvanic isolation over USB (Topping HS-01, also running itself on separated battery). Second, I get the exact same nightmare results when using optical... Besides, there would be no excuse, anyway, for any DAC in this price range to not be immunized against this kind of issues.
Things are getting even worse with IMD:
No need to comment on this: It is a complete mess, with uncontrolled distortion all over the spectrum.
Strangely, RCA Multitone is displayed as "better" than XLR, or at least, when just looking at raw TD++N number. In reality, the hum at 100Hz is still dominant and has no excuse to be that strong in this specific test. You can observe the strange shapes (aliasing?) in the N+D floor, that I personally did not witness in any Multitone graph so far.
Note: The 3.5mm jack headphones output gave me the same results as the RCA outs (exact same SINAD and output voltage for both Right and Left channels). My guess would be that there is no dedicated circuit or any "amp" to speak about, as it seems to share the same path as SE outputs. Considering how bad it is already, I did not bother to investigate any further...
Atoll does not communicate much about any dedicated preamp output stage/circuit. I then assume that we are dealing with regular digital attenuation from the ESS DAC. While we should expect deterioration of THD+N when the level moves away from 0dBr, the DAC200 behaves... differently:
THD+N Versus Level
Volume Level | Output Voltage @1kHz | THD+N |
80/80 | 4.664Vrms | -68.2dB |
78/80 | 4.161Vrms | -70.4dB |
75/80 | 3.502Vrms | -73.6dB |
70/80 | 2.626Vrms | -78.8dB |
60/80 | 1.477Vrms | -89.0dB |
50/80 | 0.831Vrms | -93.6dB |
40/80 | 0.312Vrms | -87.8dB |
30/80 | 0.098Vrms | -78.4dB |
I do not understand how the DAC200 works on that field... Maybe there are different analog stages regarding of volume, or maybe the performance is simply random. Surprisingly, we almost get half-decent results when using the 50/80 volume step, that comes with some 93.6dB SINAD. Infinitely better that what we have seen so far. Needless to say, it still is unremarkable, not only for the class, but you would not want some weak 0.8Vrms signal as your regular DAC output. But ultimately, yes, the DAC200 is doing a better job as a preamp than as a DAC.
I was still wondering if, perhaps, filters would have any impact on the performance, since I have another product next to be reviewed that show a few dB difference in THD+N depending on the Filter setting. Sure, it should absolutely not be the case in any competent product, but maybe the Standard filter could be part of the issue here.
No. Each filter measure the same, as they should. Of course, only Frequency Response acts differently, as displayed below (same amplitude for each FR, separated):
That is all, folks. I think we do have more than needed data at this point.
Conclusions
What could have happened here? So far, I would set up three hypotheses :
1) The DAC200 Signature performs exactly as it should. Atoll designed and tweaked it only by ears and finds that it sounds accordingly to what they wanted to. Then, I wonder why they are publishing any performance specs at all. By the way, all of them are plain wrong. THD+N, SNR and DR certainly are. In fact, even output voltage is: they say "2.25Vrms", while we get 4.66V with XLR and 2.34V through RCA. 2) They wanted to design a device, performing as advertised, but they simply have no clue about how to engineer it. Pity, then, not to mention at that price... 3) This particular unit I have my hands on is faulty. Honestly, I do not buy it, but... Then, I wonder how such a thing would have happened. How a 1800€ DAC would be able to leave the factory without any care for quality control? 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.
Time for my two cents: This is the worst DAC I ever measured, regardless of the price and class. Period. The DAC200 Signature is acting way off its own specifications, and I find no excuse whatsoever that could explain it. In a perfect world, local manufacturing, if paired with competent design, should be an assurance of an actually good product to buy. In that case, it is not, since it seems like the second requirement is lacking. What a bummer.
: Suited input impedance to match more DUTs. IMD VS Level without Hump of its own. Flatter Frequency Response with most DUTs I tested. Depending of the DUT output voltages you are dealing with, the RME could also be more precise, or closest to 0dBFS. Thanks to its Auto Ref Level, you know immediately which step to use (need sometimes adjustment to prevent clipping). Overall, better performance for measuring single ended signals in my experience. Better performance by quite a margin for lowest signals (because of the 1dBu-0.87Vrms lowest input sensitivity VS 1.7Vrms for Cosmos). Ability to perform measurements in digital domain (loopback recorder)... And last but not least, you got a display to observe what is going on when measuring.
