ECdesigns

[Qhwoeprktiyns]

Rest assured, I have as much difficulty as you grasping these concepts - I've probably just spend more time trying to do so. Having used ECD products for the past few years, I am always curious, as most of us probably are, to correlate what I hear with what ECD explains with each new development. It is rare to have a DAC manufacturer be so open about their designs and provide so much information. It is also rare that they explain the issues with previous models, or what they have "understood" since, and how they improve on them. 

 

To answer you, I am pretty confident that:

PowerDAC + UPL > Current DA96ETL + UPL. 

And the DA96 >> MOS16 (i have both, and other trusted ears have confirmed). 

Question will be:

PowerDAC + generic source = PowerDAC + UPL ? 

[Superdad]

7 hours ago, hopkins said:

It has a "master clock", so it does not rely on the clock signal of the source, and it therefore "re-clocks" the incoming signal. The way it does this is novel, and has not been done before (at least according to ECD). They found a way to do it "asynchronously" (as with USB), so that clock and data signals remain "separate", in a sense (my very limited understanding) and this avoids noise spreading from one to the other... . So essentially only the "data" part of the Toslink signal is used, and the rest (clock signal) is discarded. 

 

Just to be clear, asynchronously reclocking S/PDIF (or TOSLINK or AES/EBU) is not new. Many DACs, S/PDIF reclockers, and digital converters offer this. There are several ways to do it--some easy, some hard--and some of methods are better than others.

(And this is ENTIRELY different and not at all related to asynchronous isochronous USB, which has nothing to do with clocking and only is with regards to communication back to the computer about how much data to put in each packet.)

[Qhwoeprktiyns]

54 minutes ago, Superdad said:

 

Just to be clear, asynchronously reclocking S/PDIF (or TOSLINK or AES/EBU) is not new. Many DACs, S/PDIF reclockers, and digital converters offer this. There are several ways to do it--some easy, some hard--and some of methods are better than others.

(And this is ENTIRELY different and not at all related to asynchronous isochronous USB, which has nothing to do with clocking and only is with regards to communication back to the computer about how much data to put in each packet.)

 

Yes, I do think it has to do with the method, as I clearly stated, more than the fact that it is done asynchronously. However, do you know of any DACs that does this while dispensing of (high bandwidth) I2S communicaton altogether ? The PowerDAC's internals rely purely on a very low bandwidth (100kHz)parallel data signal to feed the DAC. 

 

If you achieve asynchronous reclocking of spdif using micro-processors, but fail to deal with RF issues you are back to square one and might as well use a PC and USB connection. 

 

A similar 'bandwidth limiting" technique was partially retrofitted to their player (UPL) and I actually have both the original model and the revised model, so was able to compare both, and the difference between is not subtle. A step in the right direction, which led them to develop this even further in their new DAC. 

 

Its a combination of techniques, as I understand, that have been used in ways that are "novel". 

 

I am not saying this is the only aporoach possible. Heck, I have not even heard the results yet... But I am sure you would agree (as your business does offer solutions to this) that currently no DAC offers source immunity, far from it. With USB, you cannot isolate the source noise in the DAC's reciever, or at least not completely. John Swenson himself states this in the white paper you have on your website. Once noise makes it through, it spreads...So we are left with "source optimization" in various forms, going from using your very affordable products, to spending >20.000€ on a low noise PC... An alternative approach is always welcome. 

 

In fact, not having compared their current player+DAC with many other solutions (some, but obviously not all) I am more interested in seing how it performs relative to this current model. It will also be affordable, and that's good, i don't feel too guilty talking about it. I am not going to be buying a Taiko PC to find out which one is best :) I'm happy with the sound I get now, but if I get the same SQ (or even perhaps a small improvement) with a network player, bingo! 

 

Vis à vis the parallel with USB, I understand USB is an entirely different protocol, I was just making the parallel to its asynchronous aspect, as opposed to spdif.

[Qhwoeprktiyns]

And by the way, its nice to get your input. Unlike some, I don't believe you can't have a valid point of view on an audio product unless you have heard it... Otherwise, there would be very little discussions going on in this forum. We just have to remain cautious.

[mevdinc]

On 3/31/2021 at 2:46 PM, hopkins said:

 

The technical explanations (given on their website R&D section and some added points here) are really interesting, though they are not always so easy to grasp ! I've done some reading on these topics recently, and what strikes me is how the very nuanced and cautious explanations that engineers give concerning digital audio sometimes translate into not so nuanced and cautious claims by the consumers (us) of these products (ex: reading recently that a product has "close to perfect galvanic isolation of the USB input"). So I'll remain cautious, and plan to give this new product, when it comes out, a full battery of tests/comparisons and have as many trusted ears as I can give their opinion about it.

 

As a result of their communication about this, I don't think too many people are buying their current products any longer (not that there were many to start with!), so that must be an incentive for them to come out with something fairly soon :)

 

I for one don't care much for the technical talk since mainly because it goes right over my head. I am a software person and I'd be happy to read other people's experiences and professional reviews/findings. If it all sounds interesting enough then I will listen to it myself and then decide if I like it enough or not.
There are some clever people out there, trying to do things differently to create better sounding products at reasonable price points. This is what interests me about EC. We'll see what they come up with, shouldn't be that long now, hopefully. :)
I am in no rush, as I have a decent system at the moment that gives me plenty of listening pleasure.

[Norton]

An unfortunate byproduct of EC designs‘s  approach is that the focus of opinion will always be on the next generation of product.
 

I will almost certainly go for the PowerDAC S when released, but the current DAC96+UPL combo really is outstanding. After nearly a year of ownership, the clarity, detail, lack of noise and distortion etc on plain 16/44 really puts it in another class from any other digital playback chain I’ve owned or heard, regardless of price, format, upsampling  etc.  I own a fair few SACDs and not only does the ECD combo best my Resonessence  Mirus (the latter via its onboard SD transport) on 16/44 and 24/96 files  but, when dealing with recordings from the last decade or so at least,  rips of the SACD  RBCD layer played via the ECD kit are to my ears a clear improvement on rips of the stereo DSD layer played via the Mirus.  Downsides are that preparing thumb drives will not suit many and the ECD combo is equally successful in exposing poor recordings as it is in highlighting  excellent ones.
 

It would be great if more people had listening experience with the ECD kit in their own systems and if some were able to offer comparisons vs. what are considered elsewhere to be SOTA digital replay chains.  My suspicion is (but I don’t have the breadth of listening experience to be certain) that ECD really have pulled off something very special, a categorical improvement in digital replay hiding in plain sight in the form of these modestly priced boxes.

 

 

[numlog]

 

On 4/1/2021 at 10:30 PM, hopkins said:

 

Yes, I do think it has to do with the method, as I clearly stated, more than the fact that it is done asynchronously. However, do you know of any DACs that does this while dispensing of (high bandwidth) I2S communicaton altogether ? The PowerDAC's internals rely purely on a very low bandwidth (100kHz)parallel data signal to feed the DAC. 

 

If you achieve asynchronous reclocking of spdif using micro-processors, but fail to deal with RF issues you are back to square one and might as well use a PC and USB connection. 

 

A single 44.1kHz sample has 16bits of data per channel, afaik that means you need a clock running at 705.6kHz or the data will arrive faster than the music can play.
I suppose ECdesigns data protocol means two 16 bit frames are clocked in parallel, rather than one after another like i2s which would require twice the clock speed for data to arrive in time.
I2S frames are also 32bits long, for 16bit redbook data the other half is just padding, and this also doubles the clock speed. I would assume ECdesigns match the frame length to the incoming data to avoid padding.
To support up to 24bit 192kHz they must have a master clock inside running at a few MHz at least.

I dont really see how the processing circuits used to decode and reclock SPDIF into this data protocol could be much different to I2S aside from few times lower operating speeds.

ECD is a NOS DAC too, most other DACs are naturally operating at higher speeds because of upsampling.
Most DACs are delta sigma nowadays so it's a given, but take another popular multibit dac like the Soekris DAC2541.
It is has optical input, FIFO reclocking and a similar parallel interface for the ladder, it seems pretty similar to what ECD are doing here, upsampling was an intentional design choice since it objectively has a lot of advantages though it could technically operate without it like any other ladder dac.

[Qhwoeprktiyns]

@numlog My technical summary was probably far from being comprehensive. If you have any further questions I am sure ECD will be happy to answer them: https://www.ecdesigns.nl/en/contact-us

Maybe update us after? 

Thanks. 

[numlog]

I dont have any questions, I was just responding to your question about other DACs.
 

[Qhwoeprktiyns]

3 minutes ago, numlog said:

I dont have any questions, I was just responding to your question about other DACs.
 

 

What question ? EDIT: I see, the one about dispensing of I2S (and low bandwidth). I don't know how the Soekris works, but I assume they do and when they explain that they do things differently there must be something that escapes both of us ?

 

Here is also the link to his explanations on DIYAudio: https://www.diyaudio.com/forums/digital-line-level/79452-building-ultimate-nos-dac-using-tda1541a-post6462689.html

[numlog]

"However, do you know of any DACs that does this while dispensing of (high bandwidth) I2S communicaton altogether?"

Although I also sharing some possible doubts about what you were asking here.

From the post it does sound very similar to what the Soekris DACs do, maybe there are other examples but Soekris has roots in DIY community therefore we know a lot more about how they function.



 

[numlog]

Ah I finally see where the low bandwidth thing (the '200khz' DPI bandwidth figure from ECdesigns blog graph) is coming from, it's described in next page on that diyaudio thread:
https://www.diyaudio.com/forums/digital-line-level/79452-building-ultimate-nos-dac-using-tda1541a-778.html#post6495019

yes, a high number of optical data connections is a very different and interesting approach that would allow such low speeds! 
 

[Qhwoeprktiyns]

There's another guy claiming to have achieved similar results on DIYAudio: https://www.diyaudio.com/forums/digital-line-level/348074-synchronized-asynchronous-fifo-buffer-slaved-i2s-reclocker.html

All this escapes me, but this "low bandwidth" data communication seems to be unique to ECD. The DAC design (Fractal) is also unique, but I would be incapable of explaining any of it in detail.

 

I was trying to share my understanding of the changes they made from the current DAC, and why it could improve things, for people using currently the U192 (USB interface) and even for those of us who are using the UPL (low-noise source). We'll see...

[numlog]

It does seem to be unique, what caused the confusion is most of the old multibit DAC chips from the 90s had parallel data inputs, it is a normal thing for a ladder DAC, but the ECD has parallel bit inputs, a fact that was not clear to me from the blog, nor was the fact that the bit inputs are optical connections.
 

Im actually surprised the connections themselves are optical because, as ECD explained themselves, optical is jittery, that may be the trade off for ultra low noise/bandwidth.

 

 

[Qhwoeprktiyns]

22 hours ago, numlog said:

It does seem to be unique, what caused the confusion is most of the old multibit DAC chips from the 90s had parallel data inputs, it is a normal thing for a ladder DAC, but the ECD has parallel bit inputs, a fact that was not clear to me from the blog, nor was the fact that the bit inputs are optical connections.

 

I am not sure that is the case. I really think the best thing would be to confirm your understanding with ECD, and ask them if you can post the explanations here. DiyAudio forum could be a good place to do this as well. 

[numlog]

What is not the case? That is the explanation ECD gave on diyaudio

[Qhwoeprktiyns]

10 minutes ago, numlog said:

What is not the case? That is the explanation ECD gave on diyaudio

The only  optical connection is the Toslink input going into the microprocessor that handles the decoding / reclocking and ouputs parallel data with low bandwidth (send to the DAC). Just send them a mail to clarify... 

[numlog]

1 minute ago, hopkins said:

The only  optical connection is the Toslink input going into the microprocessor that handles the decoding / reclocking and ouputs parallel data with low bandwidth (send to the DAC). Just send them a mail to clarify... 

From the ECD diyaudio post I linked:
''The ideal digital audio interface foe connecting a source to an external DAC would be a DAPI interface based on 48 separate optical fibres (one for every bit) plus one fibre for sample timing (latch). This would reduce data bandwidth to 100 ... 200KHz, 20,000 times lower bandwidth compared to USB!''
 

[Qhwoeprktiyns]

Just now, numlog said:

From the ECD diyaudio post I linked:
''The ideal digital audio interface foe connecting a source to an external DAC would be a DAPI interface based on 48 separate optical fibres (one for every bit) plus one fibre for sample timing (latch). This would reduce data bandwidth to 100 ... 200KHz, 20,000 times lower bandwidth compared to USB!''
 

 

Yes I read that but it says "would be".

 

I can't help you to clarify these fine points. I really suggest you contact them - they have the answers, and I am sure they can adress the concerns you raised.

 

Otherwise, we are left with your doubts/questions/potential misunderstandings, and unfortunately they are not on this forum, as many other vendors are, to answer you directly. 

[numlog]

I am only responding to your own claims/potential misunderstandings about the products.

If you read the last line it sounds to me that this is the requirement for their claimed 200kHz bandwidth limiting.
Based on what they said, it is the same reason the USB DAPI wouldn't work, the 8MHz input bandwidth limiting is only possible with optical isolation of Toslink, the same applies for DAPI input to limit from 8MHz to 200kHz, isolation is required.

[Qhwoeprktiyns]

ECDesigns have completed the PowerDAC-R model, and were kind enough to send me a pre-production model to beta-test. You will find my "preview" here: https://audiophilestyle.com/blogs/blog/662-hopkins-blog/

 

It should be available within a few months, and will be priced between 1100 and 1200€ excluding VAT. Shipment outside the EU will be possible. 

 

I initially planned on posting my impressions shortly before it was offered for sale on their website. Had questions been asked until then, I would have felt uncomfortable not answering them or withholding the fact that I had the opportunity to test it. I hope people interested will appreciate reading about it now, even though there's a little longer to wait before you can purchase it :)

 

 

[Huubster]

Wow Stephane, what an honour that must be! And now I want that thing too of course, thanks for the extended report!

 

And I actually like the design! Retro style, B&O would have won prices with this kind of designs back in the days 😁

[realDHT]

Great news. Very interesting! 

And interesting news that ECD are designing a complementary tube amp too.  

 

So if I understand you correctly, you think the powerdac + rasperry is better than the UPL + fractal DAC in subjective listening tests? How much better would you say it is? Is it sort of a "must-do-upgrade-immediately" :) or just a little bit better in your opinion?  

 

[Qhwoeprktiyns]

28 minutes ago, realDHT said:

Great news. Very interesting! 

And interesting news that ECD are designing a complementary tube amp too.  

 

So if I understand you correctly, you think the powerdac + rasperry is better than the UPL + fractal DAC in subjective listening tests? How much better would you say it is? Is it sort of a "must-do-upgrade-immediately" :) or just a little bit better in your opinion?  

 

 

Yes, it is interesting that they are developing a tube amp. John Brown has researched the topic extensively, as we could expect, and I will be very curious to hear more about it. He always has interesting things to say - though they are sometimes hard for me to follow :)

 

Vis à vis the UPL+Fractal DAC, it is hard to qualify things precisely... I think it is significant - not sure you will be satisfied with that answer -  but it may vary based on how you use the FractalDAC today, as there are differences in the output impedance, output level, etc...  It certainly also depends on the rest of your system - and ECD will be the first to explain that as well (though I am not saying mine is perfect - far from it). So I am not too confident giving a "universal" answer to this. The fact that the PowerDAC - to my ears - delivers on source immunity, has its own transparent volume control, and is probably a better match with some amplifiers, is reason enough to do the upgrade. Believe me, I was reluctant to put aside the UPL, and am still a bit amazed by the sound I get with such a "basic" set up now.

 

Contrary to what I understood initially, there are quite a few improvements on the "DAC" side (between the current U96 and the PowerDAC). When you open up the PowerDAC, there are two boards: the DAPI board (Toslink receiver, micro-chip for decoding, etc...) and the DAC board. The DAC board of the PowerDAC looks a lot like the Fractal DAC board. But John Brown explained at length (on the telephone - I did not take notes !) differences which are certainly significant (otherwise he would not have bothered, I assume). Their intention is for the PowerDAC to be their "final" DAC. Those who have followed them will smile, but when you listen to it, that is the feeling you get as well.

[Qhwoeprktiyns]

33 minutes ago, Huubster said:

And I actually like the design! Retro style, B&O would have won prices with this kind of designs back in the days 😁

 

I like the unit much more "in person" than in pictures. It is really small. The picture I put on the blog entry, is how I have it now, next to my turntable, so you can appreciate the size. My girlfriend likes it, and likes the fact that there is overall less "spaghetti" ! So its a winner for everyone :)