Stacked Stereo DACs for Multichannel

[jjkale]

Gents,

I asked a question here about a week or so ago about creating a multichannel system from 3 stereo DACs.

I have narrowed the search to four possible solutions. Would appreciate any comments or insight any of you have.

Short review:

1. Currently employing Oppo 105D. I like it but I'm looking to upgrade the sound. More resolution.

2. I know the Exasound e38 is an option. My thought is that for the price, 1 x 9028 and a couple op-amps is not that great a value. No need to comment on Exasound, I know it is nice gear, I am just not inclined in that direction.

 

Possible Solutions:

1. Mytek: 3 x Brooklyn (or Manhattan/192DSD for fronts/surrounds) with WCLK and USB hub

2. PinkFaun: 3 x PF DACs with the PF I2S Bridge card. I am in contact with PinkFaun and they are all over this.

3. PSAudio: 3 xPS DACs with the PinkFaun card. PinkFaun says they can do it; I am waiting for a bit more confirmation from PSAudio.

4. Wait for the 205? version of the Oppo next year and mod it. It is my belief that a modded Oppo will exceed the quality of the e38. Again all respect.

 

The PinkFaun card options would require me to build a new computer server. Not a big deal, but a price consideration. I have reasonable skills and experience.

 

Performance, price and convenience will all be a part of my decision. The secondary market is also a player.

 

Thanks very much for your help, have a great weekend

[JohnSwenson]

Have you actually tried this?

 

How are you getting the data into the DACs? If USB you might have a problem, if using asynchronous mode each DAC has its own clock and they will be exactly the same, one will run slightly faster than another and over time this can cause the different channels to get out of sync with each other.

 

You also need software that is going to specifically start the stream to all the DACs at exactly the same time, most audio systems do not guarantee this.

 

You might ultimately be better off using a multichannel S/PDIF board that uses the same clock for all outputs, such a board would have a driver that would be designed to synchronize all the channels. I don't have a specific recommendation for such a board, but I know they exist.

 

Since most S/PDIF input DACs use the timing on the S/PDIF signal rather than an internal clock the channels should all stay in sync.

 

John S.

[jjkale]

The Mytek option does work. Others have used it with DSD192 dacs. That is what WCLK is all about.

The PinkFaun card does exactly what you suggest with I2S.

I am looking for actual experience and opinions on what the best option is.

But thanks.

[Distinctive]

You need also to bring the new PS Audio Transport and 3 DS DACS over I2s into the equation. If my understanding is correct there is built in a multichannel sync feature for the combo.

[Keith_W]

How are you getting the data into the DACs? If USB you might have a problem, if using asynchronous mode each DAC has its own clock and they will be exactly the same, one will run slightly faster than another and over time this can cause the different channels to get out of sync with each other.

 

I have a question for you, John. Most people cite this (inability to synchronize clocks) as a reason to choose a multichannel DAC over stacked stereo DAC's. My question is: what is the magnitude of the change in a worst case scenario? Assuming a DAC with really poor jitter performance, say 1 x 10^-10 seconds, it would take 10^10 seconds (16.7 million minutes) for the signals to drift 1 second out of sync. Over the time of a typical audio CD, it would drift by no more than a few fractions of a microsecond.

 

I want to know if my thinking is wrong. I suspect i may have misunderstood the nature of jitter. Could you please expand on how far out of sync each channel would get, if you were to listen to music for 2 hours non stop (without reloading any files, because I assume that if you reload a file, you will synchronize the channels again?).

 

You also need software that is going to specifically start the stream to all the DACs at exactly the same time, most audio systems do not guarantee this.

 

Do you know of any software like this?

Since most S/PDIF input DACs use the timing on the S/PDIF signal rather than an internal clock the channels should all stay in sync.

 

Some DAC's have both SPDIF inputs and USB inputs. Do these DAC's typically operate in one mode in SPDIF (accepting the incoming clock signal), and another in USB (reclocking the incoming signal)?

[ted_b]

Have you actually tried this?

 

How are you getting the data into the DACs? If USB you might have a problem, if using asynchronous mode each DAC has its own clock and they will be exactly the same, one will run slightly faster than another and over time this can cause the different channels to get out of sync with each other.

 

You also need software that is going to specifically start the stream to all the DACs at exactly the same time, most audio systems do not guarantee this.

 

You might ultimately be better off using a multichannel S/PDIF board that uses the same clock for all outputs, such a board would have a driver that would be designed to synchronize all the channels. I don't have a specific recommendation for such a board, but I know they exist.

 

Since most S/PDIF input DACs use the timing on the S/PDIF signal rather than an internal clock the channels should all stay in sync.

 

John S.

 

Yeah guys I wrote about me doing this almost 4 years ago when Michal asked me to try it (sent me two more Stereo192s in addition to mine) and we (Michal, Chebon, etc) then later demoed it at RMAF 2013.

 

The ASIO driver acts as the traffic cop for channel mapping, the serial numbers of the dacs put them in ascending rf/lf, c/lfe, rs/ls order, and a simple set of cables (wordclock out to world clock in, twice) and a USB hub was all we needed. Here is my thread:

http://www.computeraudiophile.com/f6-dac-digital-analog-conversion/first-multi-channel-direct-stream-digital-playback-solution-mytek-14946/

 

BTW, I first tried it without the wordclock synching and everything sounded fine, but was much better once the clocks cabled together. Probably former setup was slowly getting out of synch.

[bmoura]

Have you actually tried this?

Ted_B is the pioneer here as noted above.

 

Sounds like an interesting adventure to try. I'm more inclined to continue to use Multichannel DACs like the exaSound e28/e38 and the Merging Technologies NADAC MC-8/NADAC Player MC-8.

 

Then there is no worry about linking, clock sync, etc. Not to mention cost - 3 PS Audio DirectStream DACs or 3 Playback Designs DACs can easily top the cost of an e38 or a NADAC!

[ted_b]

The one advantage of stacking is that one can start off slowly into multichannel. Two stereo dacs can do 3 channel (RCA Living Stereo) or 4 channel (Audio Fidelity quad, Chesky 4.0, etc).

[bmoura]

The one advantage of stacking is that one can start off slowly into multichannel. Two stereo dacs can do 3 channel (RCA Living Stereo) or 4 channel (Audio Fidelity quad, Chesky 4.0, etc).

Depends on which Stereo DAC you are stacking.

 

Now that Mytek has cut the price of the Stereo192 DSD DAC to $1,095, there are some cost savings with 2 or 3 Mytek DACs vs. the exaSound e38. But once you stack the new Mytek Broooklyn DAC ($1,995) two of them are more expensive than the e38 ($3,849) - even before you get to three Myteks.

 

The cost of stacking two or three PS Audio or Playback Designs DACs takes you to even higher numbers. Cost saving is pretty much gone at that point.

 

http://nativedsd.com/database

[jjkale]

Good points.

Thanks for jumping in.

But clearly I'm not looking to save $. If that was the case...Oppo 105 or follow-on...decent enough DAC, great bluray player, fair price. Game over.

If the next-gen Oppo is produced with 4 or 5 9038 chips then that is a new game over. But I am guessing that will not be the case. If the Exasound had more under the hood, that would be a playa also.

I want to improve the quality of the MCH DAC by employing higher end stereo DACs.

Still looking for someone who has maybe used I2S, PinkFaun and PS Audio to derive MCH.

Thanks everybody.

[bmoura]

If the next-gen Oppo is produced with 4 or 5 9038 chips then that is a new game over. But I am guessing that will not be the case. If the Exasound had more under the hood, that would be a playa also.

Of course there is more to the sound quality of a DAC than which chips it uses. Assuming, of course, that the DAC has chips vs. a direct filtered or FPGA approach.

 

As one DAC maker famously said, it's what comes after the DAC chip(s) that really tells the story on how it will sound....

[jjkale]

Agreed. I am using the specific DAC chip (and quantity) as a proxy for quality.

I assume that in most cases, with the manufacturers we are talking about here, that the supporting electronics will be excellent.

[bmoura]

Agreed. I am using the specific DAC chip (and quantity) as a proxy for quality.

 

The problem is that the chips aren't a proxy for quality.

 

For example, at the recent Rocky Mountain Audio Fest, several new DACs were demonstrated with the new 9028 and 9038 DAC chips. But they weren't delivering the sound quality of some DACs that had been on the market for awhile with what some would call the earlier or "older" ESS DAC chips.

 

Definitely a case of needing to hear each DAC before judging the sound quality - vs. the chip number on the marketing sheet.

[Kal Rubinson]

Good points.

Thanks for jumping in.

But clearly I'm not looking to save $. If that was the case...Oppo 105 or follow-on...decent enough DAC, great bluray player, fair price. Game over.

If the next-gen Oppo is produced with 4 or 5 9038 chips then that is a new game over. But I am guessing that will not be the case. If the Exasound had more under the hood, that would be a playa also.

I want to improve the quality of the MCH DAC by employing higher end stereo DACs.

Still looking for someone who has maybe used I2S, PinkFaun and PS Audio to derive MCH.

Thanks everybody.

 

Playback Designs has a new USB-XIII box that will link up 3 Merlots for this purpose.

[jjkale]

Ok, I can live with that.

I hadn't read about that. Which equipment didn't live up to the hype?

[jjkale]

Playback Designs has a new USB-XIII box that will link up 3 Merlots for this purpose.

 

Thanks Kal, I didn't know about that equipment. Good call.

Have you by any chance written about either the DirectStream and/or the Manhattan/Brooklyn and/or an I2S/USB implementation?

Thanks.

[bmoura]

Playback Designs has a new USB-XIII box that will link up 3 Merlots for this purpose.

Interesting option. A limitation is that it doesn't support Multichannel DSD 256 files in 6 channel mode.

[mansr]

I have a question for you, John. Most people cite this (inability to synchronize clocks) as a reason to choose a multichannel DAC over stacked stereo DAC's. My question is: what is the magnitude of the change in a worst case scenario? Assuming a DAC with really poor jitter performance, say 1 x 10^-10 seconds, it would take 10^10 seconds (16.7 million minutes) for the signals to drift 1 second out of sync. Over the time of a typical audio CD, it would drift by no more than a few fractions of a microsecond.

 

I want to know if my thinking is wrong. I suspect i may have misunderstood the nature of jitter. Could you please expand on how far out of sync each channel would get, if you were to listen to music for 2 hours non stop (without reloading any files, because I assume that if you reload a file, you will synchronize the channels again?).

 

Jitter isn't the problem here, long-term accuracy is. Looking at the Mytek Brooklyn, it uses an Abracon ABLNO-V 100 MHz crystal oscillator with a specified frequency stability of 18 ppm. Over an hour, that amounts to 65 milliseconds. Worst case, two DACs using this oscillator would exhibit a drift of 130 ms per hour. Listening to that would be intolerable.

 

I realise the Mytek DAC allows an external world clock which avoids such drift problems. I used it as an example since they brag about the quality of its clock and I was able to identify the exact oscillator it uses.

 

Some DAC's have both SPDIF inputs and USB inputs. Do these DAC's typically operate in one mode in SPDIF (accepting the incoming clock signal), and another in USB (reclocking the incoming signal)?

 

That would be my expectation.

[ted_b]

Depends on which Stereo DAC you are stacking.

 

Now that Mytek has cut the price of the Stereo192 DSD DAC to $1,095, there are some cost savings with 2 or 3 Mytek DACs vs. the exaSound e38. But once you stack the new Mytek Broooklyn DAC ($1,995) two of them are more expensive than the e38 ($3,849) - even before you get to three Myteks.

 

The cost of stacking two or three PS Audio or Playback Designs DACs takes you to even higher numbers. Cost saving is pretty much gone at that point.

 

http://nativedsd.com/database

 

I'm not the one saying cost is clearly an advantage with stacking; I'm simply saying you can start with two (and yes, Mytek 192StereoDSDs are a nice cost option).

[jjkale]

Playback Designs has a new USB-XIII box that will link up 3 Merlots for this purpose.

 

Kal,

This device appears to be pretty much I have been looking for.

Do you have any reason to suspect that the PbD USB-XIII would not work as exactly as advertised? I.e., any DAC with coax or AES?

Thank you sir.

[bmoura]

Kal,

This device appears to be pretty much I have been looking for.

Do you have any reason to suspect that the PbD USB-XIII would not work as exactly as advertised? I.e., any DAC with coax or AES?

Thank you sir.

The USB-XIII uses the proprietary PL-Link system. So it will work with any DAC from Playback Designs - but not from other audio manufacturers.

[jjkale]

The USB-XIII uses the proprietary PL-Link system. So it will work with any DAC from Playback Designs - but not from other audio manufacturers.

 

Bmor,

Thanks.

If you read the Playback Designs description of the product, it specifically says it will work with non-branded PbD DACs with either coax or AES.

 

"Its capabilities have also been expanded to work with any other DAC equipped with either a AES or Coax input."

 

"Any non-Playback Designs branded DAC with AES or Coax input can be

connected. For this the USB input still supports PCM up to 384kHz and quad

DSD, but any PCM format with sample rates greater than 192kHz is converted to

either 176.4kHz or 192kHz and any DSD format with sample rates greater than

1x is converted to 1x via DoP."

 

Check it.

[cmarin]

Bmor,

Thanks.

If you read the Playback Designs description of the product, it specifically says it will work with non-branded PbD DACs with either coax or AES.

 

"Its capabilities have also been expanded to work with any other DAC equipped with either a AES or Coax input."

 

"Any non-Playback Designs branded DAC with AES or Coax input can be

connected. For this the USB input still supports PCM up to 384kHz and quad

DSD, but any PCM format with sample rates greater than 192kHz is converted to

either 176.4kHz or 192kHz and any DSD format with sample rates greater than

1x is converted to 1x via DoP."

 

Check it.

 

I have the USB-XIII and can confirm that it works wonderfully with both my Playback Designs MPS-5 through the Plink interface and my Pacific Microsonics DAC via AES.

 

 

Sent from my iPhone using Computer Audiophile

[jjkale]

Looking closer at the USB-XIII manual, it appears that the device may support any other branded DAC playback on two channel only, not MCH. MCH may only be available through the PLINK system. The website description and manual are not very clearly written. I sent PbD an email asking for clarification.

[jjkale]

I have the USB-XIII and can confirm that it works wonderfully with both my Playback Designs MPS-5 through the Plink interface and my Pacific Microsonics DAC via AES.

 

 

Sent from my iPhone using Computer Audiophile

 

But two channel only correct? I am interested in MCH.

thanks