Folks,
I'm looking to build a multichannel (6 or 8 channel) DAC with differential balanced outputs...
btw: when playing movies using a HTPC and a USB - DAC - is it easy to keep the video and audio in synch or is that something that needs manual delay correction every time?
If its a pain to keep the audio in-synch with the video on a HTPC - then I'm probably better-off going for an A/V processor with HDMI - such as Sherbourn PT-7030 or Onkyo PR-SC5509 - in which case the HDMI-lip-synch-lock is automatic.
As for the DAC - I've already looked at products from Lynx, RME, Focusrite, Digigram, Motu, M-Audio, Roland, Behringer, Wadia, Matrix, Mytek, Resonessence etc and the following options seem to come closest:
The exasound e28 is well beyond my price range... but is perfect and has most of what I want (and some features I don't need) ..
The Twistedpear - Buffalo is something that will need work putting it together and could run more than $600 total
The AKM doesn't look bad - but the performance figures aren't as good as the e28 or Buffalo.
Due to lack of any other viable options... I think I may have to embark on a DIY for the Ultimate multi-channel DAC that would rival the best 2 channel DACs from Antelope and other premium stereo DACs.
Ideally - I'd like to have someone (or a few people) experienced join me in this venture and build this DAC and if sufficient interest is there possibly market and may be even sell this under a direct-internet sales model (like exasound)....
Following are the broad requirement specs:
1. Asynchronous USB input with buffer (no need for SPDIF etc) - optional Ethernet of HDBaseT input in the future.
2. Automatic PCM / DSD Switching (native DSD support)
3. Remote control for power, volume and mode selection
4. Differential DAC output to be maintained through-out analogcircuitry to get 8 channel line level true and fully balanced outputs to gold-plated 3 pin XLR jacks
5. Noise, Crosstalk and Jitter Reduction
6. USB ground isolated galvanically from the DAC and the analogue circuits
7. Digital subsystem powered independently and isolated galvanically from the rest of the board.
8. Like the e28 - does it make sense to have separate Quartz oscillators for (44.1, 88.2, 176.4, and 352.8 kHz) and (48, 96,192 and 384kHz) and another Quartz oscillator as reference master clock with 0.13ps precision for D to A conversion?
9. Need not have internal power transformer - may be we can feed DC power using external laptop-adapter style transformers...
10. Power Cleaning Stages -
11. It would be nice if its a modular design with a single USB digital input and 2ch output and the ability to upgrade by adding expansion 2ch DACs - so as to make 4ch, 6ch, 8ch,10ch or 12ch balanced analog outputs.
Some other considerations:
DAC chip: ESS Technology ES9018 Sabre Reference 32bit (even though its a 8 channel chip - I'd like to use it 4 of these as a stereo DAC for better SNR and Dynamics)
other option is a Texas Instruments - Burr Brown - PCM1794A or DSD1794A or PCM1690?
Any other DAC ships to consider that are 32bit/384kHz or at a minimum 24bit/192kHz or better?
Other considerations:
the best quality components (Nichicon etc) through-out
gold plated XLR or miniXLR outputs.
Is there anyone out there who can join me in this DIY venture to build the ultimate multichannel DAC?
How much would such a DAC cost?
What are all the components and tools I would need to complete the above project?
Can this be done for well below the price range of the exasound e28 and closer to around what the Buffalo DAC would cost?
I'm looking to build a multichannel (6 or 8 channel) DAC with differential balanced outputs...
btw: when playing movies using a HTPC and a USB - DAC - is it easy to keep the video and audio in synch or is that something that needs manual delay correction every time?
If its a pain to keep the audio in-synch with the video on a HTPC - then I'm probably better-off going for an A/V processor with HDMI - such as Sherbourn PT-7030 or Onkyo PR-SC5509 - in which case the HDMI-lip-synch-lock is automatic.
As for the DAC - I've already looked at products from Lynx, RME, Focusrite, Digigram, Motu, M-Audio, Roland, Behringer, Wadia, Matrix, Mytek, Resonessence etc and the following options seem to come closest:
The exasound e28 is well beyond my price range... but is perfect and has most of what I want (and some features I don't need) ..
The Twistedpear - Buffalo is something that will need work putting it together and could run more than $600 total
The AKM doesn't look bad - but the performance figures aren't as good as the e28 or Buffalo.
Due to lack of any other viable options... I think I may have to embark on a DIY for the Ultimate multi-channel DAC that would rival the best 2 channel DACs from Antelope and other premium stereo DACs.
Ideally - I'd like to have someone (or a few people) experienced join me in this venture and build this DAC and if sufficient interest is there possibly market and may be even sell this under a direct-internet sales model (like exasound)....
Following are the broad requirement specs:
1. Asynchronous USB input with buffer (no need for SPDIF etc) - optional Ethernet of HDBaseT input in the future.
2. Automatic PCM / DSD Switching (native DSD support)
3. Remote control for power, volume and mode selection
4. Differential DAC output to be maintained through-out analogcircuitry to get 8 channel line level true and fully balanced outputs to gold-plated 3 pin XLR jacks
5. Noise, Crosstalk and Jitter Reduction
6. USB ground isolated galvanically from the DAC and the analogue circuits
7. Digital subsystem powered independently and isolated galvanically from the rest of the board.
8. Like the e28 - does it make sense to have separate Quartz oscillators for (44.1, 88.2, 176.4, and 352.8 kHz) and (48, 96,192 and 384kHz) and another Quartz oscillator as reference master clock with 0.13ps precision for D to A conversion?
9. Need not have internal power transformer - may be we can feed DC power using external laptop-adapter style transformers...
10. Power Cleaning Stages -
11. It would be nice if its a modular design with a single USB digital input and 2ch output and the ability to upgrade by adding expansion 2ch DACs - so as to make 4ch, 6ch, 8ch,10ch or 12ch balanced analog outputs.
Some other considerations:
DAC chip: ESS Technology ES9018 Sabre Reference 32bit (even though its a 8 channel chip - I'd like to use it 4 of these as a stereo DAC for better SNR and Dynamics)
other option is a Texas Instruments - Burr Brown - PCM1794A or DSD1794A or PCM1690?
Any other DAC ships to consider that are 32bit/384kHz or at a minimum 24bit/192kHz or better?
Other considerations:
the best quality components (Nichicon etc) through-out
gold plated XLR or miniXLR outputs.
Is there anyone out there who can join me in this DIY venture to build the ultimate multichannel DAC?
How much would such a DAC cost?
What are all the components and tools I would need to complete the above project?
Can this be done for well below the price range of the exasound e28 and closer to around what the Buffalo DAC would cost?
Last edited:
I'd have to say both - because in theory both should be the same - but we all know "sounds ulti" is subjective... whereas "measures ulti" is objective (to an extent, subject to measuring equipment and environment)
therefore what trade-offs/differences/compromises are we talking about when going for either?
therefore what trade-offs/differences/compromises are we talking about when going for either?
A good question. I'd say that for subjective excellence its necessary to pay attention to noise, and noise floor modulation in particular. For objective excellence the attention turns to distortion (THD+N). Subjectively, low-level performance (-30dBfs and below) needs to be optimized because music signals aren't pure sinewaves - rather superpositions of hundreds of small sinewaves. Objectively, the attention goes to performance at full output with a single tone as stimulus.
If you look at how the noise varies for your suggested ES9018 DAC, you can measure something like this. A test tone of around 1kHz is applied, the two colours differ by just 1dB in level but check out the huge shift in the noisefloor. The levels of the two tones are -36dBfs and -35dBfs.
If you look at how the noise varies for your suggested ES9018 DAC, you can measure something like this. A test tone of around 1kHz is applied, the two colours differ by just 1dB in level but check out the huge shift in the noisefloor. The levels of the two tones are -36dBfs and -35dBfs.
Attachments
Last edited:
Thanks for that explanation... that makes sense...
at the end of the day - fidelity is the goal - as in reproducing the studio-recording transparently without extraneous noises or coloring.
i suggested the ES9018 - not out of loyalty or anything but more based on the fact that it seems to have the highest theoretical specs and has been implemented in various high-end brands.
Now to the details:
1. what about the synchronization with video? if this is an issue... an external multichannel Asynchronous USB or Ethernet DAC will only be useful for surround audio and practically meaningless for movies...
2. Can the suggested modular design of a 2Ch DAC with USB input but with ability to add additional 2 DAC baords be achieved? If its tough or too costly this requirement can be dropped.
3. What are the other requirement and design considerations?
4. What would be the estimated parts/material cost (excluding design cost and labor) etc?
at the end of the day - fidelity is the goal - as in reproducing the studio-recording transparently without extraneous noises or coloring.
i suggested the ES9018 - not out of loyalty or anything but more based on the fact that it seems to have the highest theoretical specs and has been implemented in various high-end brands.
Now to the details:
1. what about the synchronization with video? if this is an issue... an external multichannel Asynchronous USB or Ethernet DAC will only be useful for surround audio and practically meaningless for movies...
2. Can the suggested modular design of a 2Ch DAC with USB input but with ability to add additional 2 DAC baords be achieved? If its tough or too costly this requirement can be dropped.
3. What are the other requirement and design considerations?
4. What would be the estimated parts/material cost (excluding design cost and labor) etc?
Last edited:
Reproducing the studio recording transparently suggests to me you prefer the 'subjective' approach to the objective one. Am I correct?
If I am then the next stage is to ask whether the recording itself has been produced 'transparently' according to the same priorities. And I rather suspect you'll find the answer in many cases is 'no, its produced with equipment designed for the best numbers'. Hence the preponderance of S-D based ADCs in service nowadays with all their noise modulation artifacts which don't show up clearly in the traditional measurements.
If I am then the next stage is to ask whether the recording itself has been produced 'transparently' according to the same priorities. And I rather suspect you'll find the answer in many cases is 'no, its produced with equipment designed for the best numbers'. Hence the preponderance of S-D based ADCs in service nowadays with all their noise modulation artifacts which don't show up clearly in the traditional measurements.
I somehow knew you were going to talk about the recording equipment itself... lets set that aside for the moment... I'm only concerned about the transparent playback of the lossless audio files that I have (whether it was compressed correctly and without atrificats all that is a different topic) and whether my HTPC/Software (JRiver) etc decoded or uncompressed correctly can also be set aside as outside the scope of our unit...
All we need to focus on is given a set of lossless uncompressed PCM or DSD - our DAC has to stay true to the bits... and give the proper analog rendition of the same...
Rather than get bogged down in discussions over subjective/objective design considerations and parameters - I say let us take our reference product - the exasound e28 or the Buffalo III DAC kit - these are proven and popular and have established a good name ...
Can we follow the same design approach, same parameters/compromises and try to do as good a job or better?
but all this depends on answers on the video-audio-synch issue...
If audio-synch with movies is going to be a routine issue with an external Asynchronous DAC then it would be a true deal breaker for a multichannel DAC whose primary use will be for home-theater movies and not surround audio...
and may be that explains why there aren't that many standalone multichannel DACs marketed for home-theaters - they need HDMI and that means we are talking about more than just a DAC - we are getting closer to the A/V processor territory... in-fact I'm ok with that... we can even consider adding a HDMI input some DSP/decoder chips and a full A/V pass-through HDMI output...
All we need to focus on is given a set of lossless uncompressed PCM or DSD - our DAC has to stay true to the bits... and give the proper analog rendition of the same...
Rather than get bogged down in discussions over subjective/objective design considerations and parameters - I say let us take our reference product - the exasound e28 or the Buffalo III DAC kit - these are proven and popular and have established a good name ...
Can we follow the same design approach, same parameters/compromises and try to do as good a job or better?
but all this depends on answers on the video-audio-synch issue...
If audio-synch with movies is going to be a routine issue with an external Asynchronous DAC then it would be a true deal breaker for a multichannel DAC whose primary use will be for home-theater movies and not surround audio...
and may be that explains why there aren't that many standalone multichannel DACs marketed for home-theaters - they need HDMI and that means we are talking about more than just a DAC - we are getting closer to the A/V processor territory... in-fact I'm ok with that... we can even consider adding a HDMI input some DSP/decoder chips and a full A/V pass-through HDMI output...
Last edited:
I understand your goal....
and am with you .... but I was getting lost in the "design considerations" debate... ok assuming we are focussing on the ultimate DAC - in which case it has to pass both the Subjective and Objective - Sound quality tests... because the objective is what can be put in to the brochures and the subjective will be left up to the various forum and magazine reviewers and "audio-philes"
and of-course HDMI is not an option - as it would involve exorbitant fees... may be we can consider that as an add-on later if the product does well.
and am with you .... but I was getting lost in the "design considerations" debate... ok assuming we are focussing on the ultimate DAC - in which case it has to pass both the Subjective and Objective - Sound quality tests... because the objective is what can be put in to the brochures and the subjective will be left up to the various forum and magazine reviewers and "audio-philes"
and of-course HDMI is not an option - as it would involve exorbitant fees... may be we can consider that as an add-on later if the product does well.
Last edited:
So you'd like this DAC to be a 'crowd sourced' marketed item - from talking about 'brochures' ? If so I doubt you could really sell ultimate SQ in that manner - it would have to be by word of mouth amongst the users. As Krishnamurti said 'Truth cannot be brought down, it has to be ascended to'. 'Ultimate' is not a sales pitch, really.
well i meant brochures as in any description for a product... as in the "specs" that can be put up on a website.
ok we are getting bogged down in semantics and subjective vs objective...
I have nothing against making the best sounding DAC while it also looks good on paper...
that is all...
can we get in to the technical details and my other serious questions - especially AV synch?
ok we are getting bogged down in semantics and subjective vs objective...
I have nothing against making the best sounding DAC while it also looks good on paper...
that is all...
can we get in to the technical details and my other serious questions - especially AV synch?
OK if you really did mean 'ultimate' as in the best possible SQ then its an interesting project for sure. Interesting because its massively challenging. It won't be over in one or two years, it'll be seriously major. Are you sure you have the patience? 
So what are your given set of design considerations? I see you diving into the much less important (from my pov) details and omitting to notice the huge challenges. If you set a budget for the cost of the unit at the start then it might not actually be possible. One important consideration is whether we have to use current production parts or could build from 'obsolete' parts and hence only produce a restricted number.
So what are your given set of design considerations? I see you diving into the much less important (from my pov) details and omitting to notice the huge challenges. If you set a budget for the cost of the unit at the start then it might not actually be possible. One important consideration is whether we have to use current production parts or could build from 'obsolete' parts and hence only produce a restricted number.
i had always meant for it to be the best SQ, best fidelity, best specs on paper and on ear - no doubts, no compromise - otherwise I could just have gone ahead with some existing DAC or AV processor...rather than try to make my own...
time is a factor - as I think we need to take an iterative approach - come out with the best we can as soon as possible and continue the improvements... perfection is a never-ending quest... as long as we keep getting better and doing the best with each iteration...
cost is also factor - to the extent that we can build the prototype(s) with our money and without balking or having to borrow or seek investors... on the other hand - we can also approach crowd-funding sources like kickstarter for the project... I believe there is a successful DAC that was built on the kickstarter model (http://www.indiegogo.com/projects/geek-pulse-a-digital-audio-awesomifier-for-your-desktop).
for parts: i'd say we go with current production - something that we can readily get - should there be a need to build in volume... is there a cost or performance advantage to obsolete parts?
time is a factor - as I think we need to take an iterative approach - come out with the best we can as soon as possible and continue the improvements... perfection is a never-ending quest... as long as we keep getting better and doing the best with each iteration...
cost is also factor - to the extent that we can build the prototype(s) with our money and without balking or having to borrow or seek investors... on the other hand - we can also approach crowd-funding sources like kickstarter for the project... I believe there is a successful DAC that was built on the kickstarter model (http://www.indiegogo.com/projects/geek-pulse-a-digital-audio-awesomifier-for-your-desktop).
for parts: i'd say we go with current production - something that we can readily get - should there be a need to build in volume... is there a cost or performance advantage to obsolete parts?
Last edited:
There's the Light Harmonic Geek DAC which got hugely over-subscribed yes. But I doubt that approach could work here as the timescale isn't known at the outset.
The question about obsolete vs current production was related to which DAC chip could be used. Clearly ultimate SQ isn't going to be obtained with ES9018 so one of the first decisions to be made is whether a commercial chip can be used or whether something has to be engineered from first principles. The first route would be considerably quicker but the choice is wider if an out of production part could be used, rather than limiting ourselves to current production.
The question about obsolete vs current production was related to which DAC chip could be used. Clearly ultimate SQ isn't going to be obtained with ES9018 so one of the first decisions to be made is whether a commercial chip can be used or whether something has to be engineered from first principles. The first route would be considerably quicker but the choice is wider if an out of production part could be used, rather than limiting ourselves to current production.
Picture is too fuzzy to make out scales. What can been seen in white trace is terrible IMD. Could be test setup, could be software, could be poor electronics. Likely aliasing problem combined with power supply issues.
Here's an example of well behaved test set up, with well behaved soundcard. A full scale 1kHz sine is generated and used as test signal. Signal is digitally attenuated to -35dB for 1st recording, and -36dB for second recording:
For comparison, 0dBFS signal is recorded, and plotted with -36dB offset with overlay of result obtained recording at -36dB:
From above it is clear that DAC, and ADC for capture have excellent linearity.
If OP goal is building high performance HPTC system, then building eight channel DAC is likely waste of effort.
Noise floor modulation all stems from HD of system, and potentially from system timing issues.
All signals fed to system are continuous. Any length of continuous signal may be treated as periodic, and is just a collection of sine waves.
Folks, I shouldve posted the disclaimer first... my knowledge of electronics is at beginner level... therefore someone else will be a techie on this project. I do have someone who is an intermediate level electronics person who can assist with the soldering and such...
@Barleywater.... the proposed HTPC will be a fanless cabinet with SSD and no HDD. The content will be fed from a NAS or external USB - HDD. It could even be a Raspberry Pi like device...
@abraxalito... I'm not tied to any particular chip... the ES9018 seems well known and drawing a lot of attention... that's why I started out with that... but if it makes sense to go for something else... then no problem.... ideally it should sound and measure as good as the best 2channel DACs out there...
I do think we should select a chip currently in production...(be it from TI, AD, ESS, Wolfson or others) so that we are not left in the lurch, having to redesign for an entirely new chip... should there be a wildly unexpected demand for our product.
and in such a scenario - we could go for a custom DAC chip in our second iteration - if we are lucky enough to get to that stage we will have a lot more time, confidence and resources at our disposal...
I had stated my design parameters - albeit without much foresight or insight...
could you state your design parameters for the ultimate SQ DAC?
btw: another question is - I've heard of magnetic pre-amps - is that an option for the analog stages of the DAC?
@Barleywater.... the proposed HTPC will be a fanless cabinet with SSD and no HDD. The content will be fed from a NAS or external USB - HDD. It could even be a Raspberry Pi like device...
@abraxalito... I'm not tied to any particular chip... the ES9018 seems well known and drawing a lot of attention... that's why I started out with that... but if it makes sense to go for something else... then no problem.... ideally it should sound and measure as good as the best 2channel DACs out there...
I do think we should select a chip currently in production...(be it from TI, AD, ESS, Wolfson or others) so that we are not left in the lurch, having to redesign for an entirely new chip... should there be a wildly unexpected demand for our product.
and in such a scenario - we could go for a custom DAC chip in our second iteration - if we are lucky enough to get to that stage we will have a lot more time, confidence and resources at our disposal...
I had stated my design parameters - albeit without much foresight or insight...
could you state your design parameters for the ultimate SQ DAC?
btw: another question is - I've heard of magnetic pre-amps - is that an option for the analog stages of the DAC?
Last edited:
Well the design parameters come from what's necessary to give a subjectively convincing account of the recording. Bob Stuart (of Meridian) did some work on this, published an AES paper or two - the upshot was something around 64kHz and 20bits was sufficient. Therefore since 64kHz isn't a standard sample rate the technical spec of an ultimate DAC needs only to be 88k2 (also perhaps 96kHz for compatibility) and 120dB SNR. It makes no sense to aim for more bits or higher sampling - input sample rates above this can be catered for by downsampling.
Nope, never heard of a magnetic pre-amp - where did you learn of them?
Nope, never heard of a magnetic pre-amp - where did you learn of them?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.