Article: Reality Quest: Power and Digital Audio with the Sean Jacobs DC4 and Sound Application PGI TT-7 (Part 1 of 2)

[ray-dude]

View full article

[ray-dude]

100% agreed Rajiv.  There are very practical constraints (BOM, room in case, etc) with any piece of equipment, which I appreciate and respect.  That being said, modularity (especially for modular components) enables customers to calibrate their investment to their own objectives/constraints. 

[ray-dude]

6 hours ago, ted_b said:

Great storytelling Ray, and great overall writing.  This part one was a quite a teaser, too.  BTW, what does Rob say (I love Rob too) when you explain to him the differences you hear with Sean's power supplies?  Also, what/where will the Hynes fit in this architecture?  Replace the DC4?

 

Ted, Rob is pretty clear that he thinks other power supplies are not only not helpful, but detrimental.  That being said, him voicing that opinion was (generally) a couple years ago, and there has been huge progress in upsampling (mScaler), digital servers, power, etc. since then.  Alas, with Covid restrictions, it is very difficult to be gathering in the same space to hear the same things.  Hopefully soon (and again, there is literally no one in digital audio I respect more than Rob Watts).

 

I am continuing work to experiment with different power sources, build on the findings with the DC4.  A DR SR7 for the three rails may be a winner over the DC4, it may not (but it would be a replacement).  With Paul building custom SR7's again, hopefully there are some in the wild soon to test this (alas, my build is well down the queue)

 

[ray-dude]

4 hours ago, MarkusBarkus said:

Although I had studied the pics before the conversion, it was still sobering to see the little CNC-ed chamber and aluminum billet walls created to isolate the stock PS. 
 

I suspect just pulling the stock SMPS out of its coffin and putting it in an external box would improve sound, for those without a DC4 budget. I think that would be do-able.

 

And great call-out on the connector. I used blue tape to index the orientation in case I lost my mind and my nerve reconnecting it!

 

Great article, Ray. Mark

 

Mark, FWIW, the DAVE SMPS vs DC3 test I wrote up in listening tests had the stock SMPS outside the DAVE case in a cardboard box ;)

[ray-dude]

Measurements would be very helpful Archimago.  I'm seeing different trends with 5V line vs ±15V lines, but I would love to map that to measurements to see what the correlations are.  Right now, I'm having to indirectly ascertain stuff, but I do feel like a naturalist in an EE world (as an EE, not an easy thing to say).  Subjectively, the impact is very clear, but objectively, why?  I'd love to know to more directly optimize the subjective outcomes.

[ray-dude]

1 minute ago, The Computer Audiophile said:

I wish all people who want measurements would say it this way. Bravo @Archimago . 

 

I wish all of us that can't (currently) do measurements lived close enough to @Archimago so we could take our toys over and cross map measurements to subjective experience! (over an appropriate beverage (or three) of course)

 

 

[ray-dude]

52 minutes ago, PeterG said:

Great piece!  Let's hope this evolves into a plug and play solution for those faint of heart

 

Way back in the day, I got one the very first DC3's for DAVE. It was a very scary experience to crack the case and figure out stuff (and still is). I do know several people who've sent in their DAVE's to have the umbilical installed and their DAC tested with their supply, so there are options to make things more plug and play.

 

That being said, the more DAC manufacturers move to separate modular power schemes, the easier it will be to have true plug and play. Separating out the power supply from the core DAC is inconvenient from a manufacturing and sales experience, but for folks that are buying/selling modular components (particularly at the high end) I think it very necessary and potentially very impactful on SQ.

 

Modular isn't ideal for everyone though. There will always be a market for all in one type systems (streamer+DAC etc) where the convenience of one box will win out.

[ray-dude]

1 hour ago, jeffhenning said:

John Siau, who's the chief designer & CTO at Benchmark Media, said it best when, in an interview, stated that if you need a power conditioner for your system to sound its best, your equipment has really terrible power supplies in them.

 

This may be a better conversation after part 2 is posted (re the Sound Application PGI TT-7 power line conditioners)  It had a pretty stunning impact for me across all my components (and my TV).  Diving into the details of the TT-7, it would be a very steep ask for component level power supplies to do similar filtering (although of course it is in the ear of the beholder whether than level of filtering has value)

[ray-dude]

4 hours ago, Middy said:

Never imagined i'd ever have a Dave and more than content to have a Qutest but i finally get to see a Pimp my ride version of 'Dave'.

 

 

I don't have a Qutest, but I'm 99.9999% certain that the Qutest would be significantly elevated with better power.  Even with my Hugo TT2, better energizer power to the super caps has a surprisingly significant upside.  For something like the Qutest (or bypassing the battery on a Hugo2) I think it is definitely worth exploring if power makes an impact for you.

 

Re the objective/subjective topic that @Archimagoraised, unlike a lot in audio, quality and attributes of power sources are VERY measurable.  In the case of these experiments, I suspect output impedance and transient current capacity are more impactful than any noise on the output, but the experiment with and without the CX noise filter on the output regulator of the DC4 (and what I highlight in part 2 when it gets posted), noise does have a material impact on sound stage and imaging.  I'd love to quantitatively characterize these power sources under normal operating conditions, then use subjective impressions to correlate to objective power supply measurements to see what attributes and performance measures move different SQ needles.

 

 

[ray-dude]

2 hours ago, Middy said:

I have tried and been lucky with so many great sources Like John S et al and the members here.

My Qutest is on a UPS sinewave 230v out into a balanced isolation transformer. Into an older LPS1 7V into a dual reg Ldovr 5v and IFI DC IPURIFIER..

The delta wasnt as much as we always hope but that extra refinement does stand out.

Blood from a stone at the bottom of the Chord ladder..

Cost conscious small spends.

Why its so nice reading about someones end game efforts turning the great into something above and beyond.  I mentioned this superb article on the Chord owners club on FB. 

A Nick Bacon said he had the  mod done and was one of the best things he has done in his system.

 

Thanks again Ray keep up the good work

 

Dave

 

 

 

 

The Qutest (esp. when fed with good power and upsampled content from things like HQPlayer) is some of the best bang for the buck in audio!  Very glad that @Fourlegshas been promoting this gem!

[ray-dude]

It most definitely is a Quest Bob, with these write ups being more of a travelogue ;)  

 

I do think the state of the art has evolved a LOT since the DAVE was introduced back in 2015(!!). If/when a mythical DAVE2 gets released, things may be quite different.

 

I am delighted to see newer DACs taking power more seriously, supporting 32fs inputs, etc.  All these make me daydream even more about what a DAVE2 could actually do.

[ray-dude]

The DC4 umbilical I have is ~3' long.  @Nenon and I have been chatting about giving a Mundorf silver/gold cable a try, but with holiday schedules, we haven't gotten around to that yet.  Given what Nenon and others have heard in other applications, I am enthusiastic about what an upgraded cable will bring to the party.

 

If you're willing to place your DAVE on top of the DC4, the umbilical length can be quite short (<1').  For vibration control reasons, I'd prefer to keep each on its own stack, so the minimum length for me would be at least 18" (assuming side by side like I have in the photos). If you have a robust rack, then options clearly open up for a vertical layout.

 

There are certainly many paths to power nirvana, and it continues to be interesting to compare the piece parts of DC3 and DC4. The 5V line is primarily digital, and the ±15V lines are primarily for the analog stage. Does it make sense to optimize these for different dynamic loads and noise sensitivities? Different wiring to reflect these loads/sensitivies? Different classes of filters? Being able to separate out the 5V line from the ±15V line is a boon for experimentation, and I continue to learn a heck of a lot about how different power aspects impact different aspects of sound (hence my longing to bring some object measurements to the party...with full measurements the optimization/tuning process would be a heck of a lot less empirical).

 

Certainly compared to my DC3, the DC4 has advanced the state of the art considerably (to say the least). I suspect the next frontier is individual tuning of the digital and analog lines for a bit extra synergy with the particulars of the DAC, but we're definitely at the point of diminishing returns.  Of course, when we do fully optimize power for the DAVE, the audio gods will conspire to have that be the same week Rob announces his next TOTL DAC, so the cycle can begin all over again....

 

 

[ray-dude]

Believe it or not, DAVE has a display mode where the display turns on when you change something, then turns off.  Rob did call out the SQ benefits of having the display off while playing music ;)

 

 

[ray-dude]

Remember that you have -15V as well, so you would need three supplies.  These supplies have enough current for the job so they would work.  How they would compare to other options is TBD (I've tied multiple supplies on the 5V rail, but I did not have a Farad to test...they all have a different characteristic sound)

[ray-dude]

This is true.  DAVE 5V draws 0.7-1A (or so).  LPS 1.2 really shines at lower current draw.  That being said, LPS 1.2 5V >> DAVE SMPS 5V (at least in my living room).  A higher current capable device (ldovr, JS2, etc) could be higher performing still.  For example, I thought the ldovr 3A 5V supply to perform better than the LPS 1.2

[ray-dude]

Congratulations @happybob!  Within a week or two, you'll be hearing what the DC4 can really do (although it will go through a roller coaster getting there).

 

Very much looking forward to hearing your impressions with headphones direct.  I neglected to do those listening tests when the DC4 was here, so it will be interesting to if/how the DC4 helps with the headphone output for demanding cans like the Abyss's

 

[ray-dude]

In general, Chord DACs do better with PCM content.  The more upsampling you do (mScaler or HQP sinc filters), the better PCM performs.

 

That being said, each release is mastered through a different pipeline, and any digital processing will impact what you're hearing through DAVE.  As a rule of thumb, I try to purchase music as close to the original mastering format as possible (98 times out of 100, that is PCM).  

 

Alas, with SACDs, some SACDs have different masterings on the SACD DSD layer than the redbook layer.  It makes it that much harder to compare DSD64 vs redbook PCM.  The quality of DSD to PCM conversion REALLY varies based on the modulators and algorithms used, so very very hard to do a true apples to apples comparison.

[ray-dude]

I'm 99.9% 2 channel listening these days (at home, no office), but I have a Hugo TT2 for the Abyss headphones (DAVE is not quite manly enough for those cans)

 

In the upsampling experiments I've been doing, there is a definite and tangible benefit to reconstructing signal from redbook sources (Chord DACs do this natively, but one can go well beyond what Chord DACs and even mScaler can do with SW upsampling).  The main benefit is in the time domain.  We don't hear 705kHz steady state frequencies, but we do hear timing cues at a time resolution that is enhanced at those frequencies.

 

If DAVE were able to support 32fs, I'd definitely be upsampling to 32fs (vs 16fs).  I don't think we're close to topping out on time resolution yet.

 

That being said, upsampling 44/16 content to 16fs vs upsampling 192/24 content to 16fs, I generally give the nod to the 192/24 starting point.  Mathematically, they should be very very close.  I suspect the difference is damage done when the 44/16 content was decimated from a higher resolution master.  If I start with 192/24, decimate to 44/16, then upsample, the differences are very very small.

 

As always, get content in the format that is as close to the original mastering format as possible.  

 

I defer to @The Computer Audiophile on whether we should have a audiophile wine topic in addition to an audiophile tea topic, but can't wait until we all can have proper listening/tasting sessions again (it's been a long year)