HOW DOES A PERFECT DAC ANALOG SIGNAL LOOK DIFFERENT THAN A CHEAP DAC

[gstew]

1 hour ago, Charles Hansen said:

 

Hi Mansr,

 

The thing that I see over and over and over in this thread is an irrational belief in the importance of the DAC chip itself. Just about everything affect the sound of an audio product, but when it comes to DACs, I would rank (in order or sonic importance the general categories as follows:

 

1) The analog circuitry - 99.9% of all DACs are designed by digital engineers who don't know enough about analog. They just follow the app note. The specs on the op-amps are fabulous and digital engineers are inherently seduced by the beauty of the math story. There are minor differences in the sound quality between various op-amps, but it's kind of like the difference between a Duncan-Heinz cake mix and a Betty Crocker cake mix. 99.8% of the op-amps are used a current-to-voltage converters with the inverting input operating as a virtual ground. This is probably the worst way to use an op-amp as the input signal will cause the internal circuitry to go into slewing-limited distortion. http://www.edn.com/electronics-blogs/anablog/4311648/Op-amp-myths-ndash-by-Barrie-Gilbert

 

With discrete circuitry, the only limit is your imagination. You are free to adjust the topology of the circuit, the brands of the parts, the active devices, the bias current in each stage - anything you can think of. Think of this as going to a world-class patisserie in Paris and seeing all the different things that can be made.

 

2) The power supplies - 99.9% of all DACs use "3-pin" power supply regulators, which are pretty much op-amps connected to a series pass transistor. Everything in #1 applies here.

 

3) The master clock - jitter is a single number assigned to measure the phase noise of an oscillator over a fixed bandwidth. It is far more i important to know the spectral distribution of the timing variations and how they correlate to audible problems. 99.9% of all DACs use a strip-cut AT crystal in a Pierce gate oscillator circuit. It's pretty good for the money but the results will depend heavily on the implementation, particularly in the PCB layout and the power supplies (#2).

 

It's hard to rank the rest of these so I will give them a tie score.

 

4) The digital filter - 99.9% of all DACs use the digital filter built into the DAC chip. About a dozen companies know how to make a custom digital filter based on either FPGAs or DSP chips.

 

4) PCB layout - grounding and shielding, impedance-controlled traces, return currents, and return current paths are all critical. For a complex digital PCB, 8 layers is the minimum for good results.

 

4) The DAC chip - almost everything these days is delta sigma with a built-in digital filter. Differences between different chips is one of the less important aspects of D/A converter designs. Both ESS and AKM have some special tricks to reduce out-of-band noise, which can be helpful, but not dramatic.

 

4) Passive parts - the quality of these can make a large difference in overall performance, especially for analog. Not many digital engineers sit around listening to different brands of resistors to see what sounds best.

 

These are just a few of the things that make differences in the way that a DAC will sound.

 

Hope this helps,

Charles Hansen

Charles,

 

Thanks for that. DANG, it's good to see you posting again!

 

Greg in Mississippi

[gstew]

55 minutes ago, beerandmusic said:

I have recently gained respect from Barrows, and he gives you the utmost respect, so i just want to thank you for coming on the boards and taking the time to share...it is noted and appreciated.

 

I am sure you are very busy with all your work, but I truly hope that you can review the LKS specs and design and provide your input.  Since you suggest the chip isn't that important and that the analog circuitry is the most important, maybe pay just attention to that part of it if you are busy.

I think the market for digital music is continuing to grow, but that most have no desire to spend $2K plus for a dac, and the market is saturated with so many choices that for the typical buyer it is just over-whelming and frustrating...especially when like you say there is not much you can hear differently spending less than $2k...that you use the $2k as "the next level" (as i like to call it).  I want, like probably many others the next level for $1K....my hope is that the LKS may be it, but i may be once again disheartened.  Who knows, maybe the pro-ject will be it.

 

My gut feeling though is that reviewers and website promoters will bad mouth anything under $2K as to keep the high end healthy...little do they realize that if they opened up the doors for highend at a low price point, their profits may be more than they anticipated.  The fact that things can't be measured and only heard makes it even more difficult as people need to make their "short lists" based on biased reviews.

 

Anyway, off my soapbox...still looking for the "next-level" DAC after probably close to 4 years of yawn.

BeerandMusic,

 

There's a very good (but long) thread on HeadFi about the LKS MA-DA003 DAC that has recently started investigating the new one using the ESS Pro chips:

 

https://head-fi.org/threads/lks-audio-mh-da003.745032/page-51

 

Poster 'b0bb' seems to be one of those rare people who understands electronics, listens, and is willing to share his knowledge.

 

IF you compare the details of those DACs as described in that thread with what Charles wrote above (and Barrows wrote in a recent post describing his experiences with getting great sound from the ESS-based DAC chips), you'll see that it:

 

1. Uses opamps in the output stages. B0bb's recommended upgrades go to an expensive, exotic discrete opamp, but it is still an opamp.

 

2. Uses the ESS chip in ASRC mode instead of synchronous

 

3. Retains all of the on-chip filtering.

 

4. Uses mostly chip regulators. In some cases very good ones, but still...

 

5. Can stand some clock upgrades... b0bb has outlined several levels, I believe he is currently using the roughly $700 USD Pulsar clock in his.

 

6. In the earlier versions of the ones using the pro clocks, b0bb outlines some issues with the USB input implementation.

 

IMHO from reading the comments above and in the LKS threadl, if you want to get a good deal on a DAC and spend some serious time and $ to improve it, it is possibly a good choice. IF you want a DAC rivaling others in the $5-$10k range, consider looking elsewhere. If you check my sig, you'll see I DIY or modify most of my gear (as much for the fun of it as anything... and two projects I have been slowly working on are to replicate a lot of what Barrows did with his personal DIY'd DACs, but with I2S feeds instead of USB). IF I wanted to go for commercial DACs, I'd sure look at the Schitt Yggy that Charles recommended as one of my top choices along with any Ayre DAC I could possibly afford.

 

My 2 cents...

 

Greg in Mississippi 

 

[gstew]

21 minutes ago, beerandmusic said:

that is the 003, i am referrring to the 004.  I have read that many people have modded the 003 for great sound, but some of those same people say the 004 is superb out of the box.

BeerandMusic,

 

Sorry I wasn't clear, when I said 'that has recently started investigating the new one using the ESS Pro chips:', that was referring to them discussing the 004 in recent times... at least since post #542 on 3/24/2017.

 

They include internal pictures of the along with sonic and design assessments.

 

What I wrote referenced comments that either pertained to both the 003 and the 004 (very similar designs in many ways) or the 004 specifically.

 

Greg in Mississippi

 

P.S.  AND Charles comments echo my much more limited experiences with getting things from China & Hong Kong... now rarely, in limited cases with vendors I trust, if at all. Or with stuff where the quality of the design and components just doesn't matter. 

[gstew]

17 minutes ago, barrows said:

Actually, using I2S (with LVDS transmission) does add extra jitter.  And it places the master clock in the source (where it should not be).  I have never understood why the current "standard" does it this way (I believe in MSB's implementation the master clock is in the DAC).  It would be much better if the master clock was in the DAC, and sent back over the I2S lines to the source, then the I2S feed should be re-clocked directly via the masterclock right before the signal hits the DAC (chip, or discrete).

I never understood this either. The only explanation I have is that PS Audio was probably doing an asynchronous reclocking at the DAC (their Digital Lens?) which if done correctly, should resolve this issue with the I2S over LVDS PS Audio Standard. This is similar to using the DIY tool by Ian Canada, his Asynchronous FIFO.

 

And thinking about it more, it does make it more universal and easier to add on if you aren't expecting the source component to use the DAC's clocks. BUT definitely, not the best way.

 

There was another implementation that did it correctly, but more of a DIY kit effort than a product. This was the DAC produced by Chiaki & Bunpei of Japan designed to work with their SDTrans384 SD Card Player. I don't know that very many of these were ever made though.

 

Greg in Mississippi

[gstew]

44 minutes ago, barrows said:

There are ways to do it.  Mr. Hansen referred to the SPDIF receiver in the QX-5 as being essentially "jitter free".  I know a bit about Ayre's history with digital products and they have been no fan of SPDIF over the years, until now, precisely because they do not believe in the source being the master.  Mr. Hansen mentioned that they have  new, patented, tech for SPDIF reception, which I then assumed was an SPDIF receiver which operated asynchronously.  He may not want to reveal what the tech really is.

You can have an SPDIF receiver operate asynchronously if you have a deep enough buffer.  Of course with long (or endless) playlists eventually you will have a problem.

GStew mentioned the DIY product: IANFIFO, which is a deep buffer async re-clocker which will asynchronously re-clock any (I2S) source to a new masterclock domain without control back to the source: this works as long as the playlist is not too long (combined with the data rate). 

Regarding alternative S/PDIF operation modes... John Swenson did an interesting  S/PDIF implementation for a DAC he designed for the DIY kit company Bottlehead a few years back:

 

There are two different clock "topologies" for a DAC:

Source is master, DAC is slave
Source is slave, DAC is master

What matters is the jitter at the DAC chip (in the DAC). Having a fixed frequency clock in the DAC, right next to the DAC chip is the best way to implement this.  You can do this when the DAC is master; it has the "master clock". USB can do this in "asynchronous mode". The BH DAC uses asynchronous mode so it can do the DAC as master. In this mode the source (usually a computer) sends the data out, but the DAC can tell it to speed up or slow down so the average data rate matches the clock in the DAC. 

There is another USB mode called adaptive, in which DAC is the slave, but the BH DAC does not use this.  SOME other DACs use this mode.

The S/PDIF inputs (coax and optical) just work with the source as master and the DAC as slave. Thus the DAC has to somehow synchronize its clock to the data rate from the source. This is traditionally done with a device called a PLL, which is built in to all the S/PDIF receiver chips. PLLs have much higher jitter than a good fixed frequency clock. The BH DAC does not do it this way. It cleans up the S/PDIF signal, and sends it into an FPGA (field programmable gate array) which does the S/PDIF decoding. But the special part is a digitally controlled ultra low jitter clock. This is almost as good as the best fixed frequency clocks. The FPGA tells this clock to speed up or slow down so it is synchronized to the average data rate of the source. 

The result of this is that both S/PDIF and USB produce ultra low jitter to the DAC chip. This combination of ultra low jitter from BOTH S/PDIF and USB doesn't exist in any other DAC. On other DACs one or the other will be significantly worse than the other input.
 

So another way to produce a very low-jitter S/PDIF interface. More details in this thread:

 

https://bottlehead.com/smf/index.php?topic=6898.msg67123#msg67123

 

I have some Ian FIFO's here, but have not finished an implementation with them as yet. While Ian's earlier versions had a set delay (to allow some data depth in the buffer), his latest multichannel version can be set between .1s and 1s. BUT it has to have some delay.

 

OTOH, I do use the conceptually similar Kali R-Pi reclocker from Allo.com on a number of setups, along with a couple of Soekris DAM DAC setups using his own FPGA-based FIFO reclocking (using a variable-frequency clock, not dissemilar to what John did above, but on I2S setups). I'v'e been running the Soekris setups for about 1 1/2 years and the Allo.com Kalis for about a year and have not encountered any situations where the buffers audibly over or under ran. I suspect most sources are sufficiently close on clocking that this is rare. BUT my experience on this is not comprehensive or scientific. 

 

Greg in Mississippi

[gstew]

1 hour ago, barrows said:

On the various discrete R2R DACs out there (TotalDAC, Holo, MSB, Soekris, etc)  I have never understood why these guys do not get Texas Components to make them a custom resistor network with with z-foil construction.  Perhaps no one has enough volume to justify the development, but it seems to me a single foil for each network, could be more accurate.  Someone with better understanding of high speed circuits (jabbr?) might weigh in here.  It seems to me that having a resistor network comprised (say in TotalDACs case) of .01% foil resistors, which are mounted through hole on legs adds a lot of solder joints which would contribute to inaccuracies.  MSB is cagey about their approach, and their ladders are housed in shields so you cannot see them, I asked one of their engineers what they were using in there, but they wanted to keep this information to themselves.

Soren says that he uses a 'sign-magnitude' configuration in his DACs which reduces the impact of resistor inaccuracy. He discusses that early-on in the monster thread on his DAM1021 boards on DIYAudio here:

 

http://www.diyaudio.com/forums/vendor-s-bazaar/259488-reference-dac-module-discrete-2r-sign-magnitude-24-bit-384-khz.html

 

Does it work? I don't have experience with any of the reference DACs discussed here to compare. OTOH, my 2 Soekris DAC setups are the best in my collection and some comparisons have put them in Yggy territory when well-implemented. Of course, your mileage will vary based on your implementation.

 

His later versions of the boards use some custom resistor values and a lower resistor count in the ladders, lowering solder joints. 

 

At $200 - $400 USD at current exchange rates for the various resistor-precision levels of the DAM1021 or DAM1121 boards, for a DIY'er, they seem to be a good value IF you prize the sonic signature of R2R DACs. I do, though tempered with my limited experience with modern high-end DACs.

 

[gstew]

38 minutes ago, jabbr said:

 

The PCM1704 / PCM63 DACs were probably the pinnacle in traditional "simple" R2R technology. On chip laser trimmed.

 

You need crazy accurate resistors to get to 20 bits let alone more. Not only that but they need to be tightly thermocoupled.  No individual resistor has these tolerances so you *could* hand select and then lace in an oven but ... lots and lots and lots of work...

 

Soekris does a lot of handwaving about this but there are those who are suspicious.

 

<SNIP>.

 

 

I don't have the technical chops to debate Soren's technique. BUT as a DIY module, it does work ok and can be implemented fairly easily (though needing more extensive power supplies and in the case of the DAM1021, modifications to get the best out of them).

 

 

8 minutes ago, jabbr said:

 

I don't know of any DAC that would vary the clock... there are typically layers of FIFO. Let's say you have 2 DMA buffers. The FIFO reads from one buffer while the other buffer is being filled by the processor. If both buffers are filled the processor doesn't fill the DMA buffer. If both DMA buffers are empty, there is an underrun. As long as music comes into the system fast enough, it is properly clocked out to the DAC.

 

The Soekris DACs do a hybrid with a low-depth FIFO and a variable clock. IF I continue with them long-term, I'll likely do an implementation with the premium DAM1121 module with the better variable clock.

 

Greg in Mississippi

[gstew]

Just now, jabbr said:

 

See above, I prefer my fixed clocks to generate the needed audio rate clock (BCLK) via integer division or multiplication which is easy for an FPGA to do --- otherwise the music will be at the wrong rate...

 

No argument. I know the extreme experimenters at TirNaHiFi.org tried some fixed clock setups with the DAM1021 with varying results, but reported good improvements when they worked.

 

I'd prefer a FPGA FIFO with isolation afterwards and a post-isolation standard dual-clock family reclocking board with the reclocking output set in HW, not FPGA or SW, ala Ian's FIFO setup. That method could also allow for clock upgrades.

 

But Soren sadly didn't ask for my design preferences. At least the  DAM1121 does do the final reclocking in HW after the FPGA.

 

Greg in Mississippi

[gstew]

17 hours ago, barrows said:

I totally disagree with those who find the fault in ESS DACs.  I put this down to a lot of poor ESS implementations out there.  I DIY with ESS based DACs, but it takes a certain approach to get out what they are actually capable of.  (Same is mentioned by Charlie on the QX-5).  There are a lot of different ways to use the ESS DACs.  They are extremely sensitive to power supplies, for just a single example:  There are two power supply inputs of the ESS chip for just the analog section supply alone.  I have AD 715x based regulators, local, to those two supply pins.  With these regs SQ is very good.  Then I changed two 1 µF capacitors, one of which filters the voltage reference, from ceramic caps to PML film caps, big sonic gains!  I was not really expecting to hear any improvement, just did the mod as an experiment really.

That is just one small example.  Here is another thing to consider re ESS DAC implementations: how many commercial DACs run the ESS chip synchronously?  I know of one offhand (Ayre).  Most just use a single 80-MHzZ-100 MHz clock which then engages the ESS onboard ASRC.

Of course I do not really expect those who have the "DSM is evil" attitude to agree! 

 

Barrows,

 

Following your comments closely for my own ESS implementations. Great to hear on the PML caps (we discussed that line a year or so ago), I've used them on several cheapie/DIY DAC builds/mods recently and found them to be good improvements over the previous ceramics. 

 

I remember one of the first places I believe I saw the SMD acrylic film caps used was in a... wait for it... Ayre QB9. Of course the picture was pretty bad, so I might have been mistaken.  I've seen them in several other production and DIY'd DACs since.

 

Thanks for the info!

 

Greg in Mississippi

[gstew]

18 hours ago, Charles Hansen said:

 

Hello Stew,

 

I've never run a critical clock signal through any FPGA (including the dedicated clock networks built in) that didn't seriously degrade the sound of the converter. YMMV.

 

Best,

Charles Hansen

Charles,

 

Thanks for that advice. I'm just using the best (AFAIK) of the available tools for the DIY'er. 

 

What other configuration or topology would you suggest instead?

 

TIA!

Greg in Mississippi 

[gstew]

On 9/4/2017 at 2:15 PM, gmgraves said:

 

Well, the first part is right. I did modify a pair of Acoustat Electrostatic speakers and shipped them back to Acoustat when they were owned by a company in the Phoenix area called the Rockford Group (who also owned the Hafler name at the time and were selling very good pro power amps called TransNovas at the time). What I did was change the woofers (they were hybrids - you know, like M-Ls) The 8" woofers were terrible so I found a better pair, and I told them that the transformers in the ESL circuit saturated too easily, and it squashed the dynamics of the ESL panel.

 

I have been reviewing for the Audiophile Voice for the last 22 years, and working for Gene Pitts (late of Audio Magazine). Well, I never lived in Santa Fe, but I was there on occasion because that's where Stereophile was and also where Gordon Holt lived, who, as you know, was one of my closest friends. At the time I was writing for Stereophile, I was living in Silicon Valley CA (Mountain View, to be more exact), but now I'm retired from the "Silly Valley" rat-race and I'm living in Reno, NV. 

 

You bet your booties that I regret selling my Marantz 8B! I sold it because I had obtained a pair of VTL 140 Monoblocks. These powerful tube amps used SIX 807 output tubes per channel and they were very good. But when I sold them, I didn't get anywhere near what a good 8B is worth these days!  

 

The car is resto-modded 1986 Alfa Romeo GTV-6. When I got finished with that glorious overhead-cam V-6 it put out nearly 250 BHP, and in a 2500 pound car would propel it to 60 in 5.6 seconds (the stock was eight seconds) and at 7200 rpm in 5th gear, the car would hit 160 MPH and sound like a banshee from hell while doing it. Nobody does engines like the Italians!

 

Nice of you to remember me. Charlie.

Regards,

George

George,

 

DANG! You're making me miss my long-gone Alfetta & later Milano. Loved my Alfas!

 

AND my Acoustat 2s, 2+2s, and 1+1s (especially after the Phil Keck Mods). Still have the 1+1s... should break them out someday!

 

Greg in Mississippi