The State of DACSES

[Mr.Sneis]

R2R is ultra popular again, manufacturers seem to be building them but vintage stuff also.  Schiit and Metrum I think should be credited for a bit of this resurgence and making the cost somewhat affordable.  I think Audio-GD is still big pimping pcm1704.

USB is still crapshoot, it's gotten a lot better but the use of $$$$ expensive converters has gotten even more insane (like audio over ethernet).

PS Audio really jumped the shark IMO with their super expensive Directstream DAC upgrade since the perfectwave days.

Not too sure what happened to the cetoole analog upgrade board mods for older opamp based dacs, I think they disappeared or dried up.

I've relegated myself to the vinyl realm and tried to look back but then got super screwed so I'm pretty much 98% out of the dac games now.

Edited May 5, 2017 by Mr.Sneis

[mypasswordis]

What ended up happening with the Assemblage? Could pcx repair it or at least figure out the problem?

[Mr.Sneis]

Been up with them for about a month.  No progress made yet but when I spoke to their tech last he had said he needs to get the schematic from some guy in NL who may or may not have had it.  He was pretty sure the problem lied in the digital board at least.  Should probably try to hit them up again.

Edited May 5, 2017 by Mr.Sneis

[mypasswordis]

Mind sending the schematic my way if he finds it? I can't imagine it's that different from the 3.0 schematic which you can just google and it's still even on their site... would assume the main difference between them is PMD200 vs PMD100 but who knows. Definitely hit him up again. I have found for the most part that waiting for customer service to get back to you just means absolutely nothing is happening. 

[Mr.Sneis]

That 3.0 schematic on their site but is awful resolution!

[justin]

2 hours ago, dsavitsk said:

---------^ That.

Tubes are excellent as voltage amplifiers, but they are high impedance devices. For a medium amount of gain, they are about the most linear option available. But a DAC doesn't need the gain, and the impedance is a hindrance.

[gepardcv]

13 hours ago, Hopstretch said:

See page 6 below.

PCM1704.pdf

Pasting from the PCM1704 datasheet:

Quote

The sign-magnitude architecture, which steps away from zero with small steps in both directions, avoids any glitching or large linearity errors, and provides an absolute current output. The low-level performance of the PCM1704 is such that true 24-bit resolution can be realized around the critical bipolar zero point.

How bad do linearity errors get with non-sign-magnitude designs?

[K3cT]

I still use a modded Parasound 1100HD with cetoole's boards that I bought here all those years back as my main DAC. I had recapped the whole thing and put in some genuine K-grade chips as per Filburt's recommendations. 

It still can stand toe-to-toe against a Gungnir Multibit which is not bad for an ancient relic. 

Edited May 5, 2017 by K3cT

[RudeWolf]

On 04/05/2017 at 10:23 PM, The Monkey said:

Which chipses tend to sound best?

Looks like it's the age of no-chip.

I was very impressed by the sound of first edition Soekris boards. It took two extra revisions of the board and around a year of firmware updates to get there, but now it's pretty stellar. I'd love to hear what DAC1541 has to offer.

I'd also say that some of the higher end implementations of Asahi Kasei chips sound really nice. I went from DAM1021 to RME ADI-2 Pro and don't feel like I'm missing anything sound wise.

[kevin gilmore]

r2r chips that are 24 bit seem impossible to make, or not enough demand yet to develop them.

soekris,totaldac,msb,holo,denafrips etc all use the same lvc595 for the switches, and all are exactly the same idea. but the new soekris is the only one yet with a discrete class A output stage. And a rational price.

[TMoney]

So if one were in the pre-built DAC market then this guy is the one to consider? Mafia approved?

http://www.soekris.dk/dac1541.html

[kevin gilmore]

I will be buying one as soon as they are available. can't touch the price/performance

also for non r2r the teac nt503 continues to be amszing

Edited May 6, 2017 by kevin gilmore

[Dusty Chalk]

On 5/5/2017 at 0:55 PM, Pars said:

It seems that most who want a tube output DAC are not looking for the most linear stage they can get, but instead are looking for the "tube sound" in an attempt to quell the digital nasties

Yeah, but the same can be accomplished by running it through an external tube stage.

[dsavitsk]

59 minutes ago, kevin gilmore said:

also for non r2r the teac nt503 continues to be amszing

The DAC chip they use is $5 at Digikey.

1 hour ago, kevin gilmore said:

I will be buying one as soon as they are available. can't touch the price/performance

What am I missing? This just looks awful:

 

[kevin gilmore]

The nt503 uses the ak velvet sound chips for both dsd and Pcm also converts pcm to 2x dsd

as far as I know there were no specs published on the new soekris dac

Edited May 6, 2017 by kevin gilmore

[RudeWolf]

Output spectrum is actually something I would expect to see from such ladders. With that said - most of the harmonics are -100dB below the signal, so it's okay in my book. I mean - I've heard better measuring DACs which have sounded far worse. Stock Buffalo 3 for example.

Btw, this might interest the mafia members in possession of DIY Soekris modules.

[morphsci]

Whenever I pronounce Soekris I think of this ...

[nopants]

I guess the switchers were a cost/space consideration? I'd expect that to show up in the spectrum, but you'd just need to listen. 

I'm interested enough to see how it stacks up against the yggdrasil, I imagine the in-house filter is the only thing it's got over the soekris.

Edited May 7, 2017 by nopants

[DefQon]

I personally like a tube output stage or a tube buffer after the dac I/C itself because most of the solidstate outputs there in commercial dac's are garbage sounding, some overcompensated with a high or low pass filter and a chain of shitty opamps one after the other.

Kevin any chance of a properly done tubeoutput stage schematic for current or voltage output dac's?

 

[dsavitsk]

7 hours ago, nopants said:

I guess the switchers were a cost/space consideration? I'd expect that to show up in the spectrum

I wouldn't. A good switcher is no noisier than a rectifier, and it is generally easier to filter.

 

4 hours ago, DefQon said:

I personally like a tube output stage or a tube buffer after the dac I/C itself because most of the solidstate outputs there in commercial dac's are garbage sounding, some overcompensated with a high or low pass filter and a chain of shitty opamps one after the other

 

There are other alternatives to opamps than tubes.

 

4 hours ago, DefQon said:

Kevin any chance of a properly done tubeoutput stage schematic for current or voltage output dac's?

The point is that there isn't a proper way to do it as it is the wrong tool for the job. An 8741, for instance, has more drive ability than just about any small signal tube and way less noise and a lower impedance than a power tube. Even the super high gm tubes pale in comparison to a $0.03 transistor, and they are unstable and difficult to use.

Edit: The exception to this might be if you need gain. But voltage out DACs typically don't.

Edited May 7, 2017 by dsavitsk

[Hopstretch]

34 minutes ago, dsavitsk said:

There are other alternatives to opamps than tubes.

Such as? Not trolling -- just like to learn!

[kevin gilmore]

you could do a really nice tube output stage with 6c33. But you need 4 of them for balanced. Output impedance of 30 ohms. But that is way past goofy, at least 200 watts power consumption by the time you are done. And its huge.  Or 10 tubes in parallel (40 tubes total) like the big Conrad Johnson preamp which also gets you under 50 ohms.

The Conrad Johnson thing has the benefit of very low noise.

you could do transformers and use the patent from audio precision to bias the transformer in such a way as to completely remove all second harmonics, you would be surprised at the amount of distortion that even the best low level transformers have.

my favorite is a pure class A output stage.

[dsavitsk]

Colin's IV (which is a common base topology) is a good alternative for a current out DAC. It is similar to Jocko's circuit, or Nelson Pass', and it can be made differential without too much trouble. For a voltage out DAC, a simple emitter or source follower works great as a buffer. Even a diamond buffer is a good alternative, and it can be made DC couple-able with the help of a servo (unlike some off the shelf diamond buffers.) Or any of Pass' recent buffers are good options.

There is this myth of tubes as being pleasant sounding, and tooby. The reality is that tubes are very linear amplifiers with excellent bandwidth, and the bulk of the added distortion and limited bandwidth attributed to them has to do with the topologies and the surrounding components used. Similarly, the myth of DAC digital glare probably has more to do with the opamps used in the IV than the DAC chips themselves.

[The Monkey]

On 5/6/2017 at 2:09 PM, kevin gilmore said:

 

also for non r2r the teac nt503 continues to be amszing

That thing interests me greatly...

[EdipisReks1]

2 hours ago, The Monkey said:

That thing interests me greatly...

It does look like an awfully likely device, dunnit?