"Value" DAC Options

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Perhaps (certainly!) I'm a bit behind the times but I forgot I had bought a hifime U2 DAC last year and thought I'd give it an audition. I ripped some Chesky CDs into iTunes using Apple Lossless and compared the playback with my CD player. I was quite surprised that this little 50 buck DAC sounds rather good. My CD player cost me $4k when it was new more than a decade ago and it sounds really good. My kids did an A-B comparison (younger ears) and they preferred the $50 DAC. Yikes !
 
Member
Joined 2011
Paid Member
I think the main focus of what determines the "quality" of a DAC seems to be a bit narrow. Quite often DR, noise, and crosstalk are the only things that are discussed or compared. Power supply regulation plays a huge role in noise floor so often the focus in on that, especially.

I am generally happy with devices that have DR and noise figures that are quite bad in comparison to "the best DACs". I don't really care if I can hear a tiny bit of noise when I put my ear up to the speaker cones, because that isn't how I listen to music.

I am far more concerned about the more subjective qualities of the devices. What do they actually sound like? My $40 WM8740 DAC board fed by a $30 CM6631A board is very, very pleasing. Certainly more pleasing than DACs built into sound cards and mid-fi surround receivers that I have heard. Would the $300 TEAC "blow me away"? Perhaps, but I am not prepared to spend that sort of money on a pre-built unit. I am much more interested in DIY audio, and trying different DAC boards with different DAC chips. I know for a fact that I like some better than others, and often it has nothing to at all to do with the specs and everything to do with how they sound in my system.

Synergy is difficult to find, but should never be overlooked. Through experience, I have learned that synergy is found by trying lots and lots of different components.

I can't wait for my $30 ES9023 board to arrive. Maybe it is a piece of junk - maybe I will love it anyway. For my $300, I can buy 9 more DAC boards with different chips and try them all, and have 10 times as much fun.

Horses for courses, as they say.
 
Member
Joined 2008
Paid Member
I must admit that I don't think there's really a linear relationship between SQ and price so I took your 'linear relationship' to be rather a log-linear one. Where SQ is roughly going up as log(price). I reckon after the introduction of the Yggy there will be no reason to spend more than $2300 other than for bling features like DSD.

After you posted this, I thought about it a bit, and I think I mostly agree with you. However, I still think things get exponential at a certain point. E.g., the $43k DAC, even if it's the best DAC ever built, I doubt it's log(price) better than the next-best, but more like log(log(log(price))) better. That's just speculation on my part though. (I need some way to rationalize the fact that my car cost less than a DAC.)

Given that quite a decent sounding DAC can be DIY'd for under $20 in parts (based on TDA1387 and running on batteries) yeah diminishing returns are the name of the game.

Indeed, but I think I'm in the boat with the other guys who feel a project like yours is above our means. Well maybe I'm speaking for myself. :) But the point is, I lack the experience and time (and therefore confidence) to tackle something like that at this point. I'm still in the "need a kit" or "paint by numbers" stage. Or, failing that, buy pre-made. :)

...Along the way my ears also got tweaked as at the start I was used to hearing what I now call 'false detail' from a DAC.

Could you elaborate on this idea of "false detail"?

I am far more concerned about the more subjective qualities of the devices. What do they actually sound like?

I mostly agree. You can read subjective impressions of a lot of higher-end off-the-shelf DACs in this overlong thread on Head-Fi. (And then feel bad because you can't afford a >$2k DAC.)

Speaking of subjective, though, there's no shortage of folks in that thread and many others (in just about any audio forum) saying sigma-delta DACs by design simply don't deliver. I'm hoping this tda1387 x8 DAC I have on order will be good enough to give me a taste of what non-SD can do.

Abraxalito, have you any thoughts on the Soekris dam1021? Subjective reports seem rather mixed. If I can't make a go of the tda1387 x8, that might be what I try next.

I can't wait for my $30 ES9023 board to arrive. Maybe it is a piece of junk - maybe I will love it anyway. For my $300, I can buy 9 more DAC boards with different chips and try them all, and have 10 times as much fun.

I certainly applaud your attitude! I felt that way with the tpa311x series amps (of which I now have an embarrassingly large collection). But I think those were/are a little different than DACs given their simplicity. My superficial understanding of DACs is that every sub-component in the design is critical for top-notch performance. E.g., you have to have exceptionally clean power for the digital section; separate super-clean power for the analog section; board layout and grounding of the actual digital section is crucial; critical also is the design of the post-conversion I/V stage; don't forget the EMI shielding... getting all those pieces just-right appears to take a lot of knowledge/experience/time... or at least more than I have at this point! :D
 
Member
Joined 2011
Paid Member
I mostly agree. You can read subjective impressions of a lot of higher-end off-the-shelf DACs in this overlong thread on Head-Fi. (And then feel bad because you can't afford a >$2k DAC.)

Nah. I don't feel bad. More like indifferent. What people spend their money on is none of my business and it is no reflection on me. I make six figures and have done for a long time, but you'll never catch me spending 2 grand on a DAC. I can't even hear beyond about 14-15kHz, for C's sake!

Speaking of subjective, though, there's no shortage of folks in that thread and many others (in just about any audio forum) saying sigma-delta DACs by design simply don't deliver. I'm hoping this tda1387 x8 DAC I have on order will be good enough to give me a taste of what non-SD can do.

Indeed. I have noticed the fixation on NOS DACs and I too want to know what all the fuss is about. My plan is to buy this DAC NOS1 NON Over Sampling DAC TDA1543T I2S Input | eBay and go straight from my CM6631A to the DAC via i2s. I'll provide clean, regulated power to both and see how I like it.

I certainly applaud your attitude! I felt that way with the tpa311x series amps (of which I now have an embarrassingly large collection). But I think those were/are a little different than DACs given their simplicity. My superficial understanding of DACs is that every sub-component in the design is critical for top-notch performance. E.g., you have to have exceptionally clean power for the digital section; separate super-clean power for the analog section; board layout and grounding of the actual digital section is crucial; critical also is the design of the post-conversion I/V stage; don't forget the EMI shielding... getting all those pieces just-right appears to take a lot of knowledge/experience/time... or at least more than I have at this point! :D

I read/hear stuff like this ad nauseum. One of the smoothest, most "analog" sounding DACs I have ever heard had about 4 components inside it and cost $50. It is based on the WM8524 which has a built-in charge-pump - no need for any output stage, opamps, filtering, etc. What you hear is the sound of the WM8524. I have heard better DACs, but I can't say I have ever heard one that sounds more "analog" than that one. My parents love it, and I have moved on. I am particularly interested in the aforementioned ES9023 because it too requires very little in terms of "other components". Just look at this board to see what I mean: 24bit 192kHz ES9023 DAC With I2S Input AND Ultra LOW Noise Regulator 24192 | eBay

Feed that thing clean power and I bet it sounds incredible - for $30.

I am likewise interested in the equally simple PCM5102A boards out there, like: http://www.ebay.ca/itm/261695107757?_trksid=p2055119.m1438.l2649&ssPageName=STRK:MEBIDX:IT or maybe http://www.ebay.ca/itm/390939296274?_trksid=p2055119.m1438.l2649&ssPageName=STRK:MEBIDX:IT

I love the fact that I can build a nice, clean 5V linear supply and then just swap in any of these boards in minutes.

I am selling a Yamaha HT receiver and a sub and plan to buy a few more bits and bobs to play with. They'll all eventually get used in one of my planned systems.

There's no such thing as audio perfection. Realizing this has allowed me to move on and enjoy the tinkering and simply listening to music.
 
Last edited:
Member
Joined 2008
Paid Member
Nah. I don't feel bad. More like indifferent. What people spend their money on is none of my business and it is no reflection on me. I make six figures and have done for a long time, but you'll never catch me spending 2 grand on a DAC. I can't even hear beyond about 14-15kHz, for C's sake!

Just to be clear, my "you can't afford a >$2k DAC" wasn't directed at you specifically by any means! It was the "rhetorical" you; just a joke, really.

Indeed. I have noticed the fixation on NOS DACs and I too want to know what all the fuss is about. My plan is to buy this DAC NOS1 NON Over Sampling DAC TDA1543T I2S Input | eBay and go straight from my CM6631A to the DAC via i2s. I'll provide clean, regulated power to both and see how I like it.

I bought that exact same DAC a few months ago. This is back when I saw that the tda1387x8 was going to be delayed due to Chinese New Year, but was in a hurry to have another go at the "NOS sound". I powered that DAC with a 12V Sigma11 I built a while ago, and fed it i2s from an Amanero. (I figured the Amanero could also be used for the Soekris dam1021 or possibly even the tda1387x8, or something like the DIYINHK es9023 you mention below.)

To be fair, I didn't spend a lot of time with it. But it immediately took me back to the unflattering experience I had with the $50 Muse tda1543 x4 ebay special. The track I use is "Funny Little Tragedy" by Gov't Mule. That's sort of by accident, as the album with that song was just released when I got the tda1543x4. But anyway, there's a constant ride of the cymbal through most of the song, and it just sounded lousy on the two cheap NOS DACs I tried. Lousy as in how music coming out of a child's toy sounds. That's a bit of an exaggeration, but hopefully gives you a feel for what I'm trying to convey. It wasn't unlistenable, but definitely none of the magic that has been attributed to the NOS sound.

(Depending on what it would cost to mail to you, I'd be happy to let you borrow it, or possibly buy mine for less than the new price. PM me if interested.)

If the the Schiit Yggdrasil turns out to be a game-changer (as suggested by the Head-Fi thread I linked), then perhaps there could be a resurgence in non-SD DAC technology. I'm hopeful that the magic will eventually trickle-down to value-priced DACs and simpler DIY DAC kits.

I read/hear stuff like this ad nauseum...
There's no such thing as audio perfection. Realizing this has allowed me to move on and enjoy the tinkering and simply listening to music.

Indeed! I often get caught up in the search for perfection and lose perspective on the simple pleasure of enjoying music. Fortunately I always manage to slap myself back into reality.

I am particularly interested in the aforementioned ES9023 because it too requires very little in terms of "other components". Just look at this board to see what I mean: 24bit 192kHz ES9023 DAC With I2S Input AND Ultra LOW Noise Regulator 24192 | eBay

Feed that thing clean power and I bet it sounds incredible - for $30.

I've come very close to buying that on several occasions! If you get it, you'll have to report back with your thoughts.
 
Member
Joined 2011
Paid Member
Matt,

I have to be quick here because I have to go pick up my kid at school...

- No, I didn't think the 'you can't afford' comment was directed specifically at me, I understood you, so no worries.

- Thanks for the feedback on the NOS DAC. I might take you up on your offer.

- I will definitely post my opinion about the ES9023 board.
 
After you posted this, I thought about it a bit, and I think I mostly agree with you. However, I still think things get exponential at a certain point. E.g., the $43k DAC, even if it's the best DAC ever built, I doubt it's log(price) better than the next-best, but more like log(log(log(price))) better. That's just speculation on my part though. (I need some way to rationalize the fact that my car cost less than a DAC.)

Yeah I go along with this - the 'log(price)' function does have a cut-off at the top end, which I'd suggest is going to be somewhere between $2k and $5k. Above that kinda price level I can't currently think of any engineering reason to build a more expensive DAC so I think people are just paying for bling. Just my opinion.

Indeed, but I think I'm in the boat with the other guys who feel a project like yours is above our means. Well maybe I'm speaking for myself. :) But the point is, I lack the experience and time (and therefore confidence) to tackle something like that at this point. I'm still in the "need a kit" or "paint by numbers" stage. Or, failing that, buy pre-made. :)

I'm curious by nature so what is it about my project you find so daunting? Is it soldering all those caps together into hexagon shapes? Or winding trafos?

Could you elaborate on this idea of "false detail"?

Dunno if this is going to help but I'll try. I think false detail is a kind of IMD (intermodulation distortion) which brings up the perceived level of some low-HF sounds. Kind of like a certain form of sibilance - it sharpens leading edge transients, so perhaps its like increasing 'sharpness' on a video monitor. It results in ultra-sharp imaging - meaning vocals get their position in left-right space defined too sharply. How I think it happens is when there's too much HF energy (out of a DAC normally) and this gets applied to a bipolar opamp input stage. I was able to tame it on a very early version of a TDA1387 DAC I built by putting a long string of ferrite beads between the DAC and the following amplifier.

Abraxalito, have you any thoughts on the Soekris dam1021? Subjective reports seem rather mixed. If I can't make a go of the tda1387 x8, that might be what I try next.

I haven't been following it very closely but its R2R isn't it? Which means its main weakness (excluding its analog output stage) is going to be getting the timing of the switch elements matched so as not to create glitches. The current incarnation does have opamps for the output stages which is in my estimation another weakness in that when I looked at the schematics Soekris had just implemented the normal 'objectivist' style of supplies. Also I can't help thinking it might benefit from a passive (LC) filter. Is it capable of being run NOS do you know? Somewhere I got the impression that its always doing some oversampling (8X?) - which is only going to make glitching 8X worse.

There's a lot of activity on the filters thread for that DAC which I've been reading along with - but with NOS there's no need for any digital filter, just a steep analog one. So I reckon if its able to be run pure NOS then it may well have potential. But its never going to be anywhere near as cheap as a TDA1387 DAC.
 
Member
Joined 2011
Paid Member
Member
Joined 2008
Paid Member
I'm curious by nature so what is it about my project you find so daunting? Is it soldering all those caps together into hexagon shapes? Or winding trafos?

I don't want that guy in the tda1387x8 thread to make fun of me. ;) (Just kidding, couldn't resist.)

I guess at this point it's mostly that I don't have a good understanding of the entire scope of what is involved: fear of the unknown. Maybe I missed it, but I don't think there's any build thread for it, just a schematic, right? Soldering all those caps looks tedious but I'm pretty sure I could do it. Winding transformers, though... that would be a first for me, and I don't know how to check my work when I'm done.

But really, the biggest roadblock for me is simply time. I usually only get a few hours on the weekend to work on this stuff. I consider myself "time poor", so I generally choose projects where I can make some encouraging progress reasonably quickly. And/or projects that are popular, so I have confidence that I'll meet with some degree of success.

Dunno if this is going to help but I'll try. I think false detail is a kind of IMD (intermodulation distortion) which brings up the perceived level of some low-HF sounds. Kind of like a certain form of sibilance - it sharpens leading edge transients, so perhaps its like increasing 'sharpness' on a video monitor. It results in ultra-sharp imaging - meaning vocals get their position in left-right space defined too sharply. How I think it happens is when there's too much HF energy (out of a DAC normally) and this gets applied to a bipolar opamp input stage. I was able to tame it on a very early version of a TDA1387 DAC I built by putting a long string of ferrite beads between the DAC and the following amplifier.

Interesting. Any particular song(s) that demonstrate this?

I haven't been following [the Soekris dam1021] very closely but its R2R isn't it? Which means its main weakness (excluding its analog output stage) is going to be getting the timing of the switch elements matched so as not to create glitches. The current incarnation does have opamps for the output stages which is in my estimation another weakness in that when I looked at the schematics Soekris had just implemented the normal 'objectivist' style of supplies. Also I can't help thinking it might benefit from a passive (LC) filter. Is it capable of being run NOS do you know? Somewhere I got the impression that its always doing some oversampling (8X?) - which is only going to make glitching 8X worse.

Yup, it's R2R. I don't recall anyone talking about glitching in either of the big threads on it. Though to be honest, I didn't know anything about that until I read your blog post about the Schiit Yggy, so I may have glossed over any discussion on that. My understanding regarding the opamp output is that only comes into play if you want balanced outputs. You can skip the opamps if you go with unbalanced outputs. (And at least a few people commented a subjective improvement when using unbalanced outs. Also a few people asking for the ability to use two boards in parallel for a true differential output.)

One cool feature is that you can load different filters into the FPGA that sits between the input and R2R ladder. By default it does oversample, but IIRC, somebody did develop a true NOS filter for it.
 
I am "happy" with my current DAC, however its limitations are revealed when I switch from that input to my turntable. The curtain is lifted, as they say.

At least you can hear a difference so that you can have a starting point. I don't have a turntable, so I am left with comparing various similarly priced DACs, which all sound the same to me. In general, I struggle a lot to hear a difference between digital sources. I am not getting any younger either.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.