Stereo AD/DA with S/PDIF or AES-EBU needed. Evaluation boards any good for this?

Hi,

I'm looking for a Stereo AD/DA with SP-DIF or AES-EBU digital I/O. I would ideally want something like Universal Audio 2192 but it's discontinued and was too expensive for what it does anyway.

There's benchmark and lavry, but they don't even seem to have a combined AD/DA models, only separate ones, and most importantly they are all incredibly expensive!

What are my options?

What about DIY, can evaluation boards (TI?) be used to make something like this relatively easily?

Well to your point about eval boards, I'm going to try to get one of our distributor reps to just give me one of these things:

http://www.my-boardclub.com/future_blox/miniblox.php?blox=32


You know, as a favor.  At least that will give you a solid DAC and possibly an ADC.  I'm sure there's something similar out there that isn't so brand new.

I am also interested in a DIY type project along these lines.  My hope was that I could whip something together and get a board made, all SMD of course.  But what ever happened with that project that guy did for the modular ADC/DAC?  That thing was moving along pretty well wasn't it?

http://www.rme-audio.de/products_adi_2.php

volker said:

http://www.rme-audio.de/products_adi_2.php

Click to expand...

Thanks. This is what I'm looking for. So is Behringer ultramatch pro apparently. Seeing how ultramatch is like €100 a pop, it makes one wonder why Lavry and Forssel equivalents cost 30-40X as much. Is the signal path made entirely of pure gold and the PCB's embedded with flawless emeralds? It's not like they make their own converter chips.

But this is a DIY forum and I was never interested in the easy way out, I'll have to study those evaluation board options further.

I've "build" a converter using an eval board for a friend. After configuring all the dip switches testing it the results weren't that great somehow. I'd rather buy a good converter. DIY is not really an option, good converter design and build seems far more complicated than vintage clones, at least if you want stellar performance (of which most commercial ones aren't capable either IMO, there's a precious few that don't just measure well with sinewaves but actually sound great).

On the AD sound you should be able to find a used BLA Sparrow for not too much money, beats even my modified Aurora easily, although the build quality is less than stellar (I'm on my third unit now, and it still broke down and I had to fix it myself). But it comes with a great clock and you'll really appreciate how drums keep their fast transients and voices stay crisp. They've got new models now, but they're also more expensive.

For DACs there's some info by Jim Williams on gearslutz how to mod a cheap one to make it sound good, but I don't think this will deal with the clock. There's some good information here somewhere on how to put a quality transformer in the place of an op amp on the output right behind the DAC chip as well. Joe Malone from jlmaudio has done something similar with an ADC, but I don't think they ever released the specifics.

living sounds said:

mod a cheap one to make it sound good

Click to expand...

I know many companies offer this as service, but lets face it, changing opamps and electrolytics or adding transformers in the analog stage is not going to affect the most crucial part: the conversion chip itself.

I was looking at the evaluation boards because this way I get my hands on the best possible converter chip.

Which evaluation board exactly did you configure? And maybe you didn't configure the clocking correctly. It's easy to get it wrong while the board still "works".

There was a new DA chip with internal line driver released just last year by TI: pcm5102. For input pcm4222 looks as good as ever. These are the most interesting ones to me.

Yes, it was the PCM4222. I tried all possible combinations and checked the switches many times, it didn't work very well (high THD especially).

The converter chip is actually the least relevant part. The best converter unit from the mid-90s is still up there with the best of todays top contenders in terms of sound. You can get a SNR of 120db now with a 300 EUR EMU 1820m from 2006, but it still doesn't sound good. What matters is board layout, PSU, clocking and of course the analog circuitry. The major fault with most converters IMO is - untechnically speaking - a kind of smear in the time domain that takes away the fast impact of bass, clutters the midrange in an aggressive way (makes mixing very hard) and destroys the fast transients, the "snappyness" in the high end. This all gets very obvious comparing directly from an analog source to the converted signal.

living sounds said:

it didn't work very well (high THD especially).

Click to expand...

This just tells us something in the set up was very wrong. I'd rather not get into that nontechnical audiophile stuff at all.

Kingston said:

living sounds said:

it didn't work very well (high THD especially).

Click to expand...

This just tells us something in the set up was very wrong. I'd rather not get into that nontechnical audiophile stuff at all.

Click to expand...

That might be the case. I still wouldn't bother with it, converters are the one area in DIY where I think it's best to leave it to the pros with the high end test equipment and spezialized engineering skills. Minus modding the analog stage (where I've had success) and probably upgrading the PSU.

"Audiophile"? Converter chips are the smallest part of the puzzle, that shouldn't be news to anyone. If you've got the choice make sure you get one where you can switch the digital filters to different modes and also got the option to turn them off completely. That's where the biggest audible differences in the chips lie IMO.

living sounds said:

Converter chips are the smallest part of the puzzle, that shouldn't be news to anyone.

Click to expand...

You mean the part that only does all the work, the heart of the conversion process? The rest can be filed under interfacing.

Kingston said:

You mean the part that only does all the work, the heart of the conversion process? The rest can be filed under interfacing.

Click to expand...

Maybe, but the rest is the hard and expensive part. Even in the higher end. For instance, the Lynx Aurora and the Prism Orpheus both use CS5381 and even OPA2134 on the AD side, but they exhibit clearly different performance.

To chime in on this discussion, with regard to DAC's:  I have an MAudio Delta 66, and then bought a heavily modded ART DIO from someone who put a bunch of work into improving the power supply, signal caps, etc.

The result is that I cannot tell the difference from one to the other.  To my ears there is no audible difference between DAC's.  I then went into the designs of each and found they use the same converter chip...  This has lead me to believe that the single most important part of a DAC is the chip, followed by the clock.  To prove my point, when I have a minute, I'm going to upgrade the clock in the DIO to something with top of the line specs and then compare again.

millzners said:

The result is that I cannot tell the difference from one to the other.  To my ears there is no audible difference between DAC's.  I then went into the designs of each and found they use the same converter chip...  This has lead me to believe that the single most important part of a DAC is the chip, followed by the clock.

Click to expand...

Can I ask what speaker system you made these listening test on?

ruairioflaherty said:

millzners said:

The result is that I cannot tell the difference from one to the other.  To my ears there is no audible difference between DAC's.  I then went into the designs of each and found they use the same converter chip...  This has lead me to believe that the single most important part of a DAC is the chip, followed by the clock.

Click to expand...

Can I ask what speaker system you made these listening test on?

Click to expand...

Event TR8 XL's

And as a disclaimer, I don't pretend to have golden ears or top of the line speakers, but I could not detect any difference at all A/Bing songs I know extremely well.

millzners said:

ruairioflaherty said:

millzners said:

The result is that I cannot tell the difference from one to the other.  To my ears there is no audible difference between DAC's.  I then went into the designs of each and found they use the same converter chip...  This has lead me to believe that the single most important part of a DAC is the chip, followed by the clock.

Click to expand...

Can I ask what speaker system you made these listening test on?

Click to expand...

Event TR8 XL's

And as a disclaimer, I don't pretend to have golden ears or top of the line speakers, but I could not detect any difference at all A/Bing songs I know extremely well.

Click to expand...

Hello,

I tried out those monitors when I was shopping for mines. They sounded good compared to say some KRK or M-Audio stuff. But nothing like a pair of Dynaudio, Adam's or Genlec. This might be the reason why you don't notice much difference. Personally I did not notice a big difference since i went from sound cards to an RME AD/DA. I'm talking about paring the AD/DA with what I ended up buying for monitoring (Dynaudio BM15A). This was a day and night thing. Before I had and RME AIO. Then again the ART might be as bad as your M-Audio Delta (no offense).

fabriciom said:

millzners said:

ruairioflaherty said:

millzners said:

The result is that I cannot tell the difference from one to the other.  To my ears there is no audible difference between DAC's.  I then went into the designs of each and found they use the same converter chip...  This has lead me to believe that the single most important part of a DAC is the chip, followed by the clock.

Click to expand...

Can I ask what speaker system you made these listening test on?

Click to expand...

Event TR8 XL's

And as a disclaimer, I don't pretend to have golden ears or top of the line speakers, but I could not detect any difference at all A/Bing songs I know extremely well.

Click to expand...

Hello,

I tried out those monitors when I was shopping for mines. They sounded good compared to say some KRK or M-Audio stuff. But nothing like a pair of Dynaudio, Adam's or Genlec. This might be the reason why you don't notice much difference. Personally I did not notice a big difference since i went from sound cards to an RME AD/DA. I'm talking about paring the AD/DA with what I ended up buying for monitoring (Dynaudio BM15A). This was a day and night thing. Before I had and RME AIO. Then again the ART might be as bad as your M-Audio Delta (no offense).

Click to expand...

No offense taken -- I believe the ART DIO probably has very limited ability to outperform the MAudio Delta 66 given that the two use identical 10 year old chips: AK4524, and both were priced for the same market meaning I'm willing to bet they use comparable opamps and crystals. 

Back to Kingston's point which I believe my one example confirmed, is that when you're talking about converter boxes the chip really matters, and all the analog circuitry improvements that an audiophile would recommend are limited in what they can do.  At least limited to what a pair of Event TR8's can reproduce.

Some day I'll pony up for the RME ADI-2, or at least borrow one long enough to see if I can tell the difference.

Here's a link I found helpful to this discussion:
http://www.gearslutz.com/board/geekslutz-forum/542009-audio-interfaces-their-ad-da-chips-listed.html

It's interesting to compare which units use which chips...

millzners said:

when you're talking about converter boxes the chip really matters, and all the analog circuitry improvements that an audiophile would recommend are limited in what they can do.

Click to expand...

Yes, chips matter, although very often the most expensive chip in a range offers only marginal improvement over the next in line - same thing with computer microprocessors.
But the implementation is a very important factor. Performance can be ruined by a minor detail of grounding scheme or PSU decoupling. Manufacturers publish specs based on an optimised implementation; just to achieve them is a major endeavour.
This has been proved a number of times; examples that come to mind are the first Benchmark ADC1 and the Lucid 8824, which used pretty pedestrian chips but produced much better results than many competing units based on the same chips.
Clock management is probably one of the most significant aspects.
"analog circuitry improvements that an audiophile would recommend" are just a lure. What's the point of using an opamp with 20v/us slew-rate on a signal that has 9us rise-time (at 96kHz sample-rate)?

I wasn't talking about low end chips like the AK4524 by the way, but the usual suspects in the higher end.

I did some more reading up, and the mid-90s converters I was talking about appear to be fully discrete. No converter chips used, at all.


Here's a website about DIY discrete DACs, looks interesting:

http://www.sonicillusions.co.uk/discrete_dac.htm


It's easy to see why a discrete DAC like the Lavry Gold with all the matching required and complicated circuitry to achieve outstanding specs costs a little more than an IC based design.

living sounds said:

I wasn't talking about low end chips like the AK4524 by the way, but the usual suspects in the higher end.

I did some more reading up, and the mid-90s converters I was talking about appear to be fully discrete. No converter chips used, at all.


Here's a website about DIY discrete DACs, looks interesting:

http://www.sonicillusions.co.uk/discrete_dac.htm


It's easy to see why a discrete DAC like the Lavry Gold with all the matching required and complicated circuitry to achieve outstanding specs costs a little more than an IC based design.

Click to expand...

Great link, often times during designs we run into a circuit and say "well, we could just design it all with discrete components..." and the result would often be an extra week or two of bench testing and schematic/layout, or we could just toss a chip in there and spend an hour reading the datasheet...  Sure the performance of the discrete circuit would be superior and in production cheaper to make, but it feels like overkill, and we're pressed for time.

The other thing is that companies like Lavry employ the mad-scientist types that know how to do this stuff -- it takes someone with some real engineering skills to sit down and do that.  Just about anyone can pick up a couple of datasheets and copy their eval board's design...

It's my understanding that a fully discrete converter needs lot's of space, ventilation, time to heat up, has inherently high latency and is prohibiltly expensive not least because the components need matching at a crazily high standard in order to achieve a measured performance comparable to chip based solutions. I think the Lavry Gold has special heating circuitry in order to maintain the required temperature at all time during use. It sounds totally awesome anyway, smooth as silk, not "digital" at all.


I had a talk today with an engineer of a higher end converter manufacturer who blatantly told me that in his opinion engineers need to turn off their measuring equipment once in a while and use their ears to actually find things out. He tests the analog circuitry and all the op amps by ear to find the best sounding solution. I like that.