[Roam]

Quote:

Originally Posted by Faust2D /img/forum/go_quote.gif IMHO good output stage can make or break a DAC. Lots of the new up-over-under sampling is just marketing to sell new DACs. Just look at TDA1451 or TDA1453 DAC designs that sound superb.

Agreed, but with a caveat. The reason the ancient TDA chips sound relatively good is because the design implementation is so simple that it's almost impossible to screw up.

 

[regal]

Quote:

Originally Posted by n_maher /img/forum/go_quote.gif They are, afterall, exactly the same thing really. One just packages all of the components into a tiny housing for easy of use.

Here's one of my favorite posts on Head-fi from a designer I respect more than the big commercial guys.

http://www.head-fi.org/forums/2217292-post11.html

 

[granodemostasa]

Quote:

Originally Posted by regal /img/forum/go_quote.gif Here's one of my favorite posts on Head-fi from a designer I respect more than the big commercial guys. http://www.head-fi.org/forums/2217292-post11.html

i guess that link settles this dispute then, there really is little to say after this...

 

[Herandu]

Quote:

Originally Posted by granodemostasa /img/forum/go_quote.gif i guess that link settles this dispute then, there really is little to say after this...

Terrible, isn't it? Sometimes I feel like setting fire to all my certificates and diplomas and sending my staff home for good. Why waste time with a U$15 OP627 when a couple of 5 cent transistors would be a far better idea? Would make for far cheaper products..

But wait! Since transistors are cheaper, but better in a DAC output stage, why are they not found in cheaper DACs? Should we demand that cheaper DACs use them so that they too can charge an esoteric price?

Ori shows a total lack of IC designing knowledge! The circuit representation of a IC is just that. It is not a couple of transistors and resistors fused together in a nice box shaped format with pins sticking out.

 

[Cauhtemoc]

Quote:

Originally Posted by n_maher What evidence do you have to support anything that you just said? How easily could a discrete output stage be implemented? How much better, at the same price point, could it sound? It's all well and good to propose stuff like this but it's a bit naive to think that companies like benchmark, lavry, etc, are intentionally doing it the wrong way.

They use opamps because it makes them measure excellent in terms of THD. The NE5534 is an excellent example of this, it's cheap ($0.60) and it measures very good, but it sounds very cold and sterile. The DAC1 costs $1000, so you'd expect something better than $0.60 opamps.

 

[chesebert]

Quote:

Originally Posted by Herandu /img/forum/go_quote.gif Terrible, isn't it? Sometimes I feel like setting fire to all my certificates and diplomas and sending my staff home for good. Why waste time with a U$15 OP627 when a couple of 5 cent transistors would be a far better idea? Would make for far cheaper products.. But wait! Since transistors are cheaper, but better in a DAC output stage, why are they not found in cheaper DACs? Should we demand that cheaper DACs use them so that they too can charge an esoteric price? Ori shows a total lack of IC designing knowledge! The circuit representation of a IC is just that. It is not a couple of transistors and resistors fused together in a nice box shaped format with pins sticking out.

I don't think Ori said anything about transistor/resistor fused together; there are issues with common substrate resistors namely the p/n junction kind and MOS kind; although I think thin film solves most of that problem; but is expensive to make assuming you want abs, say 0.1%, tolerance. similar stuff with caps, not the most desirable kind whether you are doing revers p/n junction or slab a SiO2 between metal1 and n-well. so you run into issues like nonlinear values, parasitic cap, leakage through the substrate (although not a big problem).

This is as much as I know about analog VLSI layout

CMOS VLSI design is arguably much simpler, although SOI brings its own set of problems.

 

[Herandu]

Quote:

Originally Posted by Cauhtemoc /img/forum/go_quote.gif They use opamps because it makes them measure excellent in terms of THD. The NE5534 is an excellent example of this, it's cheap ($0.60) and it measures very good, but it sounds very cold and sterile. The DAC1 costs $1000, so you'd expect something better than $0.60 opamps.

I fully agree with you that the DAC1 at its price range isn't doing itself any favours by using cheap opamps. Where does the rest of its U$1000 price tag come from? The TC-7510 digital audio receiver, DAC chip and opamp output stage cost about the same as the DAC1, give or take a couple of U$. But at the price difference between the two I would be looking for OP627, low ESR electrolytic caps, gold plated PCB, etc in the DAC1.
The high end DAC market seems to be a mugs game for most of the time.

 

[Cauhtemoc]

Quote:

Originally Posted by Herandu I fully agree with you that the DAC1 at its price range isn't doing itself any favours by using cheap opamps. Where does the rest of its U$1000 price tag come from? The TC-7510 digital audio receiver, DAC chip and opamp output stage cost about the same as the DAC1, give or take a couple of U$. But at the price difference between the two I would be looking for OP627, low ESR electrolytic caps, gold plated PCB, etc in the DAC1. The high end DAC market seems to be a mugs game for most of the time.

Have you seen the inside of the Transporter? It costs $2000 and has an output stage made up of the same $0.60 opamps, but this time they're AC coupled with cheap electrolytic capacitors. When are people going to learn...

 

[chesebert]

Quote:

Originally Posted by Cauhtemoc /img/forum/go_quote.gif Have you seen the inside of the Transporter? It costs $2000 and has an output stage made up of the same $0.60 opamps, but this time they're AC coupled with cheap electrolytic capacitors. When are people going to learn...

I think the transporter uses opamp not for IV conversion but mere acting as a summoner/LPF for the +- output of the DAC. I don't think the opamp actually amps anything seeing how the DAC produces 2.95Vpp on either the 0 or 180, the fact that the balanced out is only 3Vpp means opamp is def used as a LPF/summing instead for amping. I am not sure if the quality of opamp matters in that case. If I am correct those cap you are seeing are part of the LPF

And I am guessing the unbalanced out gets another opamp for summing the +- signal.

 

[Cauhtemoc]

Quote:

Originally Posted by chesebert I think the transporter uses opamp not for IV conversion but mere acting as a summoner/LPF for the +- output of the DAC. I don't think the opamp actually amps anything seeing how the DAC produces 2.95Vpp on either the 0 or 180, the fact that the balanced out is only 3Vpp means opamp is def used as a LPF/summing instead for amping. I am not sure if the quality of opamp matters in that case. If I am correct those cap you are seeing are part of the LPF And I am guessing the unbalanced out gets another opamp for summing the +- signal.

Check page 4 of the AK4396 eval board manual for the circuit used in the Transporter.

http://www.asahi-kasei.co.jp/akm/en/...6-sbw2-02e.pdf

I see that you have a Transporter. If you don't believe me when I say that the output is crippled, try opening it up, bypass the coupling capacitors and replace the opamps with some LME49710 for example. You will be amazed.

 

[ezkcdude]

You don't need to spend thousands of dollars to get a great sounding DAC. You can build one yourself for a couple of hundred that will sound very respectable.

 

[chesebert]

Quote:

Originally Posted by Cauhtemoc /img/forum/go_quote.gif Check page 4 of the AK4396 eval board manual for the circuit used in the Transporter. http://www.asahi-kasei.co.jp/akm/en/...6-sbw2-02e.pdf I see that you have a Transporter. If you don't believe me when I say that the output is crippled, try opening it up, bypass the coupling capacitors and replace the opamps with some LME49710 for example. You will be amazed.

I know exactly how you feel; the transporter output is a verbatim copy of the test board from AKM. No coupling cap just load caps, but still any audiophile would just cringe looking at the schematic. Reality is a little different however, I hate to say it, but the damn thing actually sounded pretty decent. Maybe those god damn EE at AKM actually knows what sounds good

To give those jelly bean opamp some credit, they are running at +-15V, I mean pushing 30V into an opamp is nothing to sneeze at (their perf will def improve)

I would be afraid to 'mod' since AKM4396 by itself was known as a 'cold', precise/3D DAC; although that's certainly not the case in the transporter. (the thing is almost every bit as warm as my Saturn)

 

[Cauhtemoc]

Quote:

Originally Posted by chesebert I know exactly how you feel; the transporter output is a verbatim copy of the test board from AKM. No coupling cap just load caps, but still any audiophile would just cringe looking at the schematic.

I have it on good authority that there are indeed electrolyic coupling capacitors, if you don't believe me, open up your Transporter and trace pin 20-23 on the AK4396

Quote:

Originally Posted by chesebert Reality is a little different however, I hate to say it, but the damn thing actually sounded pretty decent.

I never said that it doesn't sound good, the AK4396 eval board also sounds half decent. I said that it could sound much better.

Quote:

Originally Posted by chesebert Maybe those god damn EE at AKM actually knows what sounds good

The problem with sound quality is that it is very subjective and hard to measure. The eval board is designed to measure good in terms of noise and THD. It doesn't sound bad, but it could sound much better.

Quote:

Originally Posted by chesebert I would be afraid to 'mod' since AKM4396 by itself was known as a 'cold', precise/3D DAC; although that's certainly not the case in the transporter. (the thing is almost every bit as warm as my Saturn)

The LME49710 is quite warm once properly burned in, at least more so than a NE5534. It can also handle higher supply voltages than a NE5534.

Six LME49710 would cost you less than $20, a quite cheap mod if you ask me...

 

[chesebert]

Quote:

Originally Posted by Cauhtemoc /img/forum/go_quote.gif I have it on good authority that there are indeed electrolyic coupling capacitors, if you don't believe me, open up your Transporter and trace pin 20-23 on the AK4396 I never said that it doesn't sound good, the AK4396 eval board also sounds half decent. I said that it could sound much better. The problem with sound quality is that it is very subjective and hard to measure. The eval board is designed to measure good in terms of noise and THD. It doesn't sound bad, but it could sound much better. The LME49710 is quite warm once properly burned in, at least more so than a NE5534. It can also handle higher supply voltages than a NE5534. Six LME49710 would cost you less than $20, a quite cheap mod if you ask me...

you are right the coupling cap is for the opamp input.

Verbatim copy of the eval board is not very 'engineered thing to do' on the part of SD.

its one of those could have, would have, should have kinda thing. I am not sure swapping the opamp would bring much benefits (for the amount of efforts involved; had they socketed those damn chips it would have been much easier, alas that's not the case)

While its fine sounding (it sounds great when you consider each element individually e.g. soundstage, bass, HF extension, etc), opamp based circuits lack the ease and flow of a finely designed class A discrete circuit; and there isn't anything you can do to change that (I feel the same way about G08's opamp stage too).

 

[Herandu]

The 5532 and 5534 have a maximum input level of 0.6V, before a set of back to back diodes clamp any signals above that level. Most good DAC chips produce an output above that 0.6V level however, so the input signal to a 5532 would have to be reduced in gain prior to it being fed to a 5532. If you don't reduce the gain the 5532 will clip the signal. When I had my DAC1 I was always wondering about this, since a clipped signal can sound like if the dynamic range is compressed. So faint sounds will sound far louder now, which gives the impression of a more detailed DAC.

Using transistors in a line amp design on a higher supply rail is something I also question. Given that the max expected line output level of a DAC or CD/DVD player would be no more than 2Vrms, what are the advantages of having a supply rail in excess of 12V+/-? There is ample headroom in the supply line to cope with the DAC max level output.

The use of a higher supply is an advantage with headphone amps as far as coping with high impedance cans are concerned. But again, even 12+/- is enough.