New to DACs; After Some Research, is Texas Instruments PCM1794 the best DAC?

[elg2001]

thanks for the links guys! keep em coming if you know of any.

as for what i said before, why is it rediculous? why is there such a HUGE spike around 1 kHz? something had to have been wrong.

 

[elg2001]

I wasn't trying to be offensive above, so I hope it didnt come off that way. I was simply asking a question because things didnt seem right in the graphs.

 

[Gluca]

1kHz is the test signal ... anything else is noise (so you really want to see that spike as higher as you can)

Ciao
Gianluca

 

[pooge]

For a DIY'er looking to use the PCM1794, the Rakk Dac from K&K Audio is the ticket. Very simple to put together with a couple of stuffed boards. They also have active and passive transformer I/V options, and the processor board is layed out to easily mate with a Tent clock, if desired.

Also, Channel Islands has a DAC that uses this processor.

RAKK DAC

Channel Islands DAC

 

[elg2001]

sorry about the confusion before. i just skimmed through it and figured those graphs were freq response graphs.

moving on, those are the exact two DACs that I've narrowed my search down to. Is anyone familiar with the performance difference between the two? If their two main chips are exactly the same(the TI and cirrus logic chips), can their performance really be that different?

 

[hifimaker]

For a DIY'er looking to use the PCM1794, the Rakk Dac from K&K Audio is the ticket. Very simple to put together with a couple of stuffed boards. They also have active and passive transformer I/V options, and the processor board is layed out to easily mate with a Tent clock, if desired.

Also, Channel Islands has a DAC that uses this processor.

RAKK DAC

Channel Islands DAC

The RAKK Dac looks very interesting, it's just too expensive. The powersupply is questionable, low jitter PLL is extra cost, and te passive I/V is $279. All the components cost more than a kick butt commercial DAC like the Lavry DA-10.

Another commerical DAC that looks interesting is the Bel Canto DAC3. Discounted it is available for $1800. I know a lot of money, but it does have a host of killer features like USB, TOS, and SPDIF inputs, jitter reduction, balanced out discrete class A I/V. It also uses dual 1792's in dual mono configuration.

-David

 

[Terry Demol]

You will probably find this link interesting (one would raise an eyebrown on the project btw)

http://www.audiodesignguide.com/DAC_final/DacFinal.html

Ciao
Gianluca

This DAC comparison appears flawed. The LL7903 transformer should
not have been used for I-V. The distortion added by the transformer
is significantly more than -any- of the DAC chips tested through it.

Further, using the LL7903 will result in some voltage swing at the
DAC OP which will affect the sound in varying degrees depending on
the particular dacs sensitivity to OP swing.

Use a fully balanced differential opamp such as TI OPA1632 for fully
balanced I-V and for the voltage opamps just insert suitable series
resistors before opamp IP's.

This way:

- All opamps are going through same analog stage.
- The analog stage has sufficiently low distortion in the same range
as DACs
- The current OP dacs will see their required true virtual ground.


Cheers

Terry

 

[Bernhard]

The LL7903 transformer should
not have been used for I-V. The distortion added by the transformer
is significantly more than -any- of the DAC chips tested through it.

Any suggestions for a good transformer 1:4 up to 1:8 ( 75 to 150 ohm primary / 600 ohm secondary ) ???

 

[Spartacus]

Further, using the LL7903 will result in some voltage swing at the
DAC OP which will affect the sound in varying degrees depending on
the particular dacs sensitivity to OP swing.

Hi Terry, I think you are perhaps getting mixed upbetween the RAKK DAC and the comparison test, in which all the DACs are either voltage out or use op amps for I/V.

 

[Terry Demol]

Any suggestions for a good transformer 1:4 up to 1:8 ( 75 to 150 ohm primary / 600 ohm secondary ) ???

The best transformers are Jensen, however Cinemag are also very
good and much cheaper.

They both publish complete specs for their products.

Personally I wouldn't use a transformer.

cheers

Terry

 

[rfbrw]

Hi Terry, I think you are perhaps getting mixed upbetween the RAKK DAC and the comparison test, in which all the DACs are either voltage out or use op amps for I/V.

Can you really have missed all the references to transformers on that web page ? I wouldn't have thought it possible. Only the AD1955 in dual mono has an active i/v.

 

[Terry Demol]

Any suggestions for a good transformer 1:4 up to 1:8 ( 75 to 150 ohm primary / 600 ohm secondary ) ???

Most mic transformers are specced for 150ohm prim so 1:8 will have
10k OP Z. It will really need a buffer with that high OP Z.

Jensen have a moving coil that does 1:12 with 5ohm IP Z, but to get
a nice OP level requires a few parallel dacs to get sufficient current
swing.

cheers,

T

 

[Bernhard]

Personally I wouldn't use a transformer.

The transformer + small parallel cap on secondary works good as a filter for non os DAC.

Can I get the same effect with a coil + cap ???

 

[Spartacus]

Can you really have missed all the references to transformers on that web page ? I wouldn't have thought it possible. Only the AD1955 in dual mono has an active i/v.

I'm afraid what you have missed is that the transformers aren't being used for I/V. You've also missed that the Burr-Brown based DAC also uses op-amps. Look more carefully in future.

 

[rfbrw]

I'm afraid what you have missed is that the transformers aren't being used for I/V. You've also missed that the Burr-Brown based DAC also uses op-amps. Look more carefully in future.

 

[rfbrw]

I'm afraid what you have missed is that the transformers aren't being used for I/V. You've also missed that the Burr-Brown based DAC also uses op-amps. Look more carefully in future.

Two are voltage out and the others bar the mono dacs use a resistor in combination with a transformer unless you think the transformer winding will not interact with the resistor. The chap did use an active for the '1798 but there are no measurements for that.

 

[Spartacus]

The Crystal and the Wolfson are voltage out. The AD part uses active I/V. The final DAC, the BB, uses resistive I/V (including the transformer) for the first implementation which apparantly didn't work so well, and active I/V in the second more successful one.

So, all DACs end up with active I/V.

 

[rfbrw]

So, all DACs end up with active I/V.

Not as far as the measurements shown are concerned.

 

[Spartacus]

Not as far as the measurements shown are concerned.

Sure, but that's not what you were saying originally.

 

[Terry Demol]

Sure, but that's not what you were saying originally.

I think to be fair, credit has to be given for making an effort to
do this not inconsiderable task.

We have thought many times of doing a complete dac comparison
but the logistics get pretty overwhelming if you want to do it really
well WRT power supply, grounding, decoupling etc etc.

The problem with tests like this are that the preference from one
dac to another can change dramatically depending on other factors.
To make an accurate and meaningful comparison they need to be
a) totally optimal WRT decoupling, layout, grounding, PS, I-V etc.
b) absolutely identical for all dacs (thats why I suggested a
balanced differential OPA that can be set up for I or V in with just
2 resistors
c) extremely transparent and non coloured. I have used LL790x
transformers on mic preamps and they do have a mild sonic
signature.

Even using the OP275 is not the best choice for I-V, it will impart it's
own sonic picture and that will suit some dacs more than others.

Still it was a good effort to take on this task and try to come up
with some meaningful results, so I'll give credit where credit is due.



Cheers,

Terry