What kind of Burr-Brown does SB2 have?

[GreenMan]

I know the SB2 has a Burr-Brown in it, but I don't know what calliber. I'm interested in maybe putting a Channel Islands VDA-1 external DAC on my SB2, but the Channel Islands unit itself has a Burr-Brown in it. Some in this forum claim an improvement in sound. Is that because of a 'better' Burr-Brown?

[cbemoore]

http://forums.slimdevices.com/showthread.php?t=15716

[seanadams]

There isn't a simple answer to this. SB2 uses the PCM1748, but the DAC chip alone does not dictate the performance of the whole system - it can be somewhat better or much worse than the data-sheet specs depending on the design. Since the DAC chip is only one component of the whole clocking + power + DAC + amplification process, the only way to really do a head-to-head is to listen blind, or measure the complete systems under identical test environments. However, even then, there is no such thing as an exhaustive test.

As a rule of thumb though, I would estimate that most standalone DACs more than 5 yrs old, regardless of price, will have lower performance than SB2. The price/performance of these chips, like CPUs, memory etc, improves with each iteration: performance levels that are relatively inexpensive to reach today were completely unavailable a few years ago.

[GreenMan]

There isn't a simple answer to this. SB2 uses the PCM1748, but the DAC chip alone does not dictate the performance of the whole system - it can be somewhat better or much worse than the data-sheet specs depending on the design. Since the DAC chip is only one component of the whole clocking + power + DAC + amplification process, the only way to really do a head-to-head is to listen blind, or measure the complete systems under identical test environments. However, even then, there is no such thing as an exhaustive test.

As a rule of thumb though, I would estimate that most standalone DACs more than 5 yrs old, regardless of price, will have lower performance than SB2. The price/performance of these chips, like CPUs, memory etc, improves with each iteration: performance levels that are relatively inexpensive to reach today were completely unavailable a few years ago.


The SB2 is a worthy technology to be sure. I really enjoy the sound of a Toshiba SD-9200 in my system, but I find that the Toshiba sounds only marginally better than the SB2 when I use superior interconnects with the SB2. The Toshiba uses a Advanced Multi-Bit Delta Sigma DAC (from Analog Devices). I don't know anything about this DAC, nor much else about the internals of the Toshiba, but I love the way it sounds in my system. However, it's a CD player with more than a few undesirable quirks, and it hasn't the cool factor the SB2 brings to my system. I've been auditioning the Toshiba and have decided to not get it and continue to use the SB2 with my own hard drives instead.

Nonetheless, I find myself in audiophile hell. Now I want a superior external DAC for my sb2. You know how it goes. (sigh).

[Fabrice Rossi]

seanadams a écrit :
> There isn't a simple answer to this. SB2 uses the PCM1748, but the DAC
> chip alone does not dictate the performance of the whole system - it
> can be somewhat better or much worse than the data-sheet specs
> depending on the design. Since the DAC chip is only one component of
> the whole clocking + power + DAC + amplification process, the only way
> to really do a head-to-head is to listen blind, or measure the complete
> systems under identical test environments. However, even then, there is
> no such thing as an exhaustive test.

Do you think that the SB2 could be improved by replacing (not as a mod,
for the SB3 for instance) the PCM1748 by a more expensive model from
Burr Brown? The PCM1748 is quite cheap, according to TI web site
(http://focus.ti.com/docs/prod/folders/print/pcm1748.html), so I guess
it's only a small part of the price of the SB2, even if you don't pay it
1.3$US ;-) (listed price for 1000 units). The PCM1792A
(http://focus.ti.com/docs/prod/folders/print/pcm1792a.html) has for
instance much better specs (on paper) than the PCM1748, but it costs 10
times more (less than 14$US for 1000 units). It is used in the Denon DVD
player DVD-A1XV, for instance. The PCM1794A seems also nice.

To say differently, according to your measurements, do you think that
going to this type of high end chips will be possible without a much
higher price and with a direct improvement, or do you think all the rest
of the SB2 should be improved to match the specs of the PCM1792A,
driving up the price (e.g., by including a PSU like Andrew's one)?

Fabrice

[Vinnie R.]

All,

FWIW, I recently began replacing the stock PCM1748E with the PCM1748KE. The KE version has 6dB improved Dynamic Range and SNR, as well as improved THD and channel separation specs. I now do this on all SB2s that I mod (except for those who do not want to use the analog output). I posted about it yesterday over here:
http://www.audiocircle.com/circles/viewtopic.php?t=21539

I'm not sure if there are any other dac chips that can be used in place of the 1748 w/o a board respin and/or SW changes. Sean?

Thanks,

[seanadams]

I haven't actually tested the KE - what I found was that the measured performance of the E consistently exceeded the specs for both the E and the KE. I'm guessing it's the same as CPUs, which can be marked with different speed grades to address different price points even if they all test the same.

As far as putting other DAC chips in there - it depends on the chip but generally no, it is not really practical - may as well make/buy an external DAC. A DAC with current outputs for example would probably require a completely different power supply for both the DAC and the IV conversion.

[Patrick Dixon]

I'm using the KE. I reckon I can hear the improved dynamic range, but it's probably not worth doing unless you improve the PSU and the clock first.

I don't actually think there's that much wrong with the PCM1748 - cheap or not!

[pfarrell]

On Wed, 2005-09-14 at 10:08 -0700, seanadams wrote:
> I haven't actually tested the KE - what I found was that the measured
> performance of the E consistently exceeded the specs for both the E and
> the KE. I'm guessing it's the same as CPUs, which can be marked with
> different speed grades to address different price points even if they
> all test the same.

Most (many? all?) semiconductor devices are mass produced and then
measured and labeled. By engineering standards, the specs have to be
limits, minimum gain, maximum noise, maximum departure from ideal, etc.

Of course, the manufacturers won't tell you for sure.

The idea is that you can design a circuit and get at least
the specs, maybe better. This lets the designer (Sean and others
in this case) design the end-to-end performance. There is no
guarantee at all that removing one part, and replacing it with
another that is labeled better, will actually deliver an
improvement.

I remember going through a bucket of transistors measuring the beta
(gain) and putting all below 100 in one bin, to be $0.25 each,
all over 100 in another bin to be $1.00 each, and all that measured
100 on the button in another bin for $10.00 each. It wasn't the
beta value that made them more valuable, it was that they
were all the same.

The common example of this is CPUs, where the clock rate
determines price, and makes a huge difference, but
all the parts start out the same.

All you can be sure of is that the KE version meets
higher specs than the E version, and costs about ten
times as much. Not that any given KE is ten times better
than any given E.


--
Pat
http://www.pfarrell.com/music/slimserver/slimsoftware.html

[Davey]

Pat,

Actually the "KE" version is just a few pennies more than the "E" version. $2.93 vice $2.70 at Digikey.
However, your point is a good one.

Regarding the devices and how they meet specifications....and how they are marked....and which hopper they came from....and whether the designs were slightly different....or whether they perform similarly in the larger circuit...etc. It all matters not when the evaluation method is subjective only. You can throw all the numbers or objective testing right out the window. :)

Cheers,

Davey.

[GreenMan]

is that it was built with audiophiles in mind. The Burr-Brown is 'good enough' to get by on. The folks at SlimD gave us something workable with the ability to port output to whatever DAC we want if the internal DAC falls short in our system. I suppose that is an argument against doing much more with the DAC if it keeps the price of the mechanism's function to give an intuitive and sturdy presentation of internet radio and hard drive music. I strongly sense that this device and ones like it are the future of hi fidelity stereo. The SB is on a compatible track for the MP3 masses who don't do hi fi. It is nice that it works as well for them as it does for audiophiles.