[ergo22]

I'm eagerly waiting for the release of this product. How are the thing rolling with this project

There is a kind of competition out also
http://www.kandkaudio.com/digitalaudio.html

at least cost wise as I hope the pricing of DAKSA will be about the same.

Regards,
Ergo

[Tinker]

Hi ergo22,
   in answer to your first qusestion, Aspen still has every intention of having making the just-in-time-for-Christmas target date. Most of the work now centres around two issues: Firstly, whether a really-amazing-on-paper jitter reduction idea is both measurably, audibly and emotionally (for want of a better word) than other already pretty good (and slightly cheaper) technology we are trialing. Secondly, getting the synergy of resistors, transistors and capacitors for the analogue section absolutely spot on.


For the second question, how does it compare to this other DAC for cost? Well, it appears that it is in the same price bracket, although Aspen's pricing structure is different. If I understand this other product correctly there are three or so stages/option that represent a 1) basic function board, 2) jitter attenuation circuit and 3) an upgraded output stage.
It seems you also need to buy a separate power supply, transformers and output stage to get started.

The DAKSA entry level seems more more expensive because a different design philosophy has been adopted and the level 1 kit already comes with with Aspen's ultimate output stage. We did not see the point in separately packaging something that everyone will want. Power cleanliness, accuracy and speed are also crucial for the DAC design. The entry level board also comes with five high-grade power supplies built-in for low inductance, whcih is critical for digital design. And the transformers are all included too. Like the GK-1, there is some room to tweak resistors and caps as suits taste as this is one of the things AKSA owners love. Level 2 adds some functionality for a few bucks and level 3 is the jitter reduction stage. So in short, the products have similar technical aims at a similar price, however, the division of cost and finer points of design philosophy appear quite different.
Short answer, that other DAC is entry level (basic supply and passive output stage) for $597 and "fully blown" will run $1371. Based on current parts and board size the DAKSA entry level (inc psu and active output) is looking to be around $650-700 and goes to $950-$1050 with all options. NOTE this is subject to change and could go up or down.


I need to reiterate that development involves critical listening and experimentation to create the final product. We have a number of breadboard designs that work, but we don't want a design that just works. Aspen is committed to a particular philosophy which Hugh has expounded a number of times on this forum. Much time (more than expected, and this has slowed things down a little) has gone in to extensive listening tests and revision to the 55N+ and lately the 100N+ which was just that little bit trickier. Also the AKCentre speaker for home theatre has been an enexpected concern this year. All these are competing interests which Hugh insists are perfect before release. DAKSA is on its way, but we have no intention of sacrificing quality to the builder by neglecting that extra part substitution or listening test.

Hope this long-winded response was of some help.
Cheers,
    T.

[Jens]

Thank you, Ben! Most elucidating reply!

I think a lot of us AKSAphiles are sitting on our hands waiting for the fully-fledged DAKSA (I know I am   )

I've have also had questions about it here at AKSA Scandinavia, and I've told people that when Hugh develops new products it may take time, but it's always worth waiting for  

So just keep up the good work! Meanwhile, the rest of us will to try to keep our impatience in check  

Cheers,

Jens

[ergo22]

Yes, thank you Tinker.

I'm at the moment tweaking/upgrading Behringer DCX2496 digital crossover. It uses a CS8420 receiver/asyncronous resampler in hardware resampling/reclocking mode. So it uses a fixed oscillator to clock the data out of the chip. I have successfully upgraded this clock to TentLabs XO2. I have also tweaked the spdif interface. It uses a video opamp as a buffer in transport end -> 75ohm BNC -> 75 ohm cable -> straight to Behringer PCB, no connector -> Schott digital transformer -> 75 ohm terminating resistor -> CS8420. The transport is Sony player also with TentLabs XO2 clock.

Now the basic theory about this solution is that it should be almost immune to jitter in spdif. I have made a listening test comparing with other player (low end technics) and the difference was huge! A friend and me were absolutely astonished. My transport had a much deeper soundstage with much more air.

Have you already experimented with DAKSA prototype. Especially the level 3. Is it really immune to transport or is there still some difference and the transport quality still matters Also would it basicly be possible to implement a setup where a clock is feed back to transport with DAKSA. That seems still to be an ultimate solution.

I know with such product the priority is a good performance with SPDIF, I'm just interested about the possibilities there.

Regards,
Ergo

[Tinker]

I'm at the moment tweaking/upgrading Behringer DCX2496 digital crossover. It uses a CS8420 receiver/asyncronous resampler in hardware resampling/reclocking mode. So it uses a fixed oscillator to clock the data out of the chip.

Nice unit. I have one too that Aspen uses for speaker testing and they are a lot of fun for so few $$$.

The 8420 and a similar device made by AD were contenders for the front end of the DAKSA. A good concept that has been used sucessfully in a number of high end devices. But I digress...

Is it really immune to transport or is there still some difference and the transport quality still matters Also would it basicly be possible to implement a setup where a clock is feed back to transport with DAKSA. That seems still to be an ultimate solution.

OK, this is part of that "testing process" I mentioned. The design we are currently testing and developing does not use resampling, but rather "synchrotonic" reclocking using a predictive clock correction. On paper this reduces the transport contribution below the resolution of the clock, close to the 3pS jitter that the best non-millitary oscillators provide. The question we are tackling now is measuring the empirical influence of transports (ie we have to create artificially dirty transports) and whether or not the reclocking algorithm is unconditionally stable.  Measuring jitter is really hard, not in the least because there is no firm standard, and mostly because different kinds of jitter pop up in different places and produce different sonic artifacts.


As for feeding the clock back in to the transport a-la the Linn Numerik system, the aim of the DAKSA is to avoid this sort of dependence. Different transports use different oscillators ranging in speed from a few MHz to 20+ MHz depending on design, so mating some DAC to an unknown transport requires the owner to do some serious mods. The DAKSA reclocker is designed to be an ultimate solution that doesn't require player modifications.

Cheers,
   T.

[tinears02]

Hi Ginger
Thank you for the info. I am also waiting for the DAKSA, and
am considering it vs the DAC1 from Benchmark Media (used
by some well known mastering engineers), Norm Tracy's
DAC in Audio Crafter's Guild and the K&K dac. I would think
all of these are worthy contenders and cost about the same.

The DAKSA's goal to eliminate the clock feedback line to the
transport is certainly attractive. My experience with
cd players are that the transport is always the first thing
that goes, whilst every other block still functions nicely, and
this includes some well-known names in the business.
I'd like to use a reasonably good cd-rom transport
to feed the daksa, with proper impedance terminations
of the spdif signal coming out of it.

While you're evaluating jitter and transport immunity, may I
also ask if you are evaluating with the following
i. poor (ie unmatched) interconnect
terminations
ii. high phase noise clocks in the transport?

Thanks for listening.

[Tinker]

i. poor (ie unmatched) interconnect
terminations
ii. high phase noise clocks in the transport?

i) Yes, although the main effect of this is to introduce clock skew, which should get ironed out by the reclocking provided no data is lost, and SPDIF is designed around data integrity rather than clock accuracy. Anyway, a 75R cable is a cheap upgrade, an absolute b*ll-tearer can be made for about AU$30 or so.

ii) Yes, see previous post. The problem is custom making poor transports.    Hi phase noise is easy to create thermally, or with low Q oscillators. The issue I am really interested in is transient sagging and jitter that has periodic and signal-correlated components as these sound the worst. Modulating the clock isn't too hard, but the others...

Analysis suggests that reclocking and PLLs are most sucecptable to periodic jitter as they spend their lives chasing it around and possibly intermodulating. Getting a really long time constant is the key here so that these alterations become at worst subsonic.

Cheers,
    T.