WRT the dig side of KD:
You could redesign it with strict EE practices, using multiple bypass ceramic caps + low esr electros, etc and end up with much worse sound.
The chip itself internally is not like modern Cmos IC's and so needs to be treated a little differently. I've been looking at the whole dig/DAC board now for a while
and how to get the absolute most out of it. There can certainly be improvements - or so we might think!
WRT power supplies, every body on this or that forum throws fancy adjectives around to describe the results of their latest power supply
be it shunt regs, ultra low noise super regs, simple series pass regs etc etc. So what -is- the optimum requirement for regs applied to a DAC? Low Z, wide BW, low noise,
shunt operation, series operation? I have a pretty good idea which direction to head based on experience but just following the latest and greatest fad that
is sweeping the net on DAC power supplies is not the answer.
Here's an example, years ago I was involved in designing an amp that was pretty esoteric. It had series pass regs for the front end HT supply. One night I bypassed
the regs altogether and just left raw RC filtering. Result - much better and transparent sound - in that particular application. So much for regulators!
Experience has shown that slow evolution based on careful trial and error with long term listening in a system that is of the highest quality is the way forward.
Experience has also shown that traditional EE solutions don't always translate into better sound. There have been many times when Steven has tried design
changes that I thought were technically not the best approach which translated into better sound.
cheers
T
Your points are well taken and one can get consumed by chasing the latest and greatest fad or trend and go backwards. I do not think that there would be any disagreement among most that the venerable TDA1541A chip benefits from better power supplies than the rather simple and antiquated 30 year old designs around when the chip was introduced. The question I suppose is what are the best imporvements that can be made.
As I stated above I have been looking at the KD dig board and where to go from here.
There are many options WRT discrete regulators, types of bypassing, caps used for bypassing, layout, DEM reclocking, etc.
But the bottom line is, as I did try to explain, it's not so much what all the individual elements are, it's how they work together to get a truly superlative result.
Given that a lot of water has flowed under the bridge since this chip was first produced, there are many people who have spent years trying to optimize the implementation. These include but are not limited to Thorsten Loesch, Pedja Rogic and the rest of the ususual suspects. Their work and observations are in the public domain for all to see.
More or less. Thorsten is pretty active / outspoken on forums and has graciously given a wealth of info for people to draw on.
Having said that - you won't see any schematics or intricate details on his -real- cutting edge designs which are of course AMR - and for god reason.
They al stress the importance of optmization of the digital power supplies and they seem to generally advocate the use of tight regulation, often using shunt regulators. Offhand, I am not aware of anyone now suggesting to go back 30 years with a simple power supply design. To the contrary.
I didn't suggest going back 30 years. My example of reg versus non regulated HT for power amp was just one case to show that things don't always
work the way we think they will and it's always good to think outside the box.
Again, in a general sense, I agree with your point of view. However, with respect to the power supplies for the TDA1541A dac, the overwhelming if not unanimous weight of opinion, learned from decades of research and experimentation, is to the contrary. Also, my limited experience messing about with this chip bears out the observation that these improvements reap worthy benefits.
Please go back and -read- my post again. I'm not suggesting that the KD doesn't require good power supplies - it already has them. Separate regulation
for all circuit elements and the noise is already low. Low noise shunt regs were tried KD's years ago but Stevens best results did not come as a direct result
of them or any other 1 element in isolation just the combination of elements.