Since DS2 will be a standalone Dac w/o any other function, it would not display album art or tract number. Those will be part of the streamer/transport/renderer/endpoint (pick your terminology).
It might display the file format tho.
I wasn’t being clear. I was referring to the PST.
1 Like
That I agree with you 
Let’s dig into the power trees a little more:
The DS Mk II shares the same power transformer as the TSS. The transformer has 4 windings speced at 14.5V @ 0.8A each. I use two windings in series for the digital card and two in series for the analog card. Both the analog and digital cards have discrete Schottky diode bridge rectifiers to minimize rectifier switching noise.
Next is a power factor correction circuit. This spreads out the current draw over the incoming bumpy rectified AC voltage waveform. Instead gulping lots of current near the peaks of the incoming voltage, smaller caps that are charged in series over much of the incoming waveform slurp current more evenly which both causes less noise in the rest of the system and lowers the incoming noise near the peaks of the incoming voltage waveforms. The caps are discharged in parallel which, in effect, lowers the voltage and raises the current for the rest of the box.
Next is something that might be little unexpected: a low noise switching voltage regulator. Much of the noise from switching power supplies is caused by a high current loop which is switched on and off very rapidly. It’s the change in current combined with the loop area that causes the notorious EMI from switchers. Linear Tech (now Analog) has a series of switching regulators which have capacitors in the chip’s package which make that high current loop much smaller in area. That lowers the noise considerably. They also do other things like controlling edge rates when switching to lower those changes in current. I also keep the switcher on the other side of the board from everything else. There’s a lot of board area on purpose to provide that isolation. To keep any residual noise from interfering with the audio the switching regulators run phase locked with the main oscillator. i.e. they don’t cause noise that shows up in the audio as an interference pattern between the switcher’s switching frequency and any other clocking.
The benefits of the switching regulators are both related to efficiency and to being less sensitive to anomalies/irregularities in the incoming power. With linear power supplies you need to be able to get all the current you want when the incoming AC voltage is on the low side, say at -20% of nominal, but you also need to dissipate as heat the extra power when the AC voltage is on the high side, say +20%. At nominal voltage you are throwing away 20% of the incoming power right off the top and more when the voltage is higher. A switching supply runs at a constant, say, 95% efficiency in the same situations. Spikes or dips in the incoming voltage don’t affect switchers, they continually take very small bites of current when they need them and the actual incoming voltage really doesn’t matter, they just take the current they need to provide the requested power.
Most of the “evils” of the switching regulator are rehabilitated by having very good wide band ripple rejection with the LT3045, one of the best regulators available. On the digital board the switcher is set to 5.47V and the LT3045 regulates that to 5V. On the analog board there are two switching regulators and two top level LT3045’s which set to 13.87V/12V and 7.12V/5V. Note that the voltage drops on the analog card are closer to 2V which is the sweet spot for the best voltage regulation with the LT3045’s.
Closer to where the current is needed, the voltage is regulated again by multiple LT3045’s. i.e. from 12V down to 10V on each analog channel right at the video opamps on the outputs and two more times from 5V to 3.3V for the master clock and also for the final reclocker, More economical regulators are also used to lower from 5V to 3.3V for other subsystems like the CMOS and the configuration relays.
Each of the LT3045 regulators has three quality, oversized, low ESR aluminum polymer capacitors for each of the input, the output and the voltage reference.
At the clock, the reclocker and the analog outputs there’s no sign of noise from the input switchers to the limits of my scope. (Tho I can see it if the scope leads lay across the switchers.) I.e. there’s no sign of RFI affecting the output transformers or of EMI getting thru the multiple layers of quality regulation.
Quiet regulators aren’t cheap, but its money well spent to have quality power everywhere.
15 Likes
For bigger drops with fairly well-known current draws I use a series resistor (with a cap for free filtering) on the inputs to the regulators. There’s no place I need parallel LT3045’s, the initial regulators (the switchers) and the FPGA regulators are the only place I need higher currents.
I use multiple stages of regulation for regulation more than heat spreading, so I design the drops to be as big as the current needed allows but still keep the heat under control at each level.
If heat wasn’t important, I’d consider using paralleled linear regulators on the inputs of the analog board. But they cost more and have to be capable of dissipating a lot more heat in a small area.
6 Likes
This is an example, tho not on the audio path:
The 0.10 resistor is more to dampen any resonance between the ferrite bead and the input caps, but it does add a little low pass filtering.
The 2.7V isn’t used anywhere but here, it splits the difference between 3.3V and 1.8V. 1.8V is too close to 1.0V for this regulator so I used the 2.7V instead of 1.8V to feed the 1.0V regulator.
In a previous version there was a (pair of) resistors between each level. It all depends on how much heat the package can transfer to the PCB.
6 Likes
Hi Ted, would you mind going into detail on the engineering efforts seeking vibration control on the DSII versus the DSI, or other perfect wave products for that matter?
Other companies go through lengths with high mass, double boards, dampening feet, etc. Not to speak of after-market products and tweaks customers can perform themselves. Just in case: where would these benefit the current design even more?
I’d like to learn from your insights and approach.
Thanks!
The first thing is to avoid things that are sensitive to vibration: tall caps, non class 1 MLCC (multilayer ceramic capacitors), wires to connectors, loose coils, etc. Having boards be all surface mount helps a lot with microphonics. In a handwaving sense a capacitor is a condenser microphone, you need to keep them short, solid and surface mounted rather that large, tall or thru hole (standing on legs which let things vibrate.)
You need to control vibration at the board level, we will be using 2 oz copper (rather than the standard 1/2 or 1 oz) for all of the copper layers. A stiffer board vibrates at a lower frequency, we want to keep any vibration that does happen away from the frequencies where the ear is most sensitive. I also have more mounting screws than many. I’m trying to keep the audio transformers as firmly mounted as possible, etc. There’s a balance, more solid mounting posts are good for stiffness, but shorter distances between mounting holes raises the frequency of any vibrations that do happen. You can rap a board with your knuckles to get an idea of the resonant frequencies around the board.
Having all of the connectors board mounted instead of chassis mounted with wires keeps things much stiffer. (Especially with the DS Mk II because all connectors are galvanically isolated and hence don’t have solid metal connections to the case.)
Having a case with a solid base to be as inert as possible is good. There’s some truth in having unit that’s solid sounding when you rap your knuckles on it.
In general I think about vibration resistance with parts choices, routing choices, mounting choices, etc.
On the other hand, the more revealing your system is the easier it is to hear various tweaks, so even tho the Mk II will be more inert than the Mk I, all in all you might hear differences between different vibration dampening tweaks better.
19 Likes
Reading this, it must be much more effective to damp the PCB internally or parts on the board, than to put fancy feet under a chassis.
1 Like
Some amps use ceramic circuit boards to help with cooling and vibration.
DS II on paper is so much superior to DS, I wonder who will keep their DS; at least for unmodified ones like I have.
Is there a Beta test list for DS II? If there is I would like to be added into it now! 
1 Like
If it’s really going to be available end of September-ish, beta is probably over.
If it’s available September-ish and if you get one you are the beta tester!!! Lol
3 Likes
If a beta tester is needed overseas in EU area, I’m ready, count on me!
1 Like
It is not the first time Paul has made optimistic projection. I bet no one has a unit yet, so I still have hope to be on the list. I just have no idea how they select testers. Do we have to beg?
However, a Beta tester from Europe has a less than slim chance 
A second tester for Europe would be more prudent.
2 Likes
Probably, at PSA they’re picking their favourites from the forum and first have them sign a non disclosure agreement, before shipping a beta product.
Paul may prove me wrong, here. 
1 Like
Yes, I suspect PS Audio probably won’t send beta units out of America for testing. But @paul and @jamesh, if you’d like feedback on how the beta versions of the DSII or AirLens work on 240 volts, I’m your guy!! 
While we like to tease @Paul about his prior predictions, I will point out that he’s been much closer with the FR30, 600, and the Phono Pre.
2 Likes
Indeed. More than a fair bit has been done already (thanks Ted!). Over air, into the box seems eliminated, due to sturdiness and mass. Remaining vibrations would need to come by foot (depending on quality of placement).
This makes me thinking of what someone like Vmax would/could do with sorbothane at the feet and various internal board supports, below and above. One would probably need small rings as suspension, and if successful, become Lord.
1 Like