SDTrans 384?

[nige2000]

its easier and better to use a picture host like flickr etc and post the image link (in between these brackets without spaces)

[gstew]

i cant see it been worth the hassle
and still need lots of capacitance
battery is the lazy mans super regulation and super capacitor only better

No caps according to their writeup here:

http://www.diyaudio.com/forums/digital- ... ost3580312

"So, after listening to various tracks in standard configuration (battery power) Bunpei set up the "solar power supply" that really looked like a little bright lamp. I would say about 30 cm tall, and 20 cm in each side direction. There were two solar panels facing each other, and I counted four "Krypton Light Bulbs" shining inside. Bunpei placed two CD-discs with the reflective side pointed inwards towards the light bulbs both at the top and at the bottom. He explained to me that the solar panels are natively 20V, but they have been "hacked" to deliver 10V which is then fed into a shunt regulator that brings the voltage down to 5V. Bunpei explained me that most of the energy is lost as heat; both from the lamps and from the regulator (I confirmed that everything was damned hot). He also explained me other things like that there is no capacitor whatsoever (and of course no battery) in this system. He said that the efficiency is only one (1) %. I think there was a bunch of other information as well, but I pretty much understood that this is "the purest of the pure" way to get a completely isolated 5V power source to your source component (SDTrans etc)."

why make it more complicated

LOL. Has that ever stopped you? It never stopped me!

Greg in Mississippi

[nige2000]

i cant see it been worth the hassle
and still need lots of capacitance
battery is the lazy mans super regulation and super capacitor only better

No caps according to their writeup here:

http://www.diyaudio.com/forums/digital- ... ost3580312

"So, after listening to various tracks in standard configuration (battery power) Bunpei set up the "solar power supply" that really looked like a little bright lamp. I would say about 30 cm tall, and 20 cm in each side direction. There were two solar panels facing each other, and I counted four "Krypton Light Bulbs" shining inside. Bunpei placed two CD-discs with the reflective side pointed inwards towards the light bulbs both at the top and at the bottom. He explained to me that the solar panels are natively 20V, but they have been "hacked" to deliver 10V which is then fed into a shunt regulator that brings the voltage down to 5V. Bunpei explained me that most of the energy is lost as heat; both from the lamps and from the regulator (I confirmed that everything was damned hot). He also explained me other things like that there is no capacitor whatsoever (and of course no battery) in this system. He said that the efficiency is only one (1) %. I think there was a bunch of other information as well, but I pretty much understood that this is "the purest of the pure" way to get a completely isolated 5V power source to your source component (SDTrans etc)."

why make it more complicated

LOL. Has that ever stopped you? It never stopped me!

Greg in Mississippi

yes i will admit we go too far to get little sometimes but its all part of learning

what type of battery did he test because they are not made equal
such as if i use my sla battery vs lifepo4 it is easily noticeably worse for any use i put it in

any way thats a very large footprint for a 5v supply

[gstew]

yes i will admit we go too far to get little sometimes but its all part of learning

I would never call it too far... just farther than many will go in search of the ultimate.

When I play records, they have generally been cleaned on a record-cleaning machine with a special regime of several fluids. Then I adjust the stylus rake angle to what I've found is right for that record's label. Then I change phase (in/out) if needed.

You've done motherboard mods. I have parts here for motherboard mods, but haven't done it yet.

You power your computer source and DAC (at least) with batteries. I have massive linear supplies on mine, as shown for the DAC & a linear ATX I built for my computer.

Why? Because we know it makes a difference. And of course, because we can!

what type of battery did he test because they are not made equal
such as if i use my sla battery vs lifepo4 it is easily noticeably worse for any use i put it in

Dunno, worth a question there. I can post one if you don't care to do so.

any way thats a very large footprint for a 5v supply

I've been looking at putting in a solar panel/battery setup to take my audio setups off the grid. PS Audio P10s are very good, but still don't eliminate time-of-day variations in SQ.

I won't be doing that this year, maybe year after next if things stay good. That will DWARF that setup!

Greg in Mississippi

[rickmcinnis]

Greg,

I can assure you the battery scheme of jkeny and Nigel will embarrass your linear supplies.

Consider the battery (A123 26650) a near perfect capacitor and use it as a bypass capacitor. Mount the battery as close as you can to the power input. You will have to use a switch or relay to turn it off, which is the only downside and not much of one at that.

This scheme greatly simplifies powering of digital stuff and works better than anything else you can come up with. I think regulator performance has meaning but not as much as one may think when using this battery.

I am most impressed with your mounting of uFl sockets. I have bought some of the sockets but could not think of a good way to securely mount them. I had figured I would get male and female cables and split them soldering one end to each respective board. I hate connectors in general but those seem to work well.

So, what do you think of the sonics of the SDTrans?

Edit: thought I had seen a female uFL cable end but I guess I haven't - can't find them now.

[gstew]

Hey Rick! Long time! I trust you are well. And again thanks for the phono tips over in the Soekris DAC thread, I've gotten some of those monster rectifiers to use with my Salas Folded and still would love to see pix of your modified ET arm someday.

Greg,

I can assure you the battery scheme of jkeny and Nigel will embarrass your linear supplies.

Consider the battery (A123 26650) a near perfect capacitor and use it as a bypass capacitor. Mount the battery as close as you can to the power input. You will have to use a switch or relay to turn it off, which is the only downside and not much of one at that.

This scheme greatly simplifies powering of digital stuff and works better than anything else you can come up with. I think regulator performance has meaning but not as much as one may think when using this battery.

I'm resisting battery power mightily. AND I have a number of experiments I plan to try with that digital test bed setup, a bunch of different DACs and sources AND eventually A123 cells on my DAM DAC and the other setups.

My intention on the battery trial is to get the AC-powered setup working as well as I can, then put in the batts on the DAM using Nige's, Randy's, & your setups as a guide. But that's probably out much later this year.

I am most impressed with your mounting of uFl sockets. I have bought some of the sockets but could not think of a good way to securely mount them. I had figured I would get male and female cables and split them soldering one end to each respective board. I hate connectors in general but those seem to work well.

Thx. That was a real pain to do, but I am good with small-component soldering. I soldered them onto the protruding stubs of the I2Ss output connector on the bottom of the board. I used about 8 sockets to finally get 3 good mounts, but I can probably do 1 for 1 now. Then after I got them mounted and checked the continuity, I put a bead of thick CA glue (Crazy Glue) across the back of the sockets and then dusted it with baking soda, which causes it to harden almost immediately and the baking soda becomes a binder that makes it very strong. I can remove it with a good CA-glue remover, but short of that they are on there good now.

If you find cables with female ends, let me know. I found some on Amazon for too much $, the same ones on EBay for more along with some from China that were cheap, but I didn't want to wait. And I found none on Digikey or Mouser. So I looked closely at the pictures of the higher-$ ones (about $10 USD for a 3" male-to-female cable) on Amazon and EBay, saw they just soldered a socket onto a raw cable end, and then made my own female-ended cables copying what they did.

So, what do you think of the sonics of the SDTrans?

After just one listening session, I was very impressed. Not broken in, not really warmed up, with a stock DAM DAC with no mods, original firmware and filters, it was good, sounding very analog-ish with some detailing I'd never heard from my turntable setup (which I know is not anywhere as good as yours). It still needs a lot of updating and modifications, but was very promising. After I get through the various testings and trials I've got planned, I'm strongly considering making my reference DAC using 2 DAM DAC cards in dual-mono with all the mods and the main source the SDTrans.

Later!

Greg in Mississippi

[jkeny]

yes i will admit we go too far to get little sometimes but its all part of learning

what type of battery did he test because they are not made equal
such as if i use my sla battery vs lifepo4 it is easily noticeably worse for any use i put it in

any way thats a very large footprint for a 5v supply

That was my first question & I went to the post - it looks like he's using ordinary alkaline batteries - at least there's no mention of LiFePO4 batteries that I can see
"(normally powered by 3 or 4 x 1.5V D-cells) and a Fidelix Caprice DAC, that also has an optional Chiaki made receiver and clock board built in that is fed by 3 x AA style batteries."

[gstew]

Thanks for that. Yes, I seemed you have restored my memory about it's clocking which at the time I wasn't impressed by. In my experience (although I haven't used these Silabs clocks) - any clock which is variable is in theory a higher jitter than a fixed frequency crystal based clock. These Silabs clocks have a fixed frequency clock which is then processed through an internal DSPLL in order to achieve the clock speed determined by it's I2C signals coming from the FPGA. Any feedback loop (PLL) which is used in clock handling introduces jitter. On the plus side these clocks also have a differential output clk+ & clk- which is a good design.

Thanks for the info on how those SiLabs units work. I haven't hunted down the datasheets. That they have an internal DSPLL goes a long way towards explaining the additional jitter of these units. I DO wish Soren had used the higher-speck Si570.

Well, again, from experience the Sabre DACs sound best when operated synchronously (this turns off some input processes which are detrimental to sound) but any DAC which accepts Bit clock, LRCLK & data (& optionally MCLK) I suggest you try reclocking these signals just before they enter the DAC chip - you will find this improves the sound. I haven't heard the AK4490 but I like the sound of the PCM5102 or 5122 with low latency filter

I've heard the same from others about both the ESS ASRC versus sync mode and the PCM5102/5122 (and also the 5142).

My main driver for the digital test-bed setup in my pictures is to try various powering, clocking, sourcing, and distribution configurations and get a feel for some of the sonic differences. One of these is no-reclocking versus ASRC (ESS) versus synchronous reclocking (Acko's S03) versus asynchronous reclocking (Ian's FIFO & Soren's DAM). My gut-feel is that the last two will be the best AND I suspect about equivalent.

So I have a couple of the EUVL ES9022 DAC cards, one of the Curryman ES9023 DAC cards, and a couple of BuffIIs. They'll either be fed from an Acko S03 or Ian FIFO. I'll connect each at first using their onboard clock and ASRC, then bypass that and feed the clock from the reclockers. On clocks I'll got some NDKs (22/24, 45/49, & 90/98), some Crysteks (45/49), and Ian's Si570 to swap between.

First two test-bed setups are two DAM DACs, the one in the pix above and another to feed from directly from an R-Pi. Of the next two, one will have a BBB feeding an Acko S03 and the other an R-Pi feeding an Ian FIFO. I'll have one of the ES902x cards in one and a BuffII in the other, and in addition to the clocking, will try some powering variations.

I will also try the SDTrans384 with synchronous-mode DACs. Way-low on my list is to try it with an Acko S03 with the clock being fed back from the S03 to the SDTrans, but I don't think adding isolation and the clocks really close to the DAC will matter as much on the SDTrans and it MAY be better to stick with their good NDKs and not take the hit of the isolators AS LONG AS the I2S connections are good and short... I consider the 6" ones I have on the SDTrans -> DAM to be WAY TOO LONG and most of these test setups will use 2" or 3" u.fl cables (I2S cable length comment just added).

Yes, I seem to remember Ian recommending the LifePo4 & wonder if he tried this PS on his Silabs clocks how they then compare to the Pulsar? Yes the expense is prohibitive, I find, particularly when 2 are needed.

I remember Ian liking the LiFePo4's on the Si570, but still preferring the Crysteks to the SiLabs. But his impression of the Pulsars was that they were in another world. Based on his comments, I don't think even great power will get an SiLabs to the level of the Pulsars or equivalent.

Greg in Mississippi

[randytsuch]

That was my first question & I went to the post - it looks like he's using ordinary alkaline batteries - at least there's no mention of LiFePO4 batteries that I can see
"(normally powered by 3 or 4 x 1.5V D-cells) and a Fidelix Caprice DAC, that also has an optional Chiaki made receiver and clock board built in that is fed by 3 x AA style batteries."

Random Randy rant :)
People often use the terms battery and linear power supply, kind of implying they are all equal.

They're not, in either case. For batteries, lifepo4's rock, and for LPS's, there are too many variables to mention.

Greg
A lifepo4 implementation is really pretty simple. For the soekris, you really just need 3.3V, which is just one cell, and a simple linear to keep it charged, and some kind of switching method.

After I discovered lifepo4's (thanks to Nige and John and others here), I use very simple linear chargers using walwarts (I know they are not audiophile at all) and simple regulators. I do use the old style walwart, with a transformer inside, and stay away from the new, cheap, solid state type walwart.

Much simpler, and cheaper than a good linear supply.

Randy

[gstew]

Random Randy rant :)
People often use the terms battery and linear power supply, kind of implying they are all equal.

They're not, in either case. For batteries, lifepo4's rock, and for LPS's, there are too many variables to mention.

So, so true!!!

Greg
A lifepo4 implementation is really pretty simple. For the soekris, you really just need 3.3V, which is just one cell, and a simple linear to keep it charged, and some kind of switching method.

After I discovered lifepo4's (thanks to Nige and John and others here), I use very simple linear chargers using walwarts (I know they are not audiophile at all) and simple regulators. I do use the old style walwart, with a transformer inside, and stay away from the new, cheap, solid state type walwart.

Much simpler, and cheaper than a good linear supply.

Randy

Randy,

Thanks for the information... and also very much thanks for your blog posting on your DAM setup, it is on my main DAM resource bookmark link.

I did think you needed 3 LiFePo4 cells to do the DAM justice, one for the logic and clock and one for each side of the ladder (after removing the regs and the opamp references). Correct? In my setup, I'll also likely need some sort of output stage due to not having an active preamp, but long-term that will likely be tube, so LiFePo4 won't be a good and practical option.

Honestly, I started planning this set of experiments well over a year ago (I know I bought some of the transformers in the fall of 2014) and before I'd seen the float charging setups Nige pioneered and that you and Rick use. That has wildly simplified a LiFePo4 powered-setup and we should all be very grateful for that.

OTOH, I DO want to try and get my linear supply setup as good as possible before trying LiFePo4, that way I know what I'm missing if I decide to stick with linear or know that I couldn't bear NOT to go to LiFePo4.

I'm hopeful I can get them very close... and comments from you and Nige and Rick have me doubt that.

Greg in Mississippi

P.S. My main reason for planning out this set of experiments was initially driven by a desire to really find out how well an isolated and reclocked input to a DAC was immune from the effects of upstream sources (Of course, comments here, and from others I respect such a John Swenson and Ted Smith, the designer of the PS Audio DirectStream DAC, now strongly suggest that hasn't been achieved yet).

My second reason was to hear the sonic differences between different clocking power supply schemes after Ian's and others comments on the Ian's FIFO thread (and a bunch of people poo-pooing some of those comments).

Since inception, the scope has grown immensely as new options become available (such as being able to get an SDTrans384 for about $350USD, they were originally about 2x that... and of course the DAM DAC). My goal now is to get down to 2-4 setups that I find work really well, build up the best 1-2 as reference DACs, keep the others for backups, and then have a graveyard of abandoned DAC projects (already in-work).