elwappo99 said:

   
Probably shouldn't. Looks like the U12 doesn't use the power from the USB connector, just the Data, so cleaning up the power signal won't matter.

Click to expand...

I think it does - it uses the +5V to switch a power relay if nothing else...

Wow! The second U12 arrived, I installed in my main system. The change in sound quality is not subtile. A huge leap, right off the bass is a half octave lower and very well defined. Really thunderous now, I'm sorted shocked. I know this component is important, but never realized how important. Still trying to get my mind around the exact why. Is it a reduction in jitter? PC noise filtering? The sound is cleaner, much greater clarity and transparency. Still the sound is warm, with excellent rich tone. It's like some unheard distortion has been removed. And my system was sounding really good before.

What an awesome audio bargain. I'll upgrade the caps on this one tomorrow and start looking for an upgraded power chord.



Edit typos

rb2013 said:

start looking for an upgraded power chord.
 

Click to expand...

Heh, did you get the 90-degree cable also that angles out over the back of the unit?  Not exactly a stroke of genius there.
Still, for good cheap power cables I've been happy with PS Audio Jewel 1's... Music Direct has them for $35/$40 depending on length (1m vs 1.5m).

sandab said:

Heh, did you get the 90-degree cable also that angles out over the back of the unit?  Not exactly a stroke of genius there.
Still, for good cheap power cables I've been happy with PS Audio Jewel 1's... Music Direct has them for $35/$40 depending on length (1m vs 1.5m).

Click to expand...

Actually it shipped with wrong cable, not a problem as I have a few around. I put Custom Cable Power Cord pc on the other, made a slight improvement. I like Shunyata Diamondback for a well made, good sounding pc. They're around $59 on ebay. On my main system I use Synergistic Research X2 ref but they're too expensive for this purpose.

I just ordered a Shunyata

PS The Shunyata Venom are really good too. I use one of those in my other syatem.

rb2013 said:

Wow! The second U12 arrived, I installed in my main system. The change in sound quality is not subtile. A huge leap, right off the bass is a half octave lower and very well defined. Really thunderous now, I'm sorted shocked. I know this component is important, but never realized how important. Still trying to get my mind around the exact why. Is it a reduction in jitter? PC noise filtering? The sound is cleaner, much greater clarity and transparency. Still the sound is warm, with excellent rich tone. It's like some unheard distortion has been removed. And my system was sounding really good before.

What an awesome audio bargain. I'll upgrade the caps on this one tomorrow and start looking for an upgraded power chord.

Edit typos

Click to expand...

The independent power supply helps a lot because it is no longer dependent on the weak and noisy USB 5V supply, which can only supply up to 500mA of current, that can drop to 100mA in some cases.
 
If you are interested in a better power supply consider replacing the power transformer, I have a similar unit based on the previous generation XMOS and it came with the green Bingzi transformer. I wanted something that was UL rated and replaced it with an Amgis/Alfamag, the bonus was it ran much cooler and the Alfamag was about 10% heavier (more copper in the windings)
Outside of the US, Talema is easier to find and a little cheaper.
 
I also installed ultra low jitter crystal XOs replacing the "Precision 0.1ppm" ones that it came with.
The Gustard looks like it has separate regulators for the XOs so it is probably worthwhile to put in a low jitter unit.
 
The last part is to replace some of the caps with Panasonic OSCONs and Nichicon FP Polymer caps.
The main rectifier filters  I used the Chemicon KZN. The KZN has 2x lower ESR than the supposedly Panasonic PX caps
(I think the PX are fakes)
 
Pictures of my unit.

 

b0bb said:

The independent power supply helps a lot because it is no longer dependent on the weak and noisy USB 5V supply, which can only supply up to 500mA of current, that can drop to 100mA in some cases.

If you are interested in a better power supply consider replacing the power transformer, I have a similar unit based on the previous generation XMOS and it came with the green Bingzi transformer. I wanted something that was UL rated and replaced it with an Amgis/Alfamag, the bonus was it ran much cooler and the Alfamag was about 10% heavier (more copper in the windings)
Outside of the US, Talema is easier to find and a little cheaper.

I also installed ultra low jitter crystal XOs replacing the "Precision 0.1ppm" ones that it came with.
The Gustard looks like it has separate regulators for the XOs so it is probably worthwhile to put in a low jitter unit.,

The last part is to replace some of the caps with Panasonic OSCONs and Nichicon FP Polymer caps.
The main rectifier filters  I used the Chemicon KZN. The KZN has 2x lower ESR than the supposedly Panasonic PX caps
(I think the PX are fakes)

Pictures of my unit.

Click to expand...

Hey, thanks for the mod tips!

Well my M2Tech EVO ran a very clean $300 linear PS built from a Acopian Yellow Box. And my previous interface favorite had it's power completely detached from the PC and fed from a linear ps Aqvox.

The sound improvement with unit is blowing me away. I actually listened to CLASSICAL last night, the 1st time I've done that. Usually it's Alt, Emo or Rock. A Beethoven piano sonnata - the piano just hung in space. I mean it was in the freaking room! The notes were free of any glare at all, very hard to do on solo piano. Their overtones just rang perfectly. Radiating out into the room as 3d point sources, as would a real piano, truly spooky. The conveying of emotion absolutely moving. Now I have recently recapped my amp's film caps with Mundorf Supreme Silver/Gold/Oil. These have been in few a few weeks, and are now fully run in. But they were there before the Gustard. The clarty, transparency, bass extension and definition improvement is really outstanding.

I will buy one more to do further mods. The big blue tranny is different then green one you replaced? If you go to the Hifiduino website, the reviewerhas some criticisms of the XMOS implementation in the DIYinHK board. It has to do with the reliance on the internal clock of the XMOS chip. I'm wondering if the Gustard's implementation of clacking is different. Worth a read.


Edit typos

rb2013 said:

The big blue tranny is different then green one you replaced? If you go to the Hifiduino website, the reviewerhas some criticisms of the XMOS implementation in the DIYinHK board. It has to do with the reliance on the internal clock of the XMOS chip. I'm wondering if the Gustard's implementation of clacking is different. Worth a read.

Click to expand...

Minor differences, both are 15VA units. The blue one is 14Vac vs 15Vac for the green one.

My original intent was to get a transformer with the proper safety agency approvals (UL,VDE,CE etc) as it was turned on 24x7, it turned out to improve the sound, the bass was more extended and there was overall more definition to the music, it removed a lot of smearing on complex passages.

I would be interested in a link on the internal clock issue on the XMOS, I remember reading something on the Hifiduino site.

Thanks for posting the detailed pictures on the Gustard, it helped to answer a lot of questions I had on the unit.

b0bb said:

Minor differences, both are 15VA units. The blue one is 14Vac vs 15Vac for the green one.

My original intent was to get a transformer with the proper safety agency approvals (UL,VDE,CE etc) as it was turned on 24x7, it turned out to improve the sound, the bass was more extended and there was overall more definition to the music, it removed a lot of smearing on complex passages.

I would be interested in a link on the internal clock issue on the XMOS, I remember reading something on the Hifiduino site.

Thanks for posting the detailed pictures on the Gustard, it helped to answer a lot of questions I had on the unit.

Click to expand...

Thanks for the info - will have to try the transformer swap.
 
Here's the link  http://hifiduino.wordpress.com/category/usb-audio/
 
Regarding the DIYinHK XMOS board:
 
OSCILLATORS

According to DIYINHK, this version uses NDK NZ2520SD Ultralow phase noise oscillators from Japan. There are 3 oscillators: the 48 MHz for the digital circuitry of the XMOS device, the 45.158MHz and 49.152 MHz for the audio frequencies.







NDK has some general information on their oscillators, including jitter data here: [link]. Below is the phase noise plot of the NDK oscillators. Note the phase noise curve of the “SD” models.



And here is the phase noise plot of the CCHD-957. Note the phase noise value of the CCHD-957 at 10 Hz (-97) vs the phase noise value of the NDK oscillator at 10 HZ (-113)



 

MASTER CLOCK

The Master Clock is available straight from the NDK oscillators or at the appropriate scaled-down frequency through the output pins.

Straight from the oscillators (45.158MHz and 49.152 MHz) through a U.FL socket:



Through the output pins (22.5792Mhz/24.576Mhz).

The “C7424Z” device is used as a clock divider to generate the 22.xx and 24.xx frequencies used by XMOS device and also used by the downstream device (the DAC). Seems the clock line to the XMOS device is also isolated in order to prevent any kind of noise leakage from the XMOS/USB side to the Clock/Clean side..

 

I posted there this question and received this reply

On the XMOS clocking issue, I believe you described the clock derivation issue as:
“And according to this doc [link], the input clock (say the 24.576MHz clock) is sampled to the internal processor clock. Meaning that the external clock ticks (or transitions) are mapped to the processor clock ticks. This means that the worse “jitter” is when you just miss the external clock transition and must wait for the next internal clock tick. So the worse deviation from the actual frequency is the period of the processor frequency, thus for a 400MHz internal clock part, it is 2.5nS. This is the peak jitter. Thus the RMS number is 2.5/SQR(2)=700 psec””
Was this just a implementation of the XMOS on this particular DIYINHK board using the C7424Z chip, or is it implicit to designing around the XMOS? Sounds that way.
Although it was mentioned by HML that “The Xmos processor data sheets suggest that it may be possible to use the MCLK for the clock source for the BCLK and LRCLK. This is not done in the reference design and not in the DIYINHK board.”
I will post closeups of the Gustard U12 board when I get it.



 

 

1. 

   
   BlgGear
   ​
   2014/10/07 at 17:52
   Reply
   I don’t recall completely, but I think this is inherent in the XMOS chip.
   
   
   
   

Another potential source of jitter is the clock scalers themselves.  They're edge triggered with a hysteresis relative to Vss/GND.  If Vss or GND is noisy the trigger point will move around.  With switched (e.g. CMOS) logic rather than some current-mode logic there will be switching noise.  The switching noise is synced to the switching events, which are clock-based.  Mostly the main 48M logic clock (which I assume is internal multiplied into multiple phases), but also the audio clocks around the actual scaling logic.  Decoupling is difficult as caps are hard on solid-state, and the distance to an external cap is thousands of times further than from the switch noise sources to triggers - so the decoupling is mainly of external merit, not internal.
 
By the way, my go-to first stop for all things clock and PLL related is Floyd M. Gardner: Phaselock Techniques.  (http://www.amazon.com/Phaselock-Techniques-Floyd-M-Gardner/dp/0471430633/ref=sr_1_1?ie=UTF8&qid=1413748575&sr=8-1&keywords=phaselock+techniques).  It has about 60 pages on the effects of noise on PLLs, most of it discussing nonlinear effects (which in this context the author defines as gaussian distribution in, non-gaussian out).  It helps if you have an EE background (especially the mathematical end, although some of it is purely notational like "f(z) mod 2pi"), but much of it is quite accessible. If you're an EE like me you'll love this book.

Thanks, it is a good amount of info
 
Quote:

rb2013 said:

  On the XMOS clocking issue, I believe you described the clock derivation issue as:
“And according to this doc [link], the input clock (say the 24.576MHz clock) is sampled to the internal processor clock. Meaning that the external clock ticks (or transitions) are mapped to the processor clock ticks. This means that the worse “jitter” is when you just miss the external clock transition and must wait for the next internal clock tick. So the worse deviation from the actual frequency is the period of the processor frequency, thus for a 400MHz internal clock part, it is 2.5nS. This is the peak jitter. Thus the RMS number is 2.5/SQR(2)=700 psec””
Was this just a implementation of the XMOS on this particular DIYINHK board using the C7424Z chip, or is it implicit to designing around the XMOS? Sounds that way.
Although it was mentioned by HML that “The Xmos processor data sheets suggest that it may be possible to use the MCLK for the clock source for the BCLK and LRCLK. This is not done in the reference design and not in the DIYINHK board.”
I will post closeups of the Gustard U12 board when I get it.



 

 

1. 

   
   BlgGear
   ​
   2014/10/07 at 17:52
   Reply
   I don’t recall completely, but I think this is inherent in the XMOS chip.
   
   
   
   

Click to expand...

I am thinking this is a way to slave the processor clock to the master clock. XMOS has gone to the trouble of allowing switchable master clocks, I think the samples are clocked out using this master clock.
The processor has to keep the output buffers full and must do it within the hard sampling time limits, this slaving prevents the 2 clock domains from drifting too far apart.
 
Put another way, the processor clock is always within 2.5ns of the master clock edge, which is clever way of doing things. 
They just have to make sure the output buffer is big enough to ride out any latency in the software processing inside the core.
 
BCLK and LRCLK has to be divided down internally from the internal reference clock, the question in my mind is whether they are doing it using a hardware scaler preferably linked to MCLK or some sort of software based PLL. I hope it is the first method.
 
(If you see info to the contrary, I would like to take a look)

You always run a scaler using the clock you're going to scale.  Nothing else would make sense.
 
I doubt the Gustard U12 has a PLL, but the downstream DAC will.  It has to, or its frame block buffers would otherwise eventually overrun or underrun.

sandab said:

  You always run a scaler using the clock you're going to scale.  Nothing else would make sense.
 
I doubt the Gustard U12 has a PLL, but the downstream DAC will.  It has to, or its frame block buffers would otherwise eventually overrun or underrun.

Click to expand...

Agreed.
 
The XMOS on the Gustard does a superb job of isolating the USB data + clock from the SPDIF output.
 
I see this as the modern day equivalent of the SPDIF reclockers of days past.
 
One thought I had was to take the U12 board out incorporate the I2S interface straight into twin Buffalo 3 DACS, the HDMI out has more than enough drive capability.
(It about USD$80 more than an Amanero board).

sandab said:

  Another potential source of jitter is the clock scalers themselves.  They're edge triggered with a hysteresis relative to Vss/GND.  If Vss or GND is noisy the trigger point will move around.  With switched (e.g. CMOS) logic rather than some current-mode logic there will be switching noise.  The switching noise is synced to the switching events, which are clock-based.  Mostly the main 48M logic clock (which I assume is internal multiplied into multiple phases), but also the audio clocks around the actual scaling logic.  Decoupling is difficult as caps are hard on solid-state, and the distance to an external cap is thousands of times further than from the switch noise sources to triggers - so the decoupling is mainly of external merit, not internal.
 
By the way, my go-to first stop for all things clock and PLL related is Floyd M. Gardner: Phaselock Techniques.  (http://www.amazon.com/Phaselock-Techniques-Floyd-M-Gardner/dp/0471430633/ref=sr_1_1?ie=UTF8&qid=1413748575&sr=8-1&keywords=phaselock+techniques).  It has about 60 pages on the effects of noise on PLLs, most of it discussing nonlinear effects (which in this context the author defines as gaussian distribution in, non-gaussian out).  It helps if you have an EE background (especially the mathematical end, although some of it is purely notational like "f(z) mod 2pi"), but much of it is quite accessible. If you're an EE like me you'll love this book.

Click to expand...

Great detailed info thanks!  The reason all (as much as possible) ps noise or ripple has to suppressed and or filtered.  I think the excellent ps filtering and isloation of this device is an important reason it sounds so good.  
 
I don't see a “C7424Z” device or equivlent on the Gustard board unless I'm missing something.  Blg Gear asked about the ALTERA MAX II chip and it's function. It's a dsp I believe and my thinking was it does the DSD processing.  Any thoughts there?

b0bb said:

  I am thinking this is a way to slave the processor clock to the master clock. XMOS has gone to the trouble of allowing switchable master clocks, I think the samples are clocked out using this master clock.
The processor has to keep the output buffers full and must do it within the hard sampling time limits, this slaving prevents the 2 clock domains from drifting too far apart.
 
Put another way, the processor clock is always within 2.5ns of the master clock edge, which is clever way of doing things. 
They just have to make sure the output buffer is big enough to ride out any latency in the software processing inside the core.
 
BCLK and LRCLK has to be divided down internally from the internal reference clock, the question in my mind is whether they are doing it using a hardware scaler preferably linked to MCLK or some sort of software based PLL. I hope it is the first method.
 
(If you see info to the contrary, I would like to take a look)

Click to expand...

Well that would be an better implementation, which would avoid the syncing issues Blg Gear points out in the DIYinHK board

One of my Gustard U12s is feeding a heavily upgraded and modded true R2R DAC.  
 
I took a PCM 1704 Lite DAC60 and replaced many of the caps, especially the couplers - replaced with Mundorf Supreme's (and soon to further upgraded by Mundorf Supreme Silver in Oil caps).  I swapped a critical ps resistor with a Hammond choke, upgraded the Elna caps to ELna RFS Simlic II, and replaced an important pr of elcheapo electrolytic with Mundorf M'Lytic Dual Cap, HiFi tuning fuse, upgraded cryod silver fuse holder.  I have another level of mods to do before the project is complete - 4 more Elna RFS Silmics, 4 Nichicon Fine Gold to replace the Sanyo ps filter caps, Vampire all copper CM2FCB rca sockets, Mundorf silver wire, Mundorf silver solder.
 
And the best - the finest 6922 I've had the pleasure of hearing (and I have had them all including Amperex PWs, Siemens CCa's Tele E88CC's, Lorentz Suttgards, etc..) the extraordinary '75 6n23p Reflector Silver Shield Single Wire Getter Posts (see my review on the Lyr tube rolling thread).
 
The stock Lite DAC 60 is very good and an incredible bargain at $650 -  with the upgrades it gives my APL DAC a stiff match (which costs any times more).
 
R2R dacs are known for their excellent natural full tone.  If your interested in the different DAC chip designs checkout this link: http://www.mother-of-tone.com/conversion.htm
 
My modded DAC60 before the stage III mods which I should be able to complete next weekend.
 
The stock dac:

 
After stage II mods
 

 
Edit: typos