[abartels]

Hi all,
 
I think this great USB interface deserves a seperate thread.
 
 
Just post all your questions, experiences, comparissons, modifications and tips & tricks 

We had a long way to go, started at Gustard U12 USB Interface 8 Core XMOS chip thread were you can read the beginnings,,,,,
 
 
Cheers to all 

 
 
Alex

 

[ginetto61]

Hi Alex
and here it is  for the people who don't know it  
 

 
(in this way i get the alert of new posts) 

bye gino 

 

[abartels]

  Hi Alex
and here it is  for the people who don't know it  
 

 
(in this way i get the alert of new posts) 

bye gino 

 
Hi Gino,
 
That's a good start 

 
I will order 2 R-Core transformers with copper shield for my MX-U8, one for the device itself, and one to feed the XO's.
 
Didn't anybody did any mods on their MX-U8??
 
Regards,
Alex

 

[ginetto61]

Hi Gino,
That's a good start 

I will order 2 R-Core transformers with copper shield for my MX-U8, one for the device itself, and one to feed the XO's.
Didn't anybody did any mods on their MX-U8??
Regards,
Alex

 
Hi Alex, i think that the idea is very sane.  Still i would prefer to place a panel mount DC socket to bypass the internal transformer and reach the big caps after the diodes bridge, leaving anything else untouched.
I have at hand at least 2 very low noise linear power supply, lab grade.  In this way i could provide very clean DC and the regulator on board would act as a second regulation stage.
These power supplies have around 1mV of residual ripple.   I know that this would mean to add another box ... but this unit is really key in the chain. It can make or break the sound.
And the added benefit would be to add the working transformer out of the box ... completely isolated ... no vibes ... no EMI.
I mean a socket like this one ...
 
http://img-europe.electrocomponents.com/images/R0487842-01.jpg
 
Thanks again to keep this thread alive and kicking

have a nice weekend
gino

 

[abartels]

   
Hi Alex, i think that the idea is very sane.  Still i would prefer to place a panel mount DC socket to bypass the internal transformer and reach the big caps after the diodes bridge, leaving anything else untouched.
I have at hand at least 2 very low noise linear power supply, lab grade.  In this way i could provide very clean DC and the regulator on board would act as a second regulation stage.
These power supplies have around 1mV of residual ripple.   I know that this would mean to add another box ... but this unit is really key in the chain. It can make or break the sound.
And the added benefit would be to add the working transformer out of the box ... completely isolated ... no vibes ... no EMI.
I mean a socket like this one ...
 
http://img-europe.electrocomponents.com/images/R0487842-01.jpg
 
Thanks again to keep this thread alive and kicking

have a nice weekend
gino

 
Hi Gino,
 
I still think my solution will sound much better. I'll explain why.
 
The influences of vibration of transformer and RFI/EMI will be less important than NOISE, and NOISE is NOT the same as RIPPLE.
Try finding what kind of regulators your lab supplies use, THEN you can see how much noise they produce.
 
I suppose they have AT LEAST between 40uV and 100uV (uV is micro volt, mV is millivolt) noise.
 
I am talking about 0.1 micro volt (uV) noise, feeding the XO's DIRECTLY, so NO ADP150 inbetween, which btw has 9uV noise, and NO LT1963, which btw has 40uV noise.
 
So, out of the box the (audio)crystals psu feed has 49uV noise, in my case it will have 0.8uV noise.
 
 
 
I wouldn't wonder if your solution has a total noise figure of more than hundreds of uVolts, which will degrade SQ in a huge way.
 
R-Core's will do the job, maybe mounted on rubber sheet, and EMI/RFI shield between them and the rest of the unit.
 
 
Just my 5 cents,,,,
 
 
 
Keep up the modding 

 
Cheers 

 
Alex

 

[ginetto61]

Originally Posted by abartels /img/forum/go_quote.gif
Hi Gino,
I still think my solution will sound much better. I'll explain why.
The influences of vibration of transformer and RFI/EMI will be less important than NOISE, and NOISE is NOT the same as RIPPLE.
Try finding what kind of regulators your lab supplies use, THEN you can see how much noise they produce.
I suppose they have AT LEAST between 40uV and 100uV (uV is micro volt, mV is millivolt) noise.
 
I am talking about 0.1 micro volt (uV) noise, feeding the XO's DIRECTLY,
 
so NO ADP150 inbetween, which btw has 9uV noise, and NO LT1963, which btw has 40uV noise.
So, out of the box the (audio)crystals psu feed has 49uV noise, in my case it will have 0.8uV noise.
I wouldn't wonder if your solution has a total noise figure of more than hundreds of uVolts, which will degrade SQ in a huge way.
R-Core's will do the job, maybe mounted on rubber sheet, and EMI/RFI shield between them and the rest of the unit.
Just my 5 cents,,,,
 
Keep up the modding 

Cheers 

Alex

 
Hi Alex and ok ... that is really low

My ps is a Mascot 719 power Supply
 
http://www.komteh.hr/download/radio/mascot/acdc/719.pdf
 
Probably you have said that and i missed it ... sorry 

  but what regulators are you going to use in order to reach almost zero noise ?
I need to keep mods basic ... i am not that clever at DIY.
Thanks again,  gino

 

[abartels]

uote:

   
Hi Alex and ok ... that is really low

My ps is a Mascot 719 power Supply
 
http://www.komteh.hr/download/radio/mascot/acdc/719.pdf
 
Probably you said and i have missed 

  but what regulators are you going to use in order to reach almost zero noise ?
I need to keep mods basic ... i am not that clever at DIY.
Thanks again,  gino

 
Hi Gino,
 
It's this one, it's a double version and needs 2 seperate 6~8 Volts input from transformer (transformer with 2x6V should do the trick)
 
http://www.diyinhk.com/shop/audio-kits/73-08uv-ultralow-noise-dac-power-supply-regulator-33v5v-1ax2.html
 

It's a kit, all the SMD components are soldered, but the Through-hole components need to be soldered by yourself.
 
In addition you need to exchange the Panasonic FC 5600uF-25V to the ones i described before:
I advise replacing them with Nichicon HW series 4700uF-25V ( UHW1E470MHD ) with a hight of 31,5mm
 
When starting modding I will do it in a step-by-step manner including pics.
 
Cheers and have a nice weekend!

 
Alex 

 

[ginetto61]

Hi and very impressive noise figures indeed.
But do you think that also with my minimal mod the noise should be lower or not ?
I think that the Mascot provide also a better mains noise filtering.
Anyway i will follow your project for sure.  Clearly this very good unit can be improved.
Thanks again,  gino

 

[abartels]

Hi Guys,
 
I found this interesting site, they have the WORLDS FIRST HIGH-SPEED USB ISOLATOR
 
http://intona.eu/en/products
 
Not cheap, but maybe a very nice add-on for us!
 
 
Cheers 

 
Alex

 

[kiertijai]

I am interested in this Melodious .  It has DSD256.
However DSD has to use I2S transmission to another DAC?
My DAC does not have I2S input but does have coaxial, optical, AES
Is there any adapter for this I2S to coaxial?  anything simple?

 

[MINORISUKE]

  I am interested in this Melodious .  It has DSD256.
However DSD has to use I2S transmission to another DAC?
My DAC does not have I2S input but does have coaxial, optical, AES
Is there any adapter for this I2S to coaxial?  anything simple?

Before discussiong "I2S to coax", I have to say that there is a channel mix (not swap) in DSD playback.  Melodious has not given me the test result that it did not happen for other units.  Therefore, I cannot recommend this unit for DSD playback until Melodious confirms nothing is wrong.
http://www.head-fi.org/t/736294/gustard-u12-usb-interface-8-core-xmos-chip/1395#post_11563853
 
Does your DAC accept DSD (DoP) through coax/optical S/PDIF?  Not many DACs support this feature.  Even so, I know no DAC supporting DSD256 in DoP via S/PDIF, for 705.6MHz PCM must be handled, which S/PDIF does not support anyway.
I think you have to design a circuit and logic including DoP encoding by yourself, as I do not see any market demand for such an adapter.
DSD output from RJ45 (I2S marked output) of MX-U8 is native.  It has to be encoded into DoP for such purposes.
 
I have read "DSD output through I2S" is a wrong expression, based on the definition of I2S.  "I2S and DSD native output" would be correct.
(DSD is a method of digitisation like PCM, whereas I2S is a method of interface within PCM equipment.)

 

[ginetto61]

Hi i would like to read some listening impressions about this unit with info about the source (pc) and the dac used.
 
I have been listening the Melodious with a desktop win 7 64 bit pc > belden usb cable too long i guess (3 meter) > Melodious > AES Canare cable (good and cheap) > Apogee Rosetta 200 dac > Sac headphone amp > AKG k701
and the sound is quite good.  Actually the best i have heard from my pc so far.
 
In general i am very interested about what you are using with it.  This device could very well become the corner stone in my system.
Even in stock form

 .... 

 
Thanks and regards,  gino

 

[ginetto61]

Originally Posted by abartels /img/forum/go_quote.gif
Hi Gino,
 
It's this one, it's a double version and needs 2 seperate 6~8 Volts input from transformer (transformer with 2x6V should do the trick)
http://www.diyinhk.com/shop/audio-kits/73-08uv-ultralow-noise-dac-power-supply-regulator-33v5v-1ax2.html
It's a kit, all the SMD components are soldered, but the Through-hole components need to be soldered by yourself....
When starting modding I will do it in a step-by-step manner including pics.
Cheers and have a nice weekend!

Alex

 
Hi Alex ! i have one question about the original power supply schematic/design.
 
Is there just one regulator for all the circuit ?
because i see two diode bridges for what looks like a dual voltage power supply.  So i would expect two regulators, one per each rail.
If not i do not understand the two bridges.
 
I will peep inside soon but if you already know this thing i would be very thankful.
Bye,  gino

 

[abartels]

   
Hi Alex ! i have one question about the original power supply schematic/design.
 
Is there just one regulator for all the circuit ?
because i see two diode bridges for what looks like a dual voltage power supply.  So i would expect two regulators, one per each rail.
If not i do not understand the two bridges.
 
I will peep inside soon but if you already know this thing i would be very thankful.
Bye,  gino

 
Hi Gino, I suppose it's something like the underneath picture:
 

 
 
Regards,
 
Alex

 

[ginetto61]

   
Hi Gino, I suppose it's something like the underneath picture:
 

 
 
Regards,
 
Alex

 
Hi Alex, thank you very much indeed for the very helpful advice.
So the ps is single voltage.  And just one regulator.  Very good !
Still i see 4 big blue caps

i guess the 2200uF parts ... while in you schematic i see just two. 

 
However for me the very good news is that there is only a single regulator. This simplifies things hugely.
 
I am sure your project is very good indeed, but sadly too challenging for me.
I would be more than satisfied just by increasing the mains HF noise filtering (with a filter or using a different transformer type) and adding another regulation stage upstream the on-board regulator.
Clearly i will not be able to reach your impressive noise figures ... really outstanding.
But my skills are limited and the risk to damage everything is around the corner.
To end i am extremely optimistic about the actual performances that can be obtained by this unit.
Even as stock the sound out of the AES connection is very very nice.  Very musical. I like it very much indeed.

As soon as i will be free i will peep inside and maybe take some pics.
Thanks a lot again.
Have a nice day !
gino