[ginetto61]

  i wold vote against any usb 'boosting'. i believe there is a bit too much re over subjective granularity placed on what 'could' affect your enjoyment of  audio re the devices merited in this thread.

Hi and i agree absolutely 
However i am about to evaluate a PCIe to usb adapter card to put in the PC that should provide a stronger, more stable and cleaner usb output (hopefully).
I bought this one here ... very cheap ... with a Nec chip on board as recommended to me by someone ...
 
http://www.ebay.com/itm/PCI-PCI-E-Express-to-USB3-0-USB2-0-USB1-1-2-port-NEC-chipset-Hub-Card-MINI-PC-BC-/361128303368?ssPageName=ADME:L:OC:IT:3160
 

 
 
But i agree. I intend to use only a usb power isolator to clean the usb power lines.
Regards,  gino  

 

[Sonic Defender]

If it was simpler, I would simply a/b test blind with help to determine if there is any audible difference. But as you know, in order to test against USB I would need to not just switch cables, but also go into control panel, change the default device, and do the same in JRiver. Even with help it seems like this would take a few minutes to do and it would slow down the testing process quite a bit, and with auditory memory being somewhat unreliable anyway I worry the delays would make the testing fairly unreliable anyway.

 

[ginetto61]

I would like to add just one another thing.  Using a usb power isolator in case of not isolated usb dac pays.
And also the quality of PS used pays.  
Of course much more with usb dacs relying only on usb power to work.
I am testing for instance a nice Hegel HD10 dac that does need power in the usb input to work, sadly ... 

 flaw in the design for me.
Presently i am using a cheap but nice sounding usb dac/headphone amp from Breeze Audio to find the best sounding combination of power supply, isolator and usb cables.

Once i have got the best sound i will use the same elements for the dac above mentioned that should be superior, hopefully. 
Usually very good sound sounds relaxed but still providing a great detail
and of course does not induce fatigue just desire to go on with the listening.  
And HPs are very useful in this process.  Being very transparent they let everything through
and when the sound is bad this is very evident. 

 

[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

 
 
No reaction?????
 
This one is THE ONLY ONE (till so far) which would work with highres music, so all above 96kHz.

 

[Sonic Defender]

It's quiet around here lately, for myself, I wonder how much of these tweak devices are ultimately just fun distractions that we enjoy working into our rigs? I must admit, I read posts about how awesome the MX-U8 can make things and frankly even with a decent Digiflex AES cable I can't hear a difference.

 

[ginetto61]

It's quiet around here lately, for myself, I wonder how much of these tweak devices are ultimately just fun distractions that we enjoy working into our rigs? ...

Hi i think that in the case of the Gustard U12 USB Interface, as its usb input is not isolated, some kind of isolation device is needed.
From what i have learned up to now i would focus only on power isolation from the pc of the usb port on the U12 and stop.
I have read that with longer usb cables packets loss is possible. Then that ferrite beads placed along the usb cable can distort the signal.
So if i had to use the U12 (in the meantime i have bought also MX-U8 that overall looks like a better device and has isolated usb port) i would:
1)   keep the usb cable as short as possible, best quality without ferrite beads
2)   use a usb power isolation device without any chip on the signal path, just a physical disconnection of screen, 5VDC and ground and let through only the data wires providing external clean 5VDC.
3)   try a pci-e to usb card on the pc in order to try to get a better usb signal out of the pc.
Instead with MX-U8 point 2) is needless, of course.
I thought usb signal transmission were less challenging ... instead it is difficult to perform rightly.
Thanks and regards,  gino

 

[SodaBoy]

Quote:

   
 
No reaction?????
 
This one is THE ONLY ONE (till so far) which would work with highres music, so all above 96kHz.

I am definitely interested, but September is pretty far away. The price sounds OK so long as it delivers considering I have seen many other garbage isolators in the past that were not much cheaper.

 

[Vic57]

I successfully used self-made chokes on ferrite rings for a filtration of the USB supply voltage.
Not the entire cable, as only two supply conductor is wound 5-7 turns per ring in the opposite direction ...

 

[pakultra]

Hi everyone, I've just finished building my ultra-low noise power supply based on the 0.8uV DIYinHK kit and I'd like to share what I've found about this kit.
 
First, some photos of the build:
 

 
 


 
 
Since I need 1.6A supply for my NAD D1050 dac, I paralleled the two individual 5V outputs with some changes to the original circuit. Theoretically, paralleling the regulators will reduce the noise even further, to 0.57uV in my case. However, there is no way I can measure it with my humble multimeter. What I do know is that right now, it's working perfectly with my DAC. Actual output voltage: 5.06V. Current consumption is very steady at 1.4A, exactly 0.7A measured from each output, what a perfect balance of load! I am amazed at the performance of these nice LT3042 regulators.
 
After I put my new power supply into service, my wife immediately noticed improvement in the sound. She said that the sound was SWEETER, to which I fully agreed. Previously I always start to feel a little fatigue after listening to my system for more than one hour, now that fatigue is all gone. 
 
I am now very happy with the combination of my power supply, my dac and my U12, although I did encounter quite a few issues while building the power supply:
 
1) Although DIYinHK claims that this kit is capable of supplying 1A x 2, I believe that the original design is only good for a maximum of 500mA x 2. The Murata hybrid EMI filters supplied in the kit are rated at 500mA only. In my initial tests, they got VERY hot at 500mA. Once the current went above 500mA, the filters got saturated and there were a huge voltage drop once the chokes were saturated. As a result, the output voltage plummeted to 3.7v from 5v upon saturation. This alone won't be too much of a problem if you only need 3.3V. The board can still output a steady 3.3v if you set its voltage jumper to 3.3v. But since I need 5V, the voltage drop was totally unacceptable to me. Eventually, I had to bypass the two nice-looking but useless filters to get a steady 5V output;
 
2) The default heatsinks that come with the DIYinHK are too small for current output above 500mA. The transistors got burning hot when more current was drawn. I was so alarmed at this that I checked again on the DIYinHK website to see what could be wrong. This time, I noticed DIYinHK's warning to use larger heatsinks or a fan for cooling if more than 500 mA current is needed. To make matters worse, the PCB was specifically designed to allow only a small heatsink to be installed. There are capacitors right beside the transistors, leaving no room for larger heatsinks. Luckily, I've already replaced most of the WIMA film capacitors with tiny 22 uf surface-mounted Murata ceramic capacitors, so that I have space to mount larger heatsinks above my capacitors. This tames the heat nicely, although the board doesn't look as nice and neat as before;
 
3) Regarding the choice of transformers, I originally ordered a R-core but decided to try an EI core first because of their alleged superiority in isolating high frequency noises. My EI-core does work. It has become particularly important after I had to bypass the Murata EMI filters,  but it does get hot and it HUMs when under full load. The humming is audible within a 2-foot range of the transformer. I haven't decided if I should replace the EI core with my R-core to get rid of the humming. But for now, I have learnt to use the humming as a cue to let me know whether the power supply is working under full load of not. 
 
I originally was thinking about building the same linear power supply for my Gustard U12 too, but after I modded my U12 to be fully isolated from USB power, it has become so quite that I don't think an external power supply is needed any more. Now, my U12 and NAD D1050 powered by 0.57uV hand-made power suppy work perfectly. I hear pure, relaxed, sweet and musical sound. This further confirms that an external power supply for U12 is unnecessary, though doable.
 
Regarding the use of USB isolators, I don't think they will be good for U12 once it has already been modded and isolated from USB power. External isolators can introduce jitter and distortion. I haven't tried them but I am not motivated to try them either, at least not until I am otherwise convinced.
 
Hope my little report helps.
 
Now, time to enjoy music!
Cheers

 
 
 
 
pakultra

 

[stuartmc]

If it was simpler, I would simply a/b test blind with help to determine if there is any audible difference. But as you know, in order to test against USB I would need to not just switch cables, but also go into control panel, change the default device, and do the same in JRiver. Even with help it seems like this would take a few minutes to do and it would slow down the testing process quite a bit, and with auditory memory being somewhat unreliable anyway I worry the delays would make the testing fairly unreliable anyway.

My testing regimen was a pain in the ***. I had U12 and Tanly powered and warmed up and had to switch out the USB input cable and the I2S output cable each time. I also had to level match with a white noise signal and db meter. Of course by this time my own auditory memory was questionable. Compounding the problem was my use of three separate tracks that emphasized bass response, dynamics and soundstage/imaging respectively. It would have been so much easier if I could just toggle a switch to A/B them. Faced with this time consuming and very flawed and subjective testing, the only way I could reach any sort of confident conclusions was if the differences were very obvious to me and repeatable with each changeover. In this case it was...and rather emphatically so. Not surprisingly, I noticed the biggest improvements with the Tanly in the soundstage/imaging arena. It's obviously a composite arena that includes treble response, low level resolution and a host of other factors. I am more sensitive to this than anything else and I believe it is a learned response from years of reviewing and certainly not any gifted auditory physiology.

Anywho....that's my long winded way of saying, like Mr. Bill, "I feel your pain."

 

[Sonic Defender]

My testing regimen was a pain in the ***. I had U12 and Tanly powered and warmed up and had to switch out the USB input cable and the I2S output cable each time. I also had to level match with a white noise signal and db meter. Of course by this time my own auditory memory was questionable. Compounding the problem was my use of three separate tracks that emphasized bass response, dynamics and soundstage/imaging respectively. It would have been so much easier if I could just toggle a switch to A/B them. Faced with this time consuming and very flawed and subjective testing, the only way I could reach any sort of confident conclusions was if the differences were very obvious to me and repeatable with each changeover. In this case it was...and rather emphatically so. Not surprisingly, I noticed the biggest improvements with the Tanly in the soundstage/imaging arena. It's obviously a composite arena that includes treble response, low level resolution and a host of other factors. I am more sensitive to this than anything else and I believe it is a learned response from years of reviewing and certainly not any gifted auditory physiology.

Anywho....that's my long winded way of saying, like Mr. Bill, "I feel your pain."

Wow, that was a lot of work! No wonder in this situation most people never bother to do the testing. If I can get my friend to assist I might one day try blind listening comparison with USB versus the MX-U8 AES. For now, it certainly sounds quite good, but sadly I can't say if this is any different as it sounded quite good before the U8! I decided to just take a shot as the cost wasn't very high so it was worth it. Maybe it does sound better, who knows? Cheers.

 

[ginetto61]

  I successfully used self-made chokes on ferrite rings for a filtration of the USB supply voltage.
Not the entire cable, as only two supply conductor is wound 5-7 turns per ring in the opposite direction ...

 
Hi interesting ... do you have any pics ? what kind of improvements did you get ?
I think that a very good isolation from mains is of paramount importance for digital equipment.
Thanks,  gino

 

[ginetto61]

  Hi everyone, I've just finished building my ultra-low noise power supply based on the 0.8uV DIYinHK kit and I'd like to share what I've found about this kit.
 
First, some photos of the build:
 

 
 


 
 
Since I need 1.6A supply for my NAD D1050 dac, I paralleled the two individual 5V outputs with some changes to the original circuit. Theoretically, paralleling the regulators will reduce the noise even further, to 0.57uV in my case. However, there is no way I can measure it with my humble multimeter. What I do know is that right now, it's working perfectly with my DAC. Actual output voltage: 5.06V. Current consumption is very steady at 1.4A, exactly 0.7A measured from each output, what a perfect balance of load! I am amazed at the performance of these nice LT3042 regulators.
After I put my new power supply into service, my wife immediately noticed improvement in the sound. She said that the sound was SWEETER, to which I fully agreed. Previously I always start to feel a little fatigue after listening to my system for more than one hour, now that fatigue is all gone.
I am now very happy with the combination of my power supply, my dac and my U12, although I did encounter quite a few issues while building the power supply:
1) Although DIYinHK claims that this kit is capable of supplying 1A x 2, I believe that the original design is only good for a maximum of 500mA x 2. The Murata hybrid EMI filters supplied in the kit are rated at 500mA only. In my initial tests, they got VERY hot at 500mA. Once the current went above 500mA, the filters got saturated and there were a huge voltage drop once the chokes were saturated. As a result, the output voltage plummeted to 3.7v from 5v upon saturation. This alone won't be too much of a problem if you only need 3.3V. The board can still output a steady 3.3v if you set its voltage jumper to 3.3v. But since I need 5V, the voltage drop was totally unacceptable to me. Eventually, I had to bypass the two nice-looking but useless filters to get a steady 5V output;
2) The default heatsinks that come with the DIYinHK are too small for current output above 500mA. The transistors got burning hot when more current was drawn. I was so alarmed at this that I checked again on the DIYinHK website to see what could be wrong. This time, I noticed DIYinHK's warning to use larger heatsinks or a fan for cooling if more than 500 mA current is needed. To make matters worse, the PCB was specifically designed to allow only a small heatsink to be installed. There are capacitors right beside the transistors, leaving no room for larger heatsinks. Luckily, I've already replaced most of the WIMA film capacitors with tiny 22 uf surface-mounted Murata ceramic capacitors, so that I have space to mount larger heatsinks above my capacitors. This tames the heat nicely, although the board doesn't look as nice and neat as before;

 
Hi ! you have done a great job indeed. And the noise levels you have reached are really impressive.
How did you connect the new power supply to the U12 ?  with an umbilical ? do you have any pictures ?
 

3) Regarding the choice of transformers, I originally ordered a R-core but decided to try an EI core first because of their alleged superiority in isolating high frequency noises. My EI-core does work. It has become particularly important after I had to bypass the Murata EMI filters, but it does get hot and it HUMs when under full load. The humming is audible within a 2-foot range of the transformer. I haven't decided if I should replace the EI core with my R-core to get rid of the humming.
But for now, I have learnt to use the humming as a cue to let me know whether the power supply is working under full load of not.

very very interesting.  If the problem is the transformer rating you can try a bigger EI maybe.  I can only see that in some very high end equipment they are used, like in the Schiit Audio dacs and Berkeley Audio dacs.
I have the feeling that EI transformers can even filter some kind of HF noise that passes through the regulators.   But a solution mains filter plus toroid could also work just fine.   And it should be even simpler to implement.  Just a simple mains RF filter is needed.
Or even a power cord with high capacitance can have a nice effect in suppressing HF noise in the mains.
 

I originally was thinking about building the same linear power supply for my Gustard U12 too, but after I modded my U12 to be fully isolated from USB power, it has become so quite that I don't think an external power supply is needed any more.
Now, my U12 and NAD D1050 powered by 0.57uV hand-made power suppy work perfectly. I hear pure, relaxed, sweet and musical sound. This further confirms that an external power supply for U12 is unnecessary, though doable.

 
My feeling is that the U12 is not very well isolated from noise in the mains.   A toroid transformer without any filter upstream.  A toroid is not a good filter.  EI is.
Of course USB isolation is mandatory.  Pc power is weak and dirty.  Only data should be taken out from the pc (someone says that also data are dirty ... but to clean data is more complicated.  Someone is seriously thinking about this anyway)
 

Regarding the use of USB isolators, I don't think they will be good for U12 once it has already been modded and isolated from USB power.
External isolators can introduce jitter and distortion. I haven't tried them but I am not motivated to try them either, at least not until I am otherwise convinced.
Hope my little report helps.
Now, time to enjoy music!
Cheers

pakultra

 
I guess you are right about usb isolator based on chip.  I have the same feeling. The sound is more grainy and less developed.
I would use only usb power isolator like the Teradak U9 i am actually using.   It just interrupts the power lines and provides external power.   And it works fine.
I have instead an isolator like this one here ...
 
https://www.circuitsathome.com/wp/wp-content/uploads/2012/02/USB_isolator_built.jpg
 
and i do not like the resulting sound.   Maybe it adds some jitter as you say.
It was just a try ... not very successful unfortunately ...
Congratulations again for your excellent project
Regards, gino

 

[abartels]

  Hi everyone, I've just finished building my ultra-low noise power supply based on the 0.8uV DIYinHK kit and I'd like to share what I've found about this kit.
 
First, some photos of the build:
 

 
 


 
 
Since I need 1.6A supply for my NAD D1050 dac, I paralleled the two individual 5V outputs with some changes to the original circuit. Theoretically, paralleling the regulators will reduce the noise even further, to 0.57uV in my case. However, there is no way I can measure it with my humble multimeter. What I do know is that right now, it's working perfectly with my DAC. Actual output voltage: 5.06V. Current consumption is very steady at 1.4A, exactly 0.7A measured from each output, what a perfect balance of load! I am amazed at the performance of these nice LT3042 regulators.
 
After I put my new power supply into service, my wife immediately noticed improvement in the sound. She said that the sound was SWEETER, to which I fully agreed. Previously I always start to feel a little fatigue after listening to my system for more than one hour, now that fatigue is all gone. 
 
I am now very happy with the combination of my power supply, my dac and my U12, although I did encounter quite a few issues while building the power supply:
 
1) Although DIYinHK claims that this kit is capable of supplying 1A x 2, I believe that the original design is only good for a maximum of 500mA x 2. The Murata hybrid EMI filters supplied in the kit are rated at 500mA only. In my initial tests, they got VERY hot at 500mA. Once the current went above 500mA, the filters got saturated and there were a huge voltage drop once the chokes were saturated. As a result, the output voltage plummeted to 3.7v from 5v upon saturation. This alone won't be too much of a problem if you only need 3.3V. The board can still output a steady 3.3v if you set its voltage jumper to 3.3v. But since I need 5V, the voltage drop was totally unacceptable to me. Eventually, I had to bypass the two nice-looking but useless filters to get a steady 5V output;
 
2) The default heatsinks that come with the DIYinHK are too small for current output above 500mA. The transistors got burning hot when more current was drawn. I was so alarmed at this that I checked again on the DIYinHK website to see what could be wrong. This time, I noticed DIYinHK's warning to use larger heatsinks or a fan for cooling if more than 500 mA current is needed. To make matters worse, the PCB was specifically designed to allow only a small heatsink to be installed. There are capacitors right beside the transistors, leaving no room for larger heatsinks. Luckily, I've already replaced most of the WIMA film capacitors with tiny 22 uf surface-mounted Murata ceramic capacitors, so that I have space to mount larger heatsinks above my capacitors. This tames the heat nicely, although the board doesn't look as nice and neat as before;
 
3) Regarding the choice of transformers, I originally ordered a R-core but decided to try an EI core first because of their alleged superiority in isolating high frequency noises. My EI-core does work. It has become particularly important after I had to bypass the Murata EMI filters,  but it does get hot and it HUMs when under full load. The humming is audible within a 2-foot range of the transformer. I haven't decided if I should replace the EI core with my R-core to get rid of the humming. But for now, I have learnt to use the humming as a cue to let me know whether the power supply is working under full load of not. 
 
I originally was thinking about building the same linear power supply for my Gustard U12 too, but after I modded my U12 to be fully isolated from USB power, it has become so quite that I don't think an external power supply is needed any more. Now, my U12 and NAD D1050 powered by 0.57uV hand-made power suppy work perfectly. I hear pure, relaxed, sweet and musical sound. This further confirms that an external power supply for U12 is unnecessary, though doable.
 
Regarding the use of USB isolators, I don't think they will be good for U12 once it has already been modded and isolated from USB power. External isolators can introduce jitter and distortion. I haven't tried them but I am not motivated to try them either, at least not until I am otherwise convinced.
 
Hope my little report helps.
 
Now, time to enjoy music!
Cheers

 
 
 
 
pakultra

 
Hi pakultra,
 
Thanks for sharing this with us. I didn't take the time to look at the specs of Murata 1430R5, they indeed are 500mA only, i did report at DIYINHK and got an answer to remove them if they get too hot,,, hmm, not the answer I was looking for,,,,
About the LT3042, it can deliver 1A when expanding schematics with external NPN, and that is what the people at DIYINHK did. Off course you can parallel more LT3042 to get higher output if needed.
 
For me higher output power is of no importance since I will use them only to power the NDK's, one psu for each NDK. Since the NDK's only draw 6mA it will be no problem at all.
 
I suppose those psu's are designed for the isolated I2S kit they sell, so design is not for high-power appliances.
 
 
Regards and keep up the good work 

 
Alex

 

[pakultra]

Hi ginetto61, thank you very much for your encouragement

 
I didn't connect my power supply to U12. I connected it to my NAD D1050 dac in place of its factory default external 5V switching mode power supply. I just built a DC to DC cable out of Mogami 2549 and plugged it into my NAD D1050 5V DC power socket.
 
U12 doesn't need this much power. 1A is enough for U12. To connect to U12, there are a few ways to achieve this. Here is what I'd suggest doing:
 
1) disconnect pin 1 (Adjust)  and pin 3 (Vin) of the LM317 regulator. Pin 1 and Pin 3 are the two pins at the sides. These pins can be easily desoldered or just cut. Pin 2 (Vout) is in the center and it can be left as it is. Disconnecting pin 3 effectively disconnect everything U12 uses to generate its internal 5V power, including the Toroid transformer, the diodes and the two 2200 uf capacitors;
 
2) connect the positive of your external 5V power to Pin 2 of LM317, or to Pin 3 (Vin) of LM1117, or to the 5V onboard via that is marked as 5V if you'd like to make connections at the bottom of the pcb;
 
3) connect the negative of your external 5V power to Pin 1 (Adjust) of LM1117, or to Pin 1 of any of the two PE-65612 transformers if you'd like to work from the bottom of the pcb.
 
Regarding transformers, you are right about picking a higher rating EI core. The one I used is rated at 8V x 1.25A x 2. Theoretically this is sufficient for what I need. I had thought about using its large brother that would be able to provide twice as much power, but eventually I picked the smaller one in favor of its smaller size. Next time I will definitely pick a bigger one.
 
My whole system including the U12 is sitting behind a dedicated mains filter, that's probably why I haven't noticed any RF problems so far. By the way, I've also had Stetzerizer filters installed and have measured the RF noise level of my power sockets with a Stetzerizer Microsurge Meter. According to the meter readings, my mains' RF noise level is better than ideal.
 
Regarding the USB isolator, I think the key is to have data reclocked after the isolation. Since you've already had an isolator at hand, I'd suggest that you try connecting a bus powered good usb hub between your isolator and your U12. Only connect the data lines from the hub to your U12 and keep using your Teradak U9 to supply power to U12's USB socket. The usb hub can function as a regenerator to smooth out the jitter. I don't know if this will work for you but I think it is worth trying it.
 
Regards
 
pakultra