Modding USB cable to power DAC with external PSU

[rodelius]

Hello all.
My DAC sounds fantastic through its SPDIF input but when driven through USB the sound quality strongly depends on the associated computer.
I know it should not happen with async USB but the difference is clearly audible.
With my laptop (battery powered) it sounds great, with one of my Mac Minis it sounds acceptable and with my newer Mac Mini with integrated SMPS it sounds awful.

As my DAC's USB receiver is powered by the computer and my HiFace EVO SQ clearly improved when I replaced the cheap PSU for a nice regulated supply I began to investigate the ways to power the USB receiver chip externally. I tested the DAC with a Placid supply and usb power injector cable but it did not work with any of the Macs. The DAC USB led always lights up but the computer does not recognize it.
I then modified an USB Female B to Male B adapter. Isolated the +5v coming from the computer and soldered the Placid cables to pin 1 and 4 at the USB male end leaving pin 4 connected to both ends and it did not work either. Would it be possible that the computer +5v needs to be terminated in a certain way or cannot be simply cut because it is used to negotiate with the USB device? Or should the external PSU be connected to pin 1 and ground instead?
I searched all over the forums and it seems that there is no consensus on the right way to do this.

Before going for a commercial solution e.g. the AQVOX USB Supply or purchasing a High End USB Hub I would like to have a clue of what is really needed to power an USB DAC externally in a manner that the DAC is recognized by any Mac.

Any help will be greatly appreciated.

Rod

 

[twest820]

I searched all over the forums and it seems that there is no consensus on the right way to do this.

USB supports both self and bus powered devices and provides specific timings for device enumeration in response to cable attach or bus power up/down and reset events---refer to the USB specifications for details. From your description I would guess the power splicing is breaking device enumeration. You might be able to work around this by slaving the external supply to the bus supply but there are good reasons the USB Implementor's Forum does not certify extension cables, power breakouts, and so on as USB compliant.

Assuming the DAC is operating as an asynchronous USB Audio 2.0 class device as expected you're probably better off leaving the USB interface bus powered and focusing on noise management around the A/D supplies and ground bounce. Hard to say much meaningful without a schematic of the DAC and measurements showing what the problem is---"sounds awful" is many things but precise is not one of them.

 

[counter culture]

It's possible that the timing of the power connection might be making some difference and if you plug it into the computer and then turn on the power, or vice versa if this is what you're doing already, it may work.

 

[rodelius]

Assuming the DAC is operating as an asynchronous USB Audio 2.0 class device as expected you're probably better off leaving the USB interface bus powered and focusing on noise management around the A/D supplies and ground bounce.

DAC is a commercial unit for which I don't have any shematics. I assembled several DACs but I don't feel comfortable fiddling with this DAC or with the computer.

For some unknown reason the Mac cannot be fooled or I am doing something wrong.

As there are several commercial units that do what I am trying to do (AQVOX ifi, I will keep investigating this route.

Thanks!

 

[rodelius]

It's possible that the timing of the power connection might be making some difference and if you plug it into the computer and then turn on the power, or vice versa if this is what you're doing already, it may work.

I tried that. It did not work

 

[phase]

I believe that there needs to be a 220 ohm 1/2 watt resistor on the host(mac) side, between the usb +5v supply and -5v. Pretty sure that the chassis ground/connector shell is not included in this operation. This tells the mac that there is a load/device, and to look further at that port for data.

I've done this with automotive computers to allow modern engines to be used in older cars that don't have the features, sensors that the new cars have(solenoids in the transmissions, roll-over fuel shutoff, etc). Same action, just canbus instead of USB.

I'm about to jump into this myself, but with an old PC. I want to expirement with different wire for the signal, since it's pretty apparent that the wire itself acts as a filter. Even in the recording studio forums there are rampant "discussions" on that topic.

 

[jerry g]

here is a great link...

2 Channel Audio : DIY - Solid core silver audiophile USB - Yes your cable makes a difference

What I did was modify a generic 2 foot long USB extention cable with a female and male end for multi purpose playing.
I broke the + line (red) and connected a 5volt switching supply + to the far side of the cut red lead that goes to the output plug at pin 1. I connected the - of the 5 volt switching supply to the - output (black) at pin 4 without breaking the line (scraped off the dielectric)
I didn't connect a diode and resistor. It works into my $42 hifimediy usb/dac from my laptop, but it doesn't work from my PC to my U link.

I might go buy a diode and resistor and try to fool the output from the computer, but the Bel Canto ulink is so great sounding by itself, why bother?

 

[ginetto61]

... I might go buy a diode and resistor and try to fool the output from the computer, but the Bel Canto ulink is so great sounding by itself, why bother?

Hi do you mean that you are using the Bel Canto connected only to the PC usb port ? if so it means that it is very well designed and built because power on usb is bad quality.
But it could sound even better with a solution like this one ... i am not sponsoring it just describe what i have in mind now

An externally hosted image should be here but it was not working when we last tested it.

it is clear that the better the quality of the +5VDC the better the performance.
Problem is that i would just need the piece in the attachment
because i have already a high quality +5VDC power supply ...
Any idea ? thanks a lot and kind regards, gino

 

[Julf]

it is clear that the better the quality of the +5VDC the better the performance.

Is it?

Problem is that i would just need the piece in the attachment
because i have already a high quality +5VDC power supply ...
Any idea ? thanks a lot and kind regards, gino

Get a cheap USB cable, cut it at one end, and solder it to your PS.

 

[ginetto61]

Is it?

Hi and thanks for the reply
Well i think because the quartz inside the dac/converter usually benefits from a clean and stable supply (i am guessing here)
Moreover i read many reviews of these devices and usually the conclusions are pretty positive.
So yes i think they are beneficial

Get a cheap USB cable, cut it at one end, and solder it to your PS.

Looking at the cable structure in the attachment do you mean that i can connect the red to the +5V and the black to 0 from the power supply ?
Just that ? i will try for sure
Thanks a lot. Kind regards, gino

 

[Julf]

Hi and thanks for the reply
Well i think because the quartz inside the dac/converter usually benefits from a clean and stable supply (i am guessing here)
Moreover i read many reviews of these devices and usually the conclusions are pretty positive.

For pretty much any device you can think of there will always be a bunch of positive subjective reviews.

For any benefit, there is a point of diminishing returns, where other factors start to dominate, so just going "better" on the same parameters won't give any further improvement.

Looking at the cable structure in the attachment do you mean that i can connect the red to the +5V and the black to 0 from the power supply?

If all you need is the power, then yes. If you also need the data, then you need to keep the data connections intact, but "hijack" the red and black wire.

 

[ginetto61]

For pretty much any device you can think of there will always be a bunch of positive subjective reviews.
For any benefit, there is a point of diminishing returns, where other factors start to dominate, so just going "better" on the same parameters won't give any further improvement.

I see. The only way is to try but if these people put out a product i think that i have seen a need ?

If all you need is the power, then yes. If you also need the data, then you need to keep the data connections intact, but "hijack" the red and black wire.

Yes the idea is to use only the signal from the pc and provide at the end of the cable power from an external regulated power supply
I am looking at usb cable and i see quality varying a lot.
For instance in some case the wires inside are very tiny ... i will try some cables from big name brand like Belden.
Thanks again, gino

 

[Redshift187]

You could try an isolator, something like this kit. You could skip the on-board regulator and provide your high quality 5V to the VBUS and AGND holes.

ADuM4160 USB Isolator board, bare PCB and kits « Circuits@Home

 

[Julf]

I see. The only way is to try but if these people put out a product i think that i have seen a need?

Depends on the definition of "need". A lot of audiophile accessories sell because of the insecurity and obsessiveness of audiophiles. "More must be better". Just wait for 772 kHz/64 bit recordings...

Yes the idea is to use only the signal from the pc and provide at the end of the cable power from an external regulated power supply
I am looking at usb cable and i see quality varying a lot.
For instance in some case the wires inside are very tiny ... i will try some cables from big name brand like Belden.

Agree, belden or any of the other respectable names should do it.

 

[ginetto61]

Depends on the definition of "need". A lot of audiophile accessories sell because of the insecurity and obsessiveness of audiophiles. "More must be better". Just wait for 772 kHz/64 bit recordings...

Yes i agree about the obsession ... i am also at risk of obsession
It was not like this with LPs anyway.
For instance just to try i connected a plastic turntable to my system and i like the sound It was very low quality and still musical
The digital is much more tricky
There are those who state that big money are needed to get a musical digital
Other that is a miracle ... the result of some lucky combinations
Sometimes there is musicality but no detail ... other times the opposite.
It is a mess.
May i ask you which dac you use ? are you aware of musical and not very expensive dacs ?

Agree, belden or any of the other respectable names should do it.

Perfect i will stick with this good brands and forget audiophile offers.
Thanks a lot again. Regards, gino

 

[ginetto61]

You could try an isolator, something like this kit. You could skip the on-board regulator and provide your high quality 5V to the VBUS and AGND holes.
ADuM4160 USB Isolator board, bare PCB and kits « Circuits@Home

Hi and thanks a lot for the kind suggestion
I think that i will try something much more basic ... just opening a usb cable e connecting the power wires to an external PS and listen
If it will not work i will pass to something more complicated
Thanks again, gino

 

[Julf]

It was not like this with LPs anyway.
For instance just to try i connected a plastic turntable to my system and i like the sound It was very low quality and still musical
The digital is much more tricky

I am not sure I agree. A turntable or tape deck was still pretty far from "audibly perfect", and needed a lot of trickery to get the best out of. Digital has made it both cheap and easy - any moderately priced, decently designed digital audio system provides far superior audio quality (in an objective sense) out of the box than the old analog systems.

Some people might prefer the coloration of the analog systems, but that is a different story.

There are those who state that big money are needed to get a musical digital

Usually by people who try to get their own slice of that big money - either directly, by selling gear and tweaks, or indirectly through advertising (audiophile magazines and commercial web sites such as computer audiophile).

May i ask you which dac you use? are you aware of musical and not very expensive dacs?

My main system doesn't have a separate DAC - it is an active system with the Hypex DLCP active crossover and ncore class D amps, so I am using the built-in DACs of the DLCP. Most of my listening is on my office system - active Genelec monitors driven by an audioengine D1.

I am not a believer in "musical" systems. I want my system to reproduce the original waveform as accurately as possible, no matter how musical or unmusical the original wave is. It is up to the artist and studio engineer to make it "musical" (or not) as they see fit.

 

[ginetto61]

I am not sure I agree. A turntable or tape deck was still pretty far from "audibly perfect", and needed a lot of trickery to get the best out of

Hi and i think i have to elaborate a little more.
With analog i have always had the feeling of music, the sound is flowing more easily, the virtual sound-stage is more dense, the sound has more body
Let's be frank ... with digital we are cutting in piece and then reassemble everything.
Once i read with great interest an article about digitalization of analog master tapes.
In the AD and DA process, even if performed with top equipment, something
was lost. For instance the soundstage of the copy was shrinked, smaller in every direction (width and depth). Constricted.

Digital has made it both cheap and easy - any moderately priced, decently designed digital audio system provides far superior audio quality (in an objective sense) out of the box than the old analog systems.

Yes but still i have heard digital just unbereable. This has never happened with analog ... even musicassette ...
Maybe they were not perfect at all
It is like for imaging ... when you get digital artifacts they are just unbearable ...
When digital is bad is shockingly bad. And it is easy to get it bad.

Some people might prefer the coloration of the analog systems, but that is a different story.
Usually by people who try to get their own slice of that big money - either directly, by selling gear and tweaks, or indirectly through advertising (audiophile magazines and commercial web sites such as computer audiophile).

you are very right but for me this is the evidence that they are not satisfied with their digital set-up
I am completely digital now but i have friends who have still their old Thorens or Dual turntable and they are happy of the sound like day one.
Analog starts from decent up ...
Digital can be very undecent and up ... and i have to say also that some cheap digital can be very satisfying ... one case in the picture for example
the Rotel 965bx cd player ...

An externally hosted image should be here but it was not working when we last tested it.

i still remember the first i listened to it and liked it immediately.
Now it is very cheap indeed.

My main system doesn't have a separate DAC - it is an active system with the Hypex DLCP active crossover and ncore class D amps, so I am using the built-in DACs of the DLCP. Most of my listening is on my office system - active Genelec monitors driven by an audioengine D1.
I am not a believer in "musical" systems. I want my system to reproduce the original waveform as accurately as possible, no matter how musical or unmusical the original wave is. It is up to the artist and studio engineer to make it "musical" (or not) as they see fit

Hi and thanks very interesting and i am thinking also to a digital crossover as a solution but i am in the middle of an experimenting phase.
I have just bought a clock generator to try it with a cheap dac and listen for any improvement.
I think that with digital tempo/pace/rhythm is a challenge, like also soundstage.
But i have the feeling than when these two parameters are right than the result is also musical ... very analog without the issues of the analog of course. So i am very satisfied.
Thanks a lot again. Regards, gino

 

[Julf]

With analog i have always had the feeling of music, the sound is flowing more easily, the virtual sound-stage is more dense, the sound has more body

Yes, and that is of course your personal feeling. Nothing wrong with that.

Let's be frank ... with digital we are cutting in piece and then reassemble everything.

No, we aren't. We are representing a continuous wave with regular points on that wave, allowing us to perfectly represent the original wave.

Once i read with great interest an article about digitalization of analog master tapes.

There is and endless amount of articles - some more accurate and correct than others.

In the AD and DA process, even if performed with top equipment, something
was lost. For instance the soundstage of the copy was shrinked, smaller in every direction (width and depth). Constricted.

That is interesting, considering there have been several well-controlled tests where a digital step (ADC and DAC) have been inserted in an analog chain. The listeners have not been able to tell a difference.

Yes but still i have heard digital just unbereable. This has never happened with analog ... even musicassette ...

Really? Most cassette tapes I listened to were horrible. Distortion, hiss, wow and flutter.

When digital is bad is shockingly bad. And it is easy to get it bad.

But also easy to get much better than any analog system.

I am completely digital now but i have friends who have still their old Thorens or Dual turntable and they are happy of the sound like day one.

That's because they prefer the distortion and coloration that a vinyl system adds to the sound.

Analog starts from decent up ...
Digital can be very undecent and up ...

I have to disagree. A $50 digital audio player provides a vastly superior sound compared to a $1000 vinyl player or tape deck.

I think that with digital tempo/pace/rhythm is a challenge, like also soundstage.

Both are usually a result of uneven frequency response that emphasizes certain frequencies.

If you look at real (and not just perceived) tempo and pace, a digital system gets timing exactly right (down to the stability of the crystal that measured in parts per million), whereas tape or vinyl has huge time variations caused by wow and flutter measured in parts per thousand.

 

[ginetto61]

Yes, and that is of course your personal feeling. Nothing wrong with that.
No, we aren't. We are representing a continuous wave with regular points on that wave, allowing us to perfectly represent the original wave.

Hi and thanks again and now that you say this i think i am wrong.
But in general when you copy something it cannot be completely equal to the original. Some distortions must be introduced.

There is and endless amount of articles - some more accurate and correct than others.
That is interesting, considering there have been several well-controlled tests where a digital step (ADC and DAC) have been inserted in an analog chain. The listeners have not been able to tell a difference.

Sorry .... are you saying that listeners have not been able to tell the difference between the direct signal and the same signal after a AD and DA conversion ?
If you tell me which AD/DA converter they have used for the test i would be very interested.
I am afraid with a very good playback system the difference between direct feed and the same feed after passing a AD and DA process should be maybe minimal but evident.

Really? Most cassette tapes I listened to were horrible. Distortion, hiss, wow and flutter. But also easy to get much better than any analog system. That's because they prefer the distortion and coloration that a vinyl system adds to the sound.

it could be that i am used to this kind of familiar distortion ? it is possible
digital surprised me many times ...

I have to disagree. A $50 digital audio player provides a vastly superior sound compared to a $1000 vinyl player or tape deck.

sorry but also for soundstage reproduction ? this is very difficult for digital IMHE ... a very challenging test indeed
Only the best digital sources have a nice 3D effect ... and they are usually very expensive. Very very.

Both are usually a result of uneven frequency response that emphasizes certain frequencies.
If you look at real (and not just perceived) tempo and pace, a digital system gets timing exactly right (down to the stability of the crystal that measured in parts per million), whereas tape or vinyl has huge time variations caused by wow and flutter measured in parts per thousand

I agree for the best digital source but not for the average that use low quality quartz and noise power supplies to power the clock circuit.
A good dac usually starts from 5-600 USD and it is only part of the story.
The transport and even the connection can impair the overall result
Soundstage for me it is the acid test for digital ... very very difficult to get it right because everything must be perfect ... linearity, low distortion, low noise and good timing.
Anyway i would be interested to know the name of one very transparent AD and DA units. That is a tough task indeed.
Digital seems easy but it is not so. A small problem creates huge problems in sound.
Thanks a lot again. Kind regards, gino