AC Filtering, Grounding Boxes, Linear PSU and Balanced Power.

[Cornan]

Have to add here that the USB cable with filter is currently configured as follows for testing purposes so it's not an ideal build:

 

- The filter is on a small Proto board in the air

 

- No shielded enclosure

 

- Power wires soldered at each side of the filter are in the air too, no shielding

 

- Components could be better isolated from vibration as well,

 

- Only the data lines are twisted and double-shielded

 

- The power lines are twisted but there's no double-shield

 

- The power line pair and the data line pair are physically separated by about 10 cm throughout the length of the cable.

 

So far, I haven't been able to listen with a physical VBus disconnection as if I do that, the DAC's high impedance sensor interprets it as a 'disconnect' event.

 

My guess is that if you remove the double shielding on the data wires and make sure that the 5v wire and GND is on a safe distance (atleast >4 times the diameter of the wire) you will gain some SQ rating points. Shielding have the tendency to dampen the sound of USB cables and make it sort of "dull" IMO.

 

I agree with @bogi that you´ll need to inject 5v battery power into your DAC USB port....but also not to be that afraid to disconnect the GND (after handshake)...you just have to do it the right way...and in a safe way.

Think about it for a moment. If you make sure NOT to use any shielding (ground connected) in ALL of your ICs, supply ALL your DC powered devices with battery power (always disconnected from mains before you listen to music) and make sure to have only one patch to AC safety ground (via one single power strip) there is nothing to be worry about lifting the GND anywere in the chain. There is simply no other path to ground than via the AC safety ground. If you want to be sure you could always measure the AC currents (NOT DC) in the USB ground wire. I was afriad too before I pulled the trigger...and my hands was shaking that day when I cutted off the GND wire of my USB cable. Lately I even got a Tech IT (floating secondary) connected to my one and only AC mains path...Wow!

 

Anyway, I just encourage you to take that path...but to take that path step-by-step in a safe way. Just a personal suggestion & a friendly advice!

[One and a half]

What does the forum have in mind to control the leakage currents in the shield of the USB cable? This is where most if not all the noise is coming from.

[Cornan]

What does the forum have in mind to control the leakage currents in the shield of the USB cable? This is where most if not all the noise is coming from.

 

Entreq have done this for many years...but the sad part is that it requires their own cables (with drain wire) and grounding boxes...and it do not remove it completely...it just reduces it. I personally beleive that the best way is to remove the shielding completely and use twisted wire configuration possibly with greater gauge wires and address the RFI/EMI at the plugs ends were the interfearences are worst. Same thing applies to any type of cable/IC...even AC mains cables IMHO.

[YashN]

I am using Y split cable, where computer VBUS line is permanently disconnected from the DAC and instead of that +5V from LPS is permanently connected to the DAC USB port. I used more DACs that way and I never experienced troubles. That's the AQVOX type of solution, although I bought LPS and split cable separately.

Grounds of computer USB port, DAC USB port and LPS are connected together, so some way of filtering computer noise on ground line would be useful.

 

I spent most of last night reading about a possible solution and came to a similar conclusion.

 

One of the things that was blocking me was I don't know much about the handshake. I assumed that since the VBus disconnection made the DAC go dark from the computer's perspective that there would be no way to disconnect it.

 

So today I removed the filter from my DIY cable and tested the injection of alternate power from an iPhone charger - not ideal since it's also an SMPS - put the Nano on internal battery, connected the power, then connected the computer side.

 

Success!

 

The highs are back in force as with the default computer connection, but without the glare, and a lot of micro-details revealed.

 

Sweet!

 

So we now have GND disconnection + VBus disconnection.

[YashN]

My guess is that if you remove the double shielding on the data wires and make sure that the 5v wire and GND is on a safe distance (atleast >4 times the diameter of the wire) you will gain some SQ rating points. Shielding have the tendency to dampen the sound of USB cables and make it sort of "dull" IMO.

 

Been testing this kind of thing literally for years. See an update from last year about how my DIY USB cable is geometrically - the power lines and the data lines were always physically separated in two sets of double-shielded twisted pairs. So GND disconnection isn't new to me, it just wasn't made permanent in my DIY USB cable because I needed it for the charge periods and it's only more recently that I added the switch on a more permanent basis. But based on feedback by iFi, I wanted a safer, more permanent way of having great sound and doing isolation until I re-test a proper HQ Player + NAA configuration.

 

I still have some further testing for shielding to do though.

I agree with @bogi that you´ll need to inject 5v battery power into your DAC USB port.

 

...but also not to be that afraid to disconnect the GND (after handshake)...you just have to do it the right way...and in a safe way.

Think about it for a moment. If you make sure NOT to use any shielding (ground connected) in ALL of your ICs, supply ALL your DC powered devices with battery power

 

Done today for alternate power injection while the DAC is internal battery-powered, so far so good! I still have normal ICs though and I really don't like the RCA architecture of having shield as return...

 

Some food for thought:

 

If the whole USB connection wasn't built ideally for async audio streaming, do you the actual connectors are made in the ideal way?...

[lmitche]

I spent most of last night reading about a possible solution and came to a similar conclusion.

 

One of the things that was blocking me was I don't know much about the handshake. I assumed that since the VBus disconnection made the DAC go dark from the computer's perspective that there would be no way to disconnect it.

 

So today I removed the filter from my DIY cable and tested the injection of alternate power from an iPhone charger - not ideal since it's also an SMPS - put the Nano on internal battery, connected the power, then connected the computer side.

 

Success!

 

The highs are back in force as with the default computer connection, but without the glare, and a lot of micro-details revealed.

 

Sweet!

 

So we now have GND disconnection + VBus disconnection.

YashN,

 

I am running my microIDSD with an AQVOX cable injecting power from an LPS1. This way I don't need to use the internal battery power. The shield needs to be connected to the source computer for handshake. I have yet to try disconnecting ground once done.

 

It sounds fabulous!

[YashN]

I am running my microIDSD with an AQVOX cable injecting power from an LPS1. This way I don't need to use the internal battery power. The shield needs to be connected to the source computer for handshake. I have yet to try disconnecting ground once done.

It sounds fabulous!

 

I was just on their website yesterday again and the iDSD Nano is in the supported DAC list so I figured I could try my experiment today.

 

Interesting info about the shield though. I suppose if the shield is disconnected then the DAC wouldn't be recognised at all. I do have a future experiment with modified shielding planned.

 

Would you say the LPS-1 sounds better than the internal battery?

 

Enjoy!

[YashN]

Rectified for omissions and edited for clarity:

 

If the whole USB connection wasn't built ideally for async audio streaming, do you believe the actual connectors are made in the best way possible?...

[Cornan]

YashN,

 

I am running my microIDSD with an AQVOX cable injecting power from an LPS1. This way I don't need to use the internal battery power. The shield needs to be connected to the source computer for handshake. I have yet to try disconnecting ground once done.

 

It sounds fabulous!

 

The shield does not need to be connected and are not used for DAC handshake. GND wire is used for handshake. Just think about it for a moment. Many USB cable have the shield connected at one end only and some USB cables are unshielded as well...but still works. If not the mocroiDSD is very strangely constructed (which I doubt) the handshake will work without the shield. That GND is used for handshake is confirmed by JS on the Sbooster VBus thread as well.

 

Lifting GND and keeping the shield connected to ground is a really bad idea. If your shield is connected to ground at both ends you cannot lift the GND since you are still connected to GND via the shield connection and you will end up without GND lift and decreased SQ.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[Cornan]

Rectified for omissions and edited for clarity:

The USB plugs is really bad constructed for audio purposes. The problem is that the wires are too close to each other inside the plug and a lack of good isolation. If you look inside some branded manufactorers plugs they are sealed with silicone (like Apple, Curious cable etc). It have two purposes. Hide their ideas but also to isolate the inside since silicone is a great material for that purpose.

You also need a good outside material for the plugs to isolate and minimize vibrations. Natural materials like wood are always best. Last but not least the metal barrel needs a firm fit.

 

However, these are only things that make bad better. To make USB perfect the whole plug needs a complete rework IMO.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[One and a half]

This is a great book to read, now in its fourth revision. Covers everything in audio to move one signal from one device t'other using shielded cables and avoiding ground loops and creating antennas. 210 pages of practical fact.

 

Grounding and shielding techniques - Ralph Morrison

[Middy]

You have 2 simple ways of tightening USB connectors... Pins can be gently lifted up to add more force/ spring.

The Metal ground has 2 push plates... They again can be pulled out and give more force to the incoming male Usb 2 holding it more steady...... I'll try and do pictures later if I can....

 

Sent from my SM-G900F using Computer Audiophile mobile app

[Cornan]

@YashN There is one more thing I want to add. If you want to keep the shield and lift the GND I suggest you do it the same way as @r_w have done it, ie. connect both the shield and the GND wire to a DIP-switch or similar. If you make sure to add the switch at the USB plug source end you will in reality have a semi-floating shield with GND lift. IMO a floating shield are better than a ground connected shield....but no shield is better than a floating shield IME as long as you assure a good distance from other sources of disturbences and make sure to use natural isolation materials for the data wires (like unbleached silk or cotton) and keep the cables atleast shorter than 500mm in lenght (but the shorter the better). I know r_w is using a Intona (industrial) so atleast the USB cable after his Intona is galvanicly isolated.

 

Personally I am curious how a non-conductive USB barrel would sound like om my own cables. One of my next experiments will be to cover the outside barrels with electrical tape to reduce capasitive coupling to the pins.

[tapatrick]

YashN,

 

I am running my microIDSD with an AQVOX cable injecting power from an LPS1. This way I don't need to use the internal battery power. The shield needs to be connected to the source computer for handshake. I have yet to try disconnecting ground once done.

 

It sounds fabulous!

 

Like lmitche, I'm using an LPS-1 to inject the 5v into my Aqua La Voce Dac. I've made a customised DIY USB cable using short section of silver wire for the power from LPS-1 and cutting the shield at Dac end. The sound was good before this but is now an incredible improvement!

 

Unless I'm completely wrong the ground does not run through to mains earth because of the isolation of power of the LPS-1.

 

 

Sent from my iPhone using Tapatalk

[r_w]

 

... above is what I do using balanced cable, which is a twisted pair inside a shield (cut the shield and solder a switch across it).

 

easy.

 

@YashN There is one more thing I want to add. If you want to keep the shield and lift the GND I suggest you do it the same way as @r_w have done it

[Middy]

Quick pic of USB 2 B CONNECTOR

 

With a pin or any small tool, you can stretch bent out the 2 silver ground tabs for better grip on the incoming male connector..

 

The 4 gold pins gently lift up ...

 

You could make an anti vibration seal but I will leave that for later......

 

 

 

 

Sent from my SM-G900F using Computer Audiophile mobile app

[lmitche]

I was just on their website yesterday again and the iDSD Nano is in the supported DAC list so I figured I could try my experiment today.

 

Interesting info about the shield though. I suppose if the shield is disconnected then the DAC wouldn't be recognised at all. I do have a future experiment with modified shielding planned.

 

Would you say the LPS-1 sounds better than the internal battery?

 

Enjoy!

I have a listening session today on just this topic. More later.

[lmitche]

The shield does not need to be connected and are not used for DAC handshake. GND wire is used for handshake. Just think about it for a moment. Many USB cable have the shield connected at one end only and some USB cables are unshielded as well...but still works. If not the mocroiDSD is very strangely constructed (which I doubt) the handshake will work without the shield. That GND is used for handshake is confirmed by JS on the Sbooster VBus thread as well.

 

Lifting GND and keeping the shield connected to ground is a really bad idea. If your shield is connected to ground at both ends you cannot lift the GND since you are still connected to GND via the shield connection and you will end up without GND lift and decreased SQ.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

Cornan,

 

With ground and power coming from the LPS 1, the microIDSD sees the power as the power led changes color when connected as appropriate. Nevertheless, the device does not show up in the windows device manager without the lps1 ground connected to the shield.

 

So Cornan, as this is the behavior observed here, I guess by using your logic the IFI microIDSD is strangely constructed.

[Cornan]

Cornan,

 

With ground and power coming from the LPS 1, the microIDSD sees the power as the power led changes color when connected as appropriate. Nevertheless, the device does not show up in the windows device manager without the lps1 ground connected to the shield.

 

So Cornan, as this is the behavior observed here, I guess by using your logic the IFI microIDSD is strangely constructed.

 

Hmm...I think we are talking about different things. I am talking about USB cable shield...but you are mensioning LPS-1 ground connection to shield here.

 

If the LPS-1 will cancel out the IFI microIDSDs battery power the LPS-1 is the reference to ground. I do not know that much in detail about the construction of the LPS-1 itself...but I´ll guess it is galvanicly isolated and needs the shield connected for some reason.

Personally, I would try the microIDSD with clean battery injection and unshielded USB cables with GND lift to see if it works and how that compares. One of the advatages with battery supplies IMO is the ground reference...as long as it is disconnected to AC mains.

[YashN]

The USB plugs is really bad constructed for audio purposes. The problem is that the wires are too close to each other inside the plug and a lack of good isolation.

 

That's what I concluded as well. It's funny how revisiting the whole of the initial USB connection shows how just re-using the initial design for robust data transfer but now transposed for the real-time async audiophile domain has other issues.

 

Indeed, I think the 'wires' (I'd say pins instead of wires here) are too close, but more often than not, we're just plugging in the default USB cables we're getting off-the-shelf and the default hard adapters as well, which have, internally the same issues.

 

The very internal build of a flat USB connector is problematic.

 

 

If you look inside some branded manufactorers plugs they are sealed with silicone (like Apple, Curious cable etc). It have two purposes. Hide their ideas but also to isolate the inside since silicone is a great material for that purpose.

 

I don't think they hide anything fancy inside and that's the rub, it's just a mechanical + esthetic thing, and with iFi, they mention they use a full aluminium shell for shielding.

 

However, these are only things that make bad better. To make USB perfect the whole plug needs a complete rework IMO.

 

I quite think the same.

[YashN]

The shield does not need to be connected and are not used for DAC handshake. GND wire is used for handshake. Just think about it for a moment. Many USB cable have the shield connected at one end only and some USB cables are unshielded as well...but still works. If not the mocroiDSD is very strangely constructed (which I doubt) the handshake will work without the shield.

Lifting GND and keeping the shield connected to ground is a really bad idea. If your shield is connected to ground at both ends you cannot lift the GND since you are still connected to GND via the shield connection and you will end up without GND lift and decreased SQ.

So, this really is connected (ah ah ah) to how the shielding and connectors are implemented (as well as computer Host and DAC USB Input ports) in my current configuration.

 

The test yesterday (it's playing the DALI CD Vol. 3 "OP8 - Leather" with great sound via battery mode, with the iPhone charger connected to mains) has the computer power lines totally disconnected from the DAC.

 

But the question is 'is it really?' since handshake does occur.

 

This brought me to think about the shielding and terminators, which would be cause for another experiment.

 

It's probable that in my current configuration the handshake is occurring via the shield (only the data lines are currently double-shielded).

 

If that's the case, it would be good to know whether there are long-term issues with this.

 

In any case, it sounds really good like that so far.

[Cornan]

So, this really is connected (ah ah ah) to how the shielding and connectors are implemented (as well as computer Host and DAC USB Input ports) in my current configuration.

 

The test yesterday (it's playing the DALI CD Vol. 3 "OP8 - Leather" with great sound via battery mode, with the iPhone charger connected to mains) has the computer power lines totally disconnected from the DAC.

 

But the question is 'is it really?' since handshake does occur.

 

This brought me to think about the shielding and terminators, which would be cause for another experiment.

 

It's probable that in my current configuration the handshake is occurring via the shield (only the data lines are currently double-shielded).

 

If that's the case, it would be good to know whether there are long-term issues with this.

 

In any case, it sounds really good like that so far.

 

I think you might missunderstand the DAC handshake a bit?

Quote: But the question is 'is it really?' since handshake does occur.

The handshake does nothing for SQ. It is just how a DAC checks that a USB device is connected by saying "hello" (ie. What a handshake is called as well). The "hello" (handshake) only takes a split second. It is the time from you press "play" until the music starts playing (or rather finding the device). It is first when the music have start to play you are able to lift the GND and enjoy the incresed SQ. Every time you listen to music you'll need to go through the same proceedure: press play/music starts/disconnect GND....and....press stop/music stops/re-connect GND.

 

The problem with shielding and handshake is that as long as a shield is connected to ground you cannot lift the GND wire since ground currents will flow through the shield instead...and the shield is a worse ground path than the GND wire. To effectively lift the GND both shield & GND have to be lifted at the same time...or...to use unshielded wires and just lift the GND wire.

 

I hope that helps!

 

Micael

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[YashN]

I think you might missunderstand the DAC handshake a bit?

Quote: But the question is 'is it really?' since handshake does occur.

Nope, you misunderstood that part: it meant 'but is it really disconnected since the handshake does occur?'

[YashN]

The handshake does nothing for SQ. It is just how a DAC checks that a USB device is connected by saying "hello" (ie. What a handshake is called as well). The "hello" (handshake) only takes a split second. It is the time from you press "play" until the music starts playing (or rather finding the device).

Nobody said otherwise.

[YashN]

It is the time from you press "play" until the music starts playing (or rather finding the device). It is first when the music have start to play you are able to lift the GND and enjoy the incresed SQ.

 

But can it work if we disconnect before playing the music?

 

It worked for me but the extent of the disconnection is now to be questioned though.

 

Every time you listen to music you'll need to go through the same proceedure: press play/music starts/disconnect GND....and....press stop/music stops/re-connect GND.

I don't find a need to do that in my tests. Here again, the extent to which what we're thinking it is a disconnection when the GND is cut off should be examined as alternate routes are possible, the shield and terminators being one.