USB Ground Lift

[r_w]

I've been experimenting with this for a long time now and can safely arrive at the conclusion that USB cables with a ground lift switch should be available to purchase.

 

Ground is needed for handshake in all of my USB gear, but all of it also works after the ground is lifted (post handshake) and it sounds miles better... it's so smooth, rich and musical, it is pure sonic gold.

 

5V Rail isn't needed for any of my gear, so I always remove that and where anything is USB powered I always inject 5V via a split cable fed from clean LPS.

 

The Aurender now feeds the Intona and W4S Recovery and the Auralic Vega throughout via the 2 data rails only.

 

How do I know that there is not a ground link via the mains?.. nothing handshakes until all ground switches are set to on.

 

I've been told that USB isn't supposed to work without a ground connection, but it does in my system... and it's a big SQ improvement.

 

Ground lift and high quality LPS has been the solution in my system to achieve unbelievable SQ (details in my sig).

 

Just my 2p, hope it helps if anyone wants to experiment.

 

(I don't see any reason ditch USB and go AOIP (until it becomes more domestic), if you can get USB working like mine... but I agree that AOIP is a superb and very promising technology - however all of the 'gorgeous' descriptions of AOIP SQ exactly mirror my system)

 

;-)

[sandyk]

r_w

With many USB devices the shield is connected internally to the 0 volts ("earth") wire.

Where the P.C. is connected to Mains Earth, this may mean that the earth return is less than optimum, and may even be a noisy earth .( Ground loop)

Are you sure that your return path isn't now via the shield ?

 

Regards

Alex

[r_w]

I never connect the shields if you mean the physical metal covers? I only ever connect the pins and tape them so that the wires can't touch the metal covers.

 

Balanced silver cable with only pins connected and a split in the ground braid that is bridged with a switch.

 

Have I understood you properly?

 

Also, I'm pretty confident that there is no mains earth path, which is why no handshaking takes place until the USB ground switches are on. To be fair even if there was a mains earth path, it sounds considerably better in any case.

 

Would be interesting to get some knowledgeable feedback on this.

 

 

 

r_w

With many USB devices the shield is connected internally to the 0 volts ("earth") wire.

Where the P.C. is connected to Mains Earth, this may mean that the earth return is less than optimum, and may even be a noisy earth .( Ground loop)

Are you sure that your return path isn't now via the shield ?

 

Regards

Alex

[sandyk]

I've been told that USB isn't supposed to work without a ground connection, but it does in my system...

 

But your system isn't a P.C. which needs to see an earth return. If the resistance of the earth exceeds a certain limit, the connection may drop out after a while. With the Peachtree DAC it for example, the connection would start to be lost at around 50 ohms in the earth return.

[mmerrill99]

This may well work correctly in that only USB handshaking needs a single ended electrical connection the so-called Single Ended Zero (SE0) condition, where both lines are low, can be detected - this needs a ground reference.

 

For data communication only differential signalling is used i.e the difference between D+ & D- where differential "1" is D+ high & D- low whereas differential "0" is D+ low & D- high. There should never be a Single Ended One (SE1) where both lines are high - it is illegal & a malfunction

 

There may be potential issues of the sending & receiving device grounds drifting further apart than allowed for the common mode voltage range of the USB receiver chip i.e the differential signals are designed to operate within certain voltage limits with respect to local ground - if the grounds are disconnected then the sending a receiving grounds may drift apart enough that the differential signals are no longer in their specified range. For high speed USB it's currents that are used in signalling. This ground drift is probably not going to be an issue in well designed systems although measuring the ground potential between the sending & receiving devices might be warranted?

[r_w]

Thanks for your comments above guys. I really recommend experimenting with this... obviously YMMV but it's a revelation in my system.

 

:-)

[mmerrill99]

Can you tell us about the sound with/without ground connection?

As this is the same issue as Abtr's thread where he thought it was jitter in USB audio that was his problem but now knows it's ground noise.

Can you measure any ground voltage differences between the sending & receiving USB devices? I don't expect you will as they will likely be small.

[r_w]

I wouldn't know how to measure ground voltage, I'm a bit of an unqualified chancer when it comes to electronics.

 

;-)

 

The sound is considerably fuller, richer and more 3D. It was excellent before but somewhat less luscious. I had done experimenting before with ground lift when I had a Regen to positive effect. Then moving to Intona and W4S Recovery introduced more boxes and were both sensitive to compatibility (both had to have manufacturer revisions to work even with the 4 USB rails connected).

 

SQ now is borderline ridiculous and I can't really see what could be significantly improved from here.

 

I had been thinking about making a switched USB cable for some time (I even asked Curious to make one but he didn't see the merit in it) as I had a hunch that ground was needed for handshake but maybe not for data transfer, otherwise it wouldn't explain why the Regen worked without ground. I know that previously my mains was providing an earth path, but upgrading to Swagman Labs SE LPS's provided mains ground isolation, meaning that I had to re-ground my USB cables. I tried one yesterday and was impressed and this morning I removed my Curious Link and put another DIY switched cable in, thus removing ground from the entire USB chain.

 

My guess is that 'the crap' uses ground to travel?

[Nikhil]

I have tried a USB cable with the +5v and GND removed.

It did not work and I agree with mmerrill99 that you need to have the GND.

 

I finally settled on a split USB cable (Dual head) which I would recommend.

The sound had a much better foundation with a sense of solidity in the sound stage.

The cable with only the +5v removed sounds a little thinner in comparison.

 

[Cornan]

I wouldn't know how to measure ground voltage, I'm a bit of an unqualified chancer when it comes to electronics.

 

;-)

 

The sound is considerably fuller, richer and more 3D. It was excellent before but somewhat less luscious. I had done experimenting before with ground lift when I had a Regen to positive effect. Then moving to Intona and W4S Recovery introduced more boxes and were both sensitive to compatibility (both had to have manufacturer revisions to work even with the 4 USB rails connected).

 

SQ now is borderline ridiculous and I can't really see what could be significantly improved from here.

 

I had been thinking about making a switched USB cable for some time (I even asked Curious to make one but he didn't see the merit in it) as I had a hunch that ground was needed for handshake but maybe not for data transfer, otherwise it wouldn't explain why the Regen worked without ground. I know that previously my mains was providing an earth path, but upgrading to Swagman Labs SE LPS's provided mains ground isolation, meaning that I had to re-ground my USB cables. I tried one yesterday and was impressed and this morning I removed my Curious Link and put another DIY switched cable in, thus removing ground from the entire USB chain.

 

My guess is that 'the crap' uses ground to travel?

 

I am all with you r_w! I have used external grounding (Entreq grounding boxes) for several years to clean the ground from noises/crap with great results. Especially good results for USB audio (digital chain) in my experiance...but also very awarding in the analogue side (just important to keep digital and analogue separated with external grounding). Ground is the return path of noises as well. Ensuring a good ground connection and making sure to avoid ground loops is extremely awarding.

Personally I cannot listen to music without external grounding anymore. Everytime I try without it the music becomes dull in comparison. My next mission is dedicated mains, a proper star ground and unshielded 2-wire USB cables (& XLR headphone cables) which I am absolutely sure is the way to go. I'll just need to figure everything out before I make it real!

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[r_w]

Hi Nikhil, Did you try removing GND after handshake when you tried?

 

 

I have tried a USB cable with the +5v and GND removed.

It did not work and I agree with mmerrill99 that you need to have the GND.

 

I finally settled on a split USB cable (Dual head) which I would recommend.

The sound had a much better foundation with a sense of solidity in the sound stage.

The cable with only the +5v removed sounds a little thinner in comparison.

 

[mmerrill99]

I have tried a USB cable with the +5v and GND removed.

It did not work and I agree with mmerrill99 that you need to have the GND.

Except, I didn't say that

[Cornan]

I have tried a USB cable with the +5v and GND removed.

It did not work and I agree with mmerrill99 that you need to have the GND.

 

I finally settled on a split USB cable (Dual head) which I would recommend.

The sound had a much better foundation with a sense of solidity in the sound stage.

The cable with only the +5v removed sounds a little thinner in comparison.

 

 

AFAIK there are no Y-split USB cables without shielding around. They are not made to be disconnected from 5v but to be connected to USB power (5v BPS or LPS). If you dissconnect a Y-split cable you will for sure experiance a ground loop running through the shields and/or dielectric materials. It could potensially harm the connected device as well. Several Regens have been reported damaged this way according to Superdad.

If you want to disconnect the power and GND unshielded cables is your only option...and without a star ground a 2-wire USB cable will not work properly.

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[Cornan]

Here is a very good thread to read for more info regarding USB ground http://www.computeraudiophile.com/f10-music-servers/vbus-isolator-sbooster-26228/

Read both John Swenson and Karin's (SBooster) replies.

 

Ps. Note r_w's posts there as well! Ds.

[Nikhil]

Hi Nikhil, Did you try removing GND after handshake when you tried?

 

No never really bothered to after my initial observations.

 

 

Except, I didn't say that

 

Oops! Sorry about that. But I remember reading John Swenson's comment on the thread referred.

 

 

The normal data on USB IS differential, there are two cases when this is not true and reference to the ground wire is required:

 

Startup: pull up resistors are added to either D+ or D- to specify the bus speed. The resistor is also used by the host to detect when a device is plugged in. This requires a ground connection for the host to tell if the resistor is present. If there is no ground connection, the host will net be able to detect that a device is even plugged in.

 

Single ended data: in this case both D+ and D- are in the same state, both are low or both are high. Without a ground line you cannot tell if both are low or both are high. These states are used for special purposes such as bus reset so are not very common, but ARE important for proper functioning of the system.

 

If you cut the ground and VBUS wires on a USB cable and a DAC still works, then you have some other hidden ground path between the computer and DAC. (like maybe the shield on the cable or a power connection). Without some form of ground connection the computer will never see the DAC.

 

This is actually a good way to see if you a ground loop. If you connect a DAC with a cable that has NO ground connection at all, and it works, you know for sure you have a ground loop somewhere.

[mmerrill99]

Oops! Sorry about that.

No problem!

[r_w]

The problem with his quote is that it does work without a ground loop, but only after handshake with ground connected.

 

This is a case of practice/experimentation rather than theory, which is where most of my SQ gains have come from (plus the use of good LPS) in terms of USB data transfer in my system.

 

If I was building any USB device now I would put a mechanism in to drop ground after handshake, either auto or manual.

 

Also 5V rail generated from anything less than a good LPS or the absolute best SMPS has no place in USB audio.

 

:-)

 

 

Oops! Sorry about that. But I remember reading John Swenson's comment on the thread referred.

 

[sandyk]

If I was building any USB device now I would put a mechanism in to drop ground after handshake, either auto or manual.

 

You keep stating that as if it's a universal remedy. It isn't , and won't work with USB from Windows PCs in general, or probably the later Mac Minis without the connection dropping out sooner or later, and the USB device no longer showing up in "My Computer"

This even happens with USB memory sticks !

 

Yes, USB Audio does benefit from the highest possible quality power supply, with the least coupling back to Mains Earth. Even the type of transformer used in a Linear PSU matters, with Toroidal transformers the worst due to their high capacitance between Primary and Secondary windings. R-Core transformers are better, but the best type is the split bobbin type with their side by side windings.

[Cornan]

Elijah Audio can actually make this cable that works with any setup.

However, I agree with r_w that there should be a ground lift switch on DAC's USB inputs or atleast a "VBus Isolator type of USB adapter" with a ground switch. It will not work for everyone...but will be a sound revolution for many!

 

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[r_w]

Very good, I recommend people try this cable to see if works for them.

 

Elijah Audio can actually make this cable that works with any setup.

[ATTACH]27962[/ATTACH]

However, I agree with r_w that there should be a ground lift switch on DAC's USB inputs or atleast a "VBus Isolator type of USB adapter" with a ground switch. It will not work for everyone...but will be a sound revolution for many!

 

 

Sent from my Nexus 7 using Computer Audiophile mobile app

[r_w]

What's so different with the Aurender? Isn't USB universally the same? If groundless USB audio works after handshake for one device, why not another?

 

Even so, if you can't drop ground from a PC you can still drop it from the rest of the chain (assuming you are using some fixers such as W4S recovery / Intona / etc).

 

Groundless USB audio is travelling across 4 different devices in my chain... That makes it universal enough for me and it never drops out.

 

I'm no evangelist, just saying it's well worth trying it for yourself.

 

;-)

 

You keep stating that as if it's a universal remedy. It isn't , and won't work with USB from Windows PCs in general, or probably the later Mac Minis without the connection dropping out sooner or later, and the USB device no longer showing up in "My Computer"

 

 

[sandyk]

I'm no evangelist, just saying it's well worth trying it for yourself.

 

Did you read what John Swenson posted about how USB functions with a Computer ?

 

Sometimes a USB port isn't recognised after switch on, and you may need to do a Restart for it to show in My Computer.

USB on a Computer can be highly temperamental, and incorrect protocols can result in the USB port being used , BLOCKED as FAULTY after several failed attempts to be correctly identified. It isn't always easy to unblock the marked USB port either, but one method that works for me with W7,8 and 10, is to use System Restore.Even if System Restore isn't successful, the Fault Counters are written back, and the blocked USB port becomes functional again.

[r_w]

OK, so that's probably an OS protection thing, makes sense... but you should still be able to isolate the rest of the USB chain, assuming at least 1 fixer is in-place. Earth from computer to fixer, then lift to DAC (assuming it's not just the VEGA that accepts data only)

 

 

Did you read what John Swenson posted about how USB functions with a Computer ?

 

Sometimes a USB port isn't recognised after switch on, and you may need to do a Restart for it to show in My Computer.

USB on a Computer can be highly temperamental, and incorrect protocols can result in the USB port being used , BLOCKED as FAULTY after several failed attempts to be correctly identified. It isn't always easy to unblock the marked USB port either, but one method that works for me with W7,8 and 10, is to use System Restore.Even if System Restore isn't successful, the Fault Counters are written back, and the blocked USB port becomes functional again.

[4est]

R-core is split bobbin btw, and afaik just as good in this regard.

You keep stating that as if it's a universal remedy. It isn't , and won't work with USB from Windows PCs in general, or probably the later Mac Minis without the connection dropping out sooner or later, and the USB device no longer showing up in "My Computer"

This even happens with USB memory sticks !

 

Yes, USB Audio does benefit from the highest possible quality power supply, with the least coupling back to Mains Earth. Even the type of transformer used in a Linear PSU matters, with Toroidal transformers the worst due to their high capacitance between Primary and Secondary windings. R-Core transformers are better, but the best type is the split bobbin type with their side by side windings.

[JohnSwenson]

Hi all,

this is a topic of intense research right now, some proper solutions for this will (hopefully) be coming out soon.

 

If you have been following some of my posts recently a lot of what seems to be related to the last significant chunk of SQ problems seems to be related to power supply leakage current ( see some of my other posts for descriptions of what this is) forming ground loops through various parts of the audio system.

 

The leakage current from the computer seems to be a major part of this in many systems (note: not all), the loop frequently flows through the ground of the USB cable hence blocking the ground in the USB cable CAN make things much better.

 

But not always. I'm working with one person whose system got much worse when the USB ground was blocked, the loop wound up going through the preamp instead which made it much worse sounding.

 

On USB and grounding in general, a device will not connect to a host unless there is a ground, both from SE0 and the pull up resistors that signal connection and bus speed. It is true that after the connection is established the ground is not actually needed for many USB transceivers, BUT there are some big issues with doing that.

 

Two truly ground isolated systems can develop a LOT of voltage between them, primarily from the above mentioned leakage current from the power supplies. It exists for ALL power supplies, even LPS, but SMPS is usually worse. I did some measurements on some systems around my house and found around 50V on most isolated systems I setup.

 

This means that if you setup a computer feeding a DAC with a USB cable with all grounds cut and there is no other ground connections between them, the ground between those two systems will probably have around 50V between them. This means the DAC's USB chip is going to be seeing 50V on its pins, this is NOT a good thing!!!! The fact that a DAC still runs at all under these conditions is due to ESD protection circuits designed to protect against "static" discharges, unfortunately these circuits are only designed to be used for very short duration events, not continuous large DC offsets. They are protecting the USB receiver from being fried, but they will not continue to do so for a long time. Eventually they will fail and PFFT! your DAC goes up in smoke.

 

There are some proper solutions being worked on right now do deal with these issues, leakage current, ground loops in general etc, and in particular with regards to USB.

 

If you want to experiment with USB ground isolation techniques be aware that you might be risking damaging your equipment, only you can decide if the SQ improvements are worth the risk. If you do have significant voltage difference between grounds the failure of the ESD protection circuits may happen in a couple weeks, 6 months, or maybe even as long as a year, but it will eventually happen.

 

How do you measure it? You need a battery operated meter (or scope) that has a decent bandwidth on its AC setting. This voltage is AC, not DC, so use an AC range. Do not use a meter or scope that plugs into the wall, this will generate its OWN leakage current which will completely mess up anything you are trying to measure. I personally use a scope with a differential probe run off batteries.

 

John S.