• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Cleaning USB for Bus Powered Audio Devices: Discuss

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,296
Likes
1,583
Location
Le Mans, France
@L5730 so basically the GND is used only at the DAC start up, right ? If yes :
PC on ---> DAC always connected or do you plug it out ?
PC off ---> DAC off ---> PC on ---> DAC on with GND closed than GND open
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
Yeah, the USB ground is only required for the initialisation of the device.

For me, I do not leave the DAC connected, I only plug it in when needed. I am thinking about building a more bespoke circuit for my needs, as soon as I know that what I am doing with respect to adding external power and cutting USB grounds is correct.

So, everything is hooked up, and the switch completes the circuit - which lets the ground from PC connect to DAC and external PSU.
(1) Turn on the external PSU, and the DAC will power up, the PC will talk to the DAC and it will display a sample rate - the good thing with the D10 is that we know it's initialised when we see the sample rate.
(2) Set the switch to the other position, breaking the ground to the PC.
The hum goes away, the DAC still functions.

Play music, switch sample rate etc. it's fine, no noises.

If your computer goes to sleep, or if you turn off the PSU or disconnect the DAC or reboot your computer - anything that would require re-initialisation of the DAC, then the ground will need to be reconnected, else USB handshaking won't happen and it won't do anything.

Please, and I cannot stress this enough, do not go and build this cable and use it without understanding what it does or doesn't do.
I am still trying to work out if I may be causing damage with current rushes and voltage bumps and other nasty things. There are no components at all here, no filtering or prevention or any of that, and as such, unless someone who knows how things work can confirm that this isn't bad for the USB port or the DAC, then I can't tell people that it's a good solution.
My findings are just that it did work, it did fix the noise, and that my USB port and DAC still both function fine. But I only tested for a few minutes.
 

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,296
Likes
1,583
Location
Le Mans, France
@L5730 thank you, however I am currently using my simpler layout with the battery power bank to power my KORG DS-DAC100 without any problem. The DAC does not need to be initialized, it powers up when I power up the battery pack. The data cable is always connected to the PC in either state the latter is, ON, OFF, or Stand-By.
But your layout is nice implementation, which is not going to jeopardize mine.
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
@graz_lag So you have Data +/- only to the Korg DAC from your PC? What about shield?
I think the ground connection of some form or other is required by USB to actually make the handshake work, although saying this, I think there are 'audiophile' USB cables that are data only, no ground, no shield.
 

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,296
Likes
1,583
Location
Le Mans, France
@L5730 just a second ... I am using a "Y" USB Cable I bought from Cooler Castle Ltd., UK.
I do not know how it is wired inside, it is used for application when an USB device needs to be powered from two ports, for example.
The thing I know for sure I have interrupted the contacts #1 & #4 in the DATA A-connector that goes into the PC.
They call it "Cables to Go 2 meters, USB 2.0 One B Male to Two A Male Y-Cable".
www.coolercastle.com
£7.39 on July-2017

Edit : the DAC does not power up unless I power up the battery pack, so I do not see how different from my basic layout you drawn the actual wiring might be ...
 
Last edited:
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
That site is no longer up.
But searching gives this on amazon:
https://www.amazon.co.uk/C2G-81578-USB-Male-Y-Cable-Black/dp/B001U3ZP1K

Ah, yeah, you disconnected GND and VBus, but there is still a shield connected.
Although the shield won't be connected to anything in the cable itself, it may be connected to ground in the DAC, and is connected to a ground in the PC for sure.
I would not be surprised to find that USB enumeration is occurring through the shield in your instance.

You could test, by taking some thin tape and wrapping it around the 'data only' connector USB A plug outside. This would insulate the plug (and shield) from the PC.
 

graz_lag

Major Contributor
Joined
Oct 13, 2018
Messages
1,296
Likes
1,583
Location
Le Mans, France
@L5730 ... "and wrapping it around the 'data only' connector USB A plug outside" ... you mean data only after my disconnection of pin #1 & #4 correct ?

Yeah, my Y Cable is exactly the one you pulled out from Amazon.
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
Yes, that's what I mean.
Take the USB A plug that goes into the PC, the one with 1 and 4 disconnected, and insulate the outside of the metal plug so it can't make contact with the socket on the PC. This way Data + and Data - only are the connection from the PC.

A USB cable can actually have 6 possible connections.
VBus = +5v Power
GND
Data +
Data -
Shield = linked to ground somewhere
ID = for on-the-go connections
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Have you tried the following:

Instead of the switch in the used circuit use a resistor. As in this circuit:
index.php

Try to determine the value needed for the D10 to initialize and see if the common mode current is low enough that you aren't bothered by it.
Use a 100 Ohm or 1k potmeter and see what value is needed for proper operation.
It may very well be that this is the simplest (and cheapest) solution for your specific case.
You already have the cables, just a matter of seeing what a resistor does.

The common mode current runs through the RCA and USB ground which is low impedance. This causes a small voltage to form over the internal ground in the D10 which becomes a very small voltage which is heard.
The lower that current the lower the noise level.
Current is lowered by adding resistance (isolators add Giga Ohms).
Consider a voltage of 1V runs through 1Ohm, 1A of current is flowing. With 100 Ohm 0.01A is flowing.
This is a 40dB reduction.
100 Ohm is nothing for the USB circuit and may operate normally yet the common mode current may be lowered enough to become inaudible.
 
Last edited:

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
I wonder how tough it might be to have a chip auto-cut the ground or both VBus and ground when enumeration is complete?
I built such a device based on a Texas Instruments TUSB213 chip.

The trouble with cutting the ground is that the data lines are terminated with 45 Ω resistors to ground. This means that if some other ground return path exists, some signal current will flow there, causing an increase in 8 kHz packet noise on the analogue DAC outputs. However, if the PC and DAC are otherwise completely isolated (e.g. a laptop running off battery), cutting the USB ground can have a positive effect since it becomes impossible for any USB signal current to escape the transceiver and pollute the local ground plane on its way back to the PC.
 

mansr

Major Contributor
Joined
Oct 5, 2018
Messages
4,685
Likes
10,700
Location
Hampshire
I think the ground connection of some form or other is required by USB to actually make the handshake work,
Correct, devices indicate their presence by pulling one of the data lines high. Without a ground connection, this isn't possible. If a device is detected without the ground being connected, it means there is another ground path somewhere.
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
I don't have any resistors to hand - although I have some guitar pots around somewhere. Maybe these are totally the wrong type, I'll take a look. So far I have managed to do everything without touching a soldering iron! I'm playing a bodge here.

Interesting thought. So cutting the USB ground, instead of using an audio ground isolator is causing any packet noise to go down the audio route. I never saw a peep at 8k of anywhere on the VST analyser looking at the motherboard ADC. The D10 added no additional noise above the ADC noise floor. Something I should be aware of though, probably bad practice, right?

I have a list of parts for a USB ground switcher, with micro USB for external PSU:
SPDT switch = USB Host GND on/cut
DPDT switch = External PSU +5v and GND on/off
Micro USB socket breakout board = for external PSU
ABS box
Red and Black 16 strand / 0.2mm ~20 AWG / 3A 1kV wire
Less than £8 shipped!

@solderdude
Multimeter resistance:
PC connected to mains (and running) > USB shield
D10 DAC connected to speakers (all turned off), DAC not connected to PC > USB shield
OL - overload, so I am guessing it's a huge resistance, because the D10 is off.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Try it with the speakers on.

Also measure the AC voltage between these points.
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
PC powered on.
D10 not connected to USB, but to active monitors (3 pin mains) turned on.
PC USB shield > D10 USB shield = 1.1 Meg ohm / 0.005 V
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Looks like the current is not flowing through the USB shield but through the ground wire.

Try to do the resistor trick.
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
Seeing the existence of a very few, but all the same, real and for sale "Data Only" USB cables, I wondered what was in them.
There is one here:
https://www.ebay.com/itm/Data-Only-...Audiophile-5-feet-1-5-meter-5ft-/261859695896

"Data only USB cable with the VCC +5V wire disconnected so no power will be passed (no power line noise for higher quality audio).
This version has the ground wire disconnected as well."

So, no +5v and no Ground - how can this even work then?
Note there is no mention of the shield!

I just tested the D10 with a bit of paper over the ground pin on a normal USB plug.
It still turns on - which one would expect it not to as the +5 has no where to go - except that the shield is connected and that is, although a higher resistance, now the return flow/ground/-ve.
In my case, I still get buzz, basically the same although subtly different in grittiness.

For my ground lift box, I am going to want the shield disconnect too, I think.
 
OP
L5730

L5730

Addicted to Fun and Learning
Joined
Oct 6, 2018
Messages
667
Likes
434
Location
East of England
OK, more tests before buying stuff:

As @mansr has confirmed, there is some 8 kHz noise components that appear. I could only see them on an FFT Spectrum analyser VST, and the highest one was at -108dB and not really audible. I can accept that without issue.

The test now only connects Data + and Data - to the PC, power (+5v/GND) come from external PSU. Shield is broken in the middle and insulated.

(1) D10 connected to mains earthed active speakers: DAC powers on, enumerates, music plays.
(2) D10 connected to onboard ADC: All works, no buzz.
(3) D10 connected to 12v DC desktop speakers: No enumeration.
(4) as above, but connect the GNDs in the USB cable: It works.

So, whatever happens, the D10 needs to find a ground of some kind, it isn't enough just for a PSU to power it.
Why is it the case the (3) doesn't work? It doesn't work when connected to an un-grounded laptop either - requiring the USB ground temporarily connected for enumeration. But, the opposite direction is fine, an ungrounded laptop does make the DAC enumerate, when using a proper USB cable - because the 'ground' reference is common between the two?

But really, I suppose the biggest question is:
Is there any possibility that removing the ground, and shield, from a USB cable could cause damage to the Host USB port or other equipment?
If there is a fault, most likely it'll go to ground, which be the equipment that is grounded.

If I am at no real risk of screwing anything up, then I will carry on with this and order a bunch of parts.
Makes me wonder why no one has made something like this for sale, an external powered manual USB ground lift box - it probably breaks all USB rules so that won't help actually selling it - but USB extension cables break the rule too, and they are aplenty.
 

solderdude

Grand Contributor
Joined
Jul 21, 2018
Messages
15,891
Likes
35,912
Location
The Neitherlands
Makes me wonder why no one has made something like this for sale, an external powered manual USB ground lift box - it probably breaks all USB rules so that won't help actually selling it - but USB extension cables break the rule too, and they are aplenty.

Its what the iFi defender does when it detects a ground loop. It just references to the ground on the receiver side.
The other (technically better) solutions are the isolators.
So there are good solutions available, just not as cheap but less problematic in usage.
 

DDF

Addicted to Fun and Learning
Joined
Dec 31, 2018
Messages
617
Likes
1,355
If your PC uses an earth grounded (3 prong) power brick, try a doubly insulated 2 prong power brick. Breaks earth ground while meeting regulatory safety requirements.

I tried a 5v disconnected usb cable and two different usb optical isolators (both based on the analog devices chip) without success before this solved my usb issues (noise and bit hits)
 
Top Bottom