USB related hum problem [SOLVED]

 
After I have got second "instrument" with the same problem, it is time to find a solution...
 
1) TC VoiceLive Play GTX connected to computer with USB, connected with Kawai CN43 from headphone output (1/8'' stereo) to line input (2xTS)
2) Roland TD-11K connected to XENYX QX1202USB (2xTS), which in turn is connected to computer with USB.
 
Both combinations are working fine for me (there are in different rooms, so different computers and the rest) till I connect either of the instruments (e-drums or DP) to computer with USB. The result is the same in both cases: significant noise/hum in middle/hi frequency (not 50Hz).
 
I have tested several USB cables, no difference. I have tested other USB ports (throw hub, without, etc.), there is a difference, but it is still not acceptable. Notebook on place of any of 2 PCs (especially on accu) works fine.
 
Since I conclude I am getting some ground loop (PC->usb->mixer/VliceLive->DP/e-drum->usb->PC) and I do not like an idea to cut ground on USB cable (I have seen many smoked devices at work after using such "solutions"), I am thinking about some trafos:
 
VoiceLive has balances XLR outputs (mic level), so I think about XLR to TS:
http://www.thomann.de/gb/neutrik_na2mjtx.htm
 
TD-11K has unbalanced TS (also mic level? I am forces to use mixer preamp inputs to get on normal one), so the idea is to "save a bit" with:
http://www.thomann.de/gb/behringer_microhd_hd400.htm
 
Does someone had the same/solved such problem?
 
EDIT:
 Solution: Behringer HD 400 helps in both cases.

Thanks for the comment. I think I will try HD400 first.
 
As a piano player, I am more sensitive to the noise on DP. That is why I have thought about Neutrik there, but may be I can somehow test HD400 for piano as well.
 
You are right, my dinner yesterday was more expensive then all these devices together... my wife has enjoyed the dinner, while she is still skeptical about my hobby 

My previous setup :
Quad Capture (unbalanced) -> HD400 (unbalanced) -> Behringer Truth monitors (active).
 
My current setup :
Quad Capture (unbalanced) -->>
-> Nad Amp (driving a JBL Sub woofer)
-> Nad Amp Rec out (unbalanced) -> HD400 (unbalanced) -> Behringer Truth monitors.
 
There is no noise from the Quad Capture headphones, only when Quad Capture -> Truth direct. I wonder if using a pair of balanced 1/4" cable would solve the problem too..?

If it really is a ground loop, the way I always handle it is to use one of those 3 prongs to 2 prongs plug converters (is this an American thing?) on various plugs in the signal chain and check to see where the hum gets louder or softer based on the change.
 
Sometimes it gets MUCH worse, sometimes much better.
 
As long as there is a path to ground, you should be ok, but you don't want to lift the ground if there is only one path to ground. So for example, if you have a power strip plugged into the wall and everything plugged into the power strip, don't lift the ground at the wall. Lift it for the items being plugged into the power strip.
 
ymmv.
 

The Behringer looks like a knockoff of the Ebtech hum eliminator.   I keep an Ebtech on hand in case its needed.

fireberd
The Behringer looks like a knockoff of the Ebtech hum eliminator.   I keep an Ebtech on hand in case its needed.

 
I actually bought the Behringer HD400 together with the Ebtech HD-2-XLR at the same time, in case one doesn't work. While the HD400 has been operating without any problem, the Ebtech gives a small buzz from day one... must be a faulty unit as the buzz came out from both channels.

Grounding in electronic is even more confusing then the meaning of ACT in SONAR.
 
We have several "grounds":
1) "ground" connector on power socket. It should be "real" ground (well connected with the earth). It should not be used to deliver elictricity. It is there just in case the power wire is by occasion touch metal parts we can touch as well, so "overcurrent" somehow switch electricity off and prevent our death. The theory is good, but the practice, especially in consumer area, is far from perfect. I was born in Russia, where outlets had no such connector.
2) "ground" (0) wire on power socket. That is used to deliver electricity (together with the power wire). It is normally "close" to the real ground, so for example we do not feel us bad in case we touch it. Well, if everything is correctly connected and it is not in use by some powerful device, in which case the wire own resistance produce significant potential shift. For that reason, it should not be connected with the "real" ground.
3) circuit "ground", as a reference for other (not power) signals inside the device. Depending on the power supply of the device, it can have virtually any potential compare to (1) or (2).
4) signal "ground" for transferring the signal (for example audio)
5) shield "ground", just to have the shield potential on some stable level.
 
So, for XLR it is quite logical:
* if both ends have the "ground" pin connected to the shield (inside device), it make no different how we connect it in the cable
* in all other cases of it's connection ((3) or (1)) we are looking for troubles (till at one end it is disconnected, inside the device or in the cable)
 
Theoretically, the same is valid for USB.
 
But there is one small problem... all "grounds" except (1) can be different between powered devices. The shield on all devices is (should be) connected such way, that the difference on it can not "kill" the device. It can create hum/noise or even switch off the device, but it should not kill it. From information in the Internet, audio circuits (at least in (semi)pro level) are also prepared for such "surprise". But for USB I am not so sure...
 
I must admit that my background in electronics has not helped me eliminate my hum, so I am waiting for the trafo
 
PS. Sorry for so long post.

Unfortunately no luck with ferrite... Absolutely no difference in my case (tested 2 cables in Piano setup).

Back in the mid 70's, I worked on "cleaning" a recording studio that was near an AM radio transmitter site.  The radio station was coming in on the AC power line, recorders, console, etc.  and headphones.  I was able to get the radio out, or at a low enough level that it didn't matter in everything except the headphone system.  I couldn't get it completely out of the headphones. 
 
Most of what I did was bypass capacitors. There was some shielding but the bypass caps fixed most of it.

Grounding is not mysterious, nor is it voodoo. It is, however, difficult to manage because of the myriad ways manufacturers apply basic grounding rules, and that's complicated by the mix of balanced and single-ended interfaces.
 
The basics:
 
A balanced output provides an equal impedance from each leg to reference ground. There is no requirement that both sides of the line are driven, although the most common topology applies a signal of equal amplitude, and opposite polarity, to the two legs. Signal symmetry has nothing to do with balancing a line, other than you can't have signal symmetry on a single-ended output.
 
A balanced input takes the difference between the two signal lines, which is probably why this is confusing. But it is the impedance balance between the two legs that provides the noise immunity. If each leg is impedance balanced, and noise is impressed on both legs it will cancel out. But in order to accomplish this both input pins must also have equal impedance to the reference ground. The cool thing is that the reference does not have to be remote!
 
Another thing people trip over - the shield on a balanced line is not required to transmit, and recover the signal. In fact, it is a really bad idea to connect the shield anywhere other than the chassis itself. If the shield enters the device you have problems - most commonly referred to as the "Pin 1" problem.
 
And now you probably see the problem... a single ended interface usually uses the shield as one of the signal conductors, which means that it necessarily has to enter the chassis. Well, ok, there are ways around even that, but they are not terribly effective.
 
The really bad news? It takes only one poorly implemented interconnection in an entire system to drag it into noise heaven. And USB is very often that culprit.
 
The good news is that most audio manufacturers seem to have figured out how to properly handle USB. My new Presonus 1818VSL exhibits no pin 1 problems.
 
There is a lot more to audio interfacing, but my fingers are getting tired, and I sense many eyes are glazing over.