USB noise - USB isolator or new audio interface?

[toroto]

Hi,
I have got a hiss noise issue coming from the speakers with my current setup that I use for music production (recording guitar and voice), music listening and calls. I think it is due to a USB power leakage but I would like confirmation from experts
Attached file is a recording of the hiss with a smartphone at ~5 cm from the speaker at normal listenning volume.


Context :
My setup is : Computer (description here) -> Focusrite 2i4 2nd GEN (using lastest ASIO drivers) -> Yamaha HS5 and Sennheiser H560s.

I don't hear the hiss with the HD560s plugged on the headphone output. only with the monitors.
All my powered equipement (both monitors and computer, screen) are connected to the same power trip. None of the power sockets in the room are grounded.


Tests :

• Change USB cable -> Same issue
• Unplugged all usb device except the Focusrite -> no change
• unplugging sound card -> no more hiss
• Changing sample rate & buffer size on focusrite configuration when ableton started -> no change
• Increasing sample rate or decreasing buffer size when Ableton Live is on -> sound becomes higher pitched
• Try with my old laptop on battery -> no more issue
• high graphic card usage (tested witch benchmark) -> causes quircks

Questions :

Do you think my assumption is correct ?

If so it seems that a solution could be an USB Isolator. However the "Topping HS1" is currently 70€, "Hifime High-Speed USB Isolator v2" to while I bought the sound card for 100€ so I wonder if it really makes sense... Cheaper alternatives (10€) like this exists but it seems they also introduce their own issues. Javier used this solution for a DAC, however I wonder if it will impact recording/latency with the interface ?

As Amir wrote about the Modi, a "Better designed DACs don't have that problem." so maybe I should sell the Focusrite and try with an other interface ?

Thank you very much for your help

 

[staticV3]

Do you think my assumption is correct ?

Yes.

However the "Topping HS1" is currently 70€, "Hifime High-Speed USB Isolator v2" to while I bought the sound card for 100€ so I wonder if it really makes sense...

I think that a USB isolator is a good tool to have in your arsenal.
You never know when it may fix a ground loop where just replacing the culprit isn't so easy.

Fwiw, there's another high speed isolator on AliExpress, which you can buy for 53€ shipped: https://aliexpress.com/item/1005001927948863.html

 

[msmucr]

It is quite common issue with certain combinations of USB interfaces and monitors.
I'd lift the ground on cables coming from the interface to your monitors.
To do that, open XLR connectors at speaker side of the cable (male plug). Then break connection on wire coming to pin 1 - https://en.wikipedia.org/wiki/XLR_connector#Three-pin_in_audio_use
You can either desolder that (if you have an iron) or simply cut the wire with pliers. There is quite high chance, it will solve your issue. If not, you can solder the wire back.

 

[msmucr]

Fwiw, there's another high speed isolator on AliExpress, which you can buy for 53€ shipped: https://aliexpress.com/item/1005001927948863.html

Nope, those cheap isolators are 12Mbps only, based on old ADUM4160 isolator chip from Analog Devices. It is unsuitable for vast majority of current USB audio interfaces, which require 480Mbs mode to operate.

Lifting ground is by far easiest and least expensive first step IMO/IME.

 

[staticV3]

Nope, those cheap isolators are 12Mbps only, based on old ADUM4160 isolator chip from Analog Devices. It is unsuitable for vast majority of current USB audio interfaces, which require 480Mbs mode to operate.

Lifting ground is by far easiest and least expensive first step IMO/IME.

I take it you didn't actually open the link?

The isolator I linked to fully supports 480Mbps High speed transmission, just like the Topping HS01 and Hifime Isolator V2.

 

[msmucr]

I take it you didn't actually open the link?

The isolator I linked to fully supports 480Mbps High speed transmission, just like the Topping HS01 and Hifime Isolator V2.

Sorry, my bad.
I've overlooked it, typed it quickly just after I spotted the picture. Come to exactly same looking isolator recently, which was ADUM based. Again sorry for that.
If it really works at 480 Mbit, then it will most likely based on recent TI ISOUSB211 chip.

 

[somebodyelse]

If it really works at 480 Mbit, then it will most likely based on recent TI ISOUSB211 chip.

Or ADUM4165/6 which IIRC was announced around the same time.

 

[pderousse]

Can you try a different sound card? It seems like you narrowed it down right there. When you got no hiss, were you outputting directly from the MB?

 

[msmucr]

Or ADUM4165/6 which IIRC was announced around the same time.

Yes, that could be also an option. I almost forget on that one. Have you seen or used some real-world device with that newer AD isolator? As I recall, there was requirement for external 24M clock for its use with 480Mbit HS mode for internal re-clock.

Generally I think it is very good those ICs for 480Mbit USB isolation became readily available. Before that there was only Intona's custom solution with FPGAs, which was naturally pretty pricey. I wonder if it could find its way into better bus powered audio interfaces and DACs as a standard, because ground loops can be quite common issue there. Especially given the variety of computers and environments, where such devices could be used. I had to solve those issues so many times for my pals, sometimes ground lift is an option, but often unbalanced lines are used (like domestic stuff or say usual DJ gear) so isolation is really necessary. It would be very nice if vendors could think about that from start.

 

[somebodyelse]

Until now I've only seen EVMs for either the TI or AD. I've been waiting for something using them to hit mass market for uses like this.

 

[DVDdoug]

If you're getting USB noise it's usually whine from the computer's power supply. If you get an isolator, get one that isolates power. Data isolation is optional.

 

[toroto]

First of all, thank you everybody for your feedbacks. It is very appreciated.

It is quite common issue with certain combinations of USB interfaces and monitors.
I'd lift the ground on cables coming from the interface to your monitors.
To do that, open XLR connectors at speaker side of the cable (male plug). Then break connection on wire coming to pin 1 - https://en.wikipedia.org/wiki/XLR_connector#Three-pin_in_audio_use
You can either desolder that (if you have an iron) or simply cut the wire with pliers. There is quite high chance, it will solve your issue. If not, you can solder the wire back.

This would remove the ground right ? I am not sure to understand why in theory it could help get rid of the hiss, can you explain ?
I understand that removing the ground could solve a ground loop issue (hum at 60hz) but not how it can help on mine ?

I tried with an "unbalanced" connection :
Desktop computer -> Focusrite (usb powered) -> TS (from TRS output of the focusrite) to TS (in TRS input of the HS5) cables -> HS5 = the hiss is was louder

Can you try a different sound card? It seems like you narrowed it down right there. When you got no hiss, were you outputting directly from the MB?

Today I tried with a Topping E30. Unfortunatly it has only RCA outputs and I have no RCA to XLR or RCA to TRS cable or adapter. However I tried from the Forcusrite to another pair of studio monitors :

• Desktop computer -> Focusrite Scarlett (usb powered) -> TRS to XLR (balanced) cables -> Focal Alpha 65 = No HISS
• Desktop computer -> Focusrite Scarlett (usb powered) -> RCA Cables (unbalanced) -> Focal Alpha 65 = No HISS
• Desktop computer -> Topping E30 -> RCA Cables (unbalanced) -> Focal Alpha 65 = No HISS

Which indicates that the Focal are maybe not sensible to the issue or that the HS5 are maybe be a part of the problem too ?

It could be interesting to test :
Desktop computer -> Topping E30 -> RCA to TRS OR RCA to XLR -> HS5 to check if the E30 helps with the issue. If so maybe i'll use an other audio interface.

Yes.


I think that a USB isolator is a good tool to have in your arsenal.
You never know when it may fix a ground loop where just replacing the culprit isn't so easy.

Fwiw, there's another high speed isolator on AliExpress, which you can buy for 53€ shipped: https://aliexpress.com/item/1005001927948863.html

If you're getting USB noise it's usually whine from the computer's power supply. If you get an isolator, get one that isolates power. Data isolation is optional.

This seems a promising solution. Before to pull the trigger :

1. Does adding an USB isolator between can impact the stability/latency/performance of the ADC as some suggest ?
2. The USB Isolator that StaticV3 linked does not allow for plugging in an external power supply to the isolator, so I wonder if the Scarlett (which is usb powered) will still work. Will the isolation be effective without the "DC filtering" that the HS02 does ?
3. Why isn't this feature included directly by the manufacturers? Does any audio interface include it as standard? I am considering replacing the C, which already has other issues.

 

[KSTR]

Try with my old laptop on battery -> no more issue

--> Classic ground loop issue. The computer is mains-earthed and the monitors are also mains-earthed and you are using unbalanced connections (note that error is created in the unbalanced cable, not in the interface or anywhere else).
Isolation the computer with an USB isolator is the best solution as it fixes the root of the problem.

 

[KSTR]

This seems a promising solution. Before to pull the trigger :

1. Does adding an USB isolator between can impact the stability/latency/performance of the ADC as some suggest ?
2. The USB Isolator that StaticV3 linked does not allow for plugging in an external power supply to the isolator, so I wonder if the Scarlett (which is usb powered) will still work. Will the isolation be effective without the "DC filtering" that the HS02 does ?
3. Why isn't this feature included directly by the manufacturers? Does any audio interface include it as standard? I am considering replacing the C, which already has other issues.

1. No, but of course there is no guarantee. Some USB chip sets have been reported to be problematic when mixed with certain USB isolators.
2. No, if your interface is USB powered you need, well, USB power.
3. USB isolation adds cost and complexity and may not be required for most applications, that's why it is not standard. And it's only a short time that USB isolator chips that work at USB2 high-speed have been available.

 

[somebodyelse]

The USB Isolator that StaticV3 linked does not allow for plugging in an external power supply to the isolator, so I wonder if the Scarlett (which is usb powered) will still work. Will the isolation be effective without the "DC filtering" that the HS02 does ?

That one says it includes a 2W isolating power supply which probably isn't enough to power the Scarlett, although the manual doesn't say how much power it needs. There are others that use an external 5V supply - should be ok on paper but I've not actually used that type.

 

[anmpr1]

Be careful what you plug your USB cable in to. Not that I have (or know anyone who has) one of these. It's getting weird out there...

[updated] Vibrator virus steals your personal information | Malwarebytes

One of our customers found their vibrator was buzzing with a hint of malware.

www.malwarebytes.com

 

[Lambda]

"Better designed DACs don't have that problem."

Well the Focusrite is Better designed and still has this problem.
He just claims its like this but never tests this hypothesis. Actually not many interfaces isolate the USB input for common mode noise.

Your computer has some noise on the USB output and the Topping HS1 is a easy fix for that.

 

[toroto]

First, thanks again to all of you for your contributions

--> Classic ground loop issue. The computer is mains-earthed and the monitors are also mains-earthed and you are using unbalanced connections (note that error is created in the unbalanced cable, not in the interface or anywhere else).

I am not sure to understand why you think it is a ground loop. To me, it sounds more like, as DVDdoug wrote, a "whine from the computer's power supply" :

• The computer and the studio monitors are both connected to the same power strip, which is plugged into an outlet that does not have a ground .
• The noise is not 50hz or 60hz (which seems to be the "hum" symptom of ground loops) but higher. In my case the pitch of the noise is higher and depends on the activity of the PC (scrolling, high CPU/GPU usage, ...)
• The link between the interface and the monitors is a balanced TRS (from the balanced outupt of the scarlett) to XLR cable (to the balanced xlr input of the hs5)

According to Wikipedia : "A ground loop is caused by the interconnection of electrical devices that results in multiple paths to ground, thereby forming closed conductive loops through the ground connections."
In my setup there is no "multiple paths to ground" as there is no path to the ground at all as the outlet into which is plugged the power strip does not have a ground pin.
Or maybe should I understand that the "closed conductive loop" is created between the monitors, the PC (through the interface ) and the power strip and is not related to the outlet ..?
I am confused

not many interfaces isolate the USB input for common mode noise.

Do you have examples of audio interfaces that isolate the USB input? I am already afraid of their price

Isolation the computer with an USB isolator is the best solution as it fixes the root of the problem.

No, if your interface is USB powered you need, well, USB power.

That one says it includes a 2W isolating power supply which probably isn't enough to power the Scarlett, although the manual doesn't say how much power it needs. There are others that use an external 5V supply - should be ok on paper but I've not actually used that type.

I guess I'll have to do some trial & error tests... Here are the potential solutions :

#	Solution	Speed	Power	€
1	Adum3160 based USB Isolator -> USB Hub with power supply	Full	Should be ok	5 + 10 = 15
2	USB 2.0 USB Isolator -> USB Hub with power supply	High	Should be ok	53+10 = 63
3	Hifime High-Speed USB Isolator v2	High	400mA, don't know if it's enough	65
4	Externaly powered USB2.0 USB Isolator + USB power supply	High	Should be ok	75 + 4 = 79
5	Topping HS02 + USB C power supply	High	OK	130 + 4 = 134
6	Optional Cheap Adum3160 based USB Isolator -> Externally powered audio Interface* *not sure it will reject the USB noise though... **The Scarlett already has other issues (balance, crackling knobs, randomly stops) and ideally I would also like an internal dsp,a compressor an more gain)	(Full speed) High	OK	5 + ? =

I'll try with the #1 and will let you know.

Thanks again

 

[KSTR]

I am not sure to understand why you think it is a ground loop. To me, it sounds more like, as DVDdoug wrote, a "whine from the computer's power supply" :

• The computer and the studio monitors are both connected to the same power strip, which is plugged into an outlet that does not have a ground .
• The noise is not 50hz or 60hz (which seems to be the "hum" symptom of ground loops) but higher. In my case the pitch of the noise is higher and depends on the activity of the PC (scrolling, high CPU/GPU usage, ...)
• The link between the interface and the monitors is a balanced TRS (from the balanced outupt of the scarlett) to XLR cable (to the balanced xlr input of the hs5)

Does the strip itself have earth contacts? If so, you can still have a ground loop.

Using an ungrounded outlet is dangerous (not conforming to electrical code in most countries) and electrically not a good idea for low noise. I guess you don't have any option...

I see now you are using balanced interconnects but still having the whine... which might indicate the interference is crazy strong and still sneaks through.

I would recommend fixing the missing earth grounding problem first if any possible.
After that I would expect the normal scenario of no or almost no hum/buzz/whine, many people have similar setups and no problem. In doubt, an USB isolator should fully fix any remaining small noise issues.

 

[Lambda]

In my setup there is no "multiple paths to ground" as there is no path to the ground

You seam to misunderstand how "ground" is used Electrically and what a ground loop is.

ground or earth may be a reference point in an electrical circuit from which voltages are measured, a common return path for electric current, or a direct physical connection to the Earth.

So there is no reason a ground loop curren needs to be 50/60Hz.

The Yamaha want a Ground and so is a normal commuter. if you don’t connect them to earth ground and neutral you can get all sorts of EMI problems and potentially dangerous situations.