USB bus noise

[Dilettante]

I am seeing and hearing a fair bit of unintended noises coming from the USB-powered Audient ID4 mk. II

Silence, normalised to 0dB:

Zoom-in

Observations so far:

1. The noise can be heard when PC is idle.
2. It gets worse when GPU is busy
3. It gets better when CPU is stressed.
   Even then, there are some regular peaks which can be seen on the spectrogram.

Questions:

1. Will using some kind of a USB-isolator help? AliExpress has various devices based on Adum4160 or 3160.
   If yes, which one is recommended?
   
2. Is there a good reason to splurge on Topping HS01 which looks about the same, but ~5x times the price?

I have tried various USB-C and USB-A cables as well as USB ports ranging from 2.0 to 3.1
Conjuring @mansr and @KSTR as they seem to know something about the subject matter.

WAV file attached.

 

[Dilettante]

Sometimes interface would also do this. Normalised to -1dB

Note the (much) higher levels. CPU is idle.

.WAV is in the .ZIP

 

[Dunring]

I had the problem with an over kicked machine too I got the FebSmart 2 Ports USB 3.0 card for $15 on Amazon. My case has vents on the bottom so I flipped the power supply and put the USB card in the bottom slot, against the heavily shielded side of the power supply.
The Amazon basics USB cable is really good, now I have complete USB silence even when revving it with a stress test while pausing a song and turning the volume up. Best $15 I ever spent.

 

[Dilettante]

FebSmart 2 Ports USB 3.0 card

Interesting. I thought PCIe cards will suffer from the same noise given that it exacerbates with the load on GPU.

 

[Lambda]

Will using some kind of a USB-isolator help

they would work... but they probably don't because they are limited to 12Mbit.
the topping isolator is full speed and would work

TOPPING HS01 USB Isolator

Descriptions: HS01 USB 2.0 HIGH-SPEED AUDIO ISOLATOR High speed, Low latency, isolation withstands up to 1kVRMS True USB 2.0 High Speed HS01 supports USB2.0 High-Speed transmission, which is compatible with high-resolution audio transmissions such as PCM32bit/768kHz and DSD512 Native. Low...

shenzhenaudio.com

Is there a good reason to splurge on Topping HS01 which looks about the same, but ~5x times the price?

• USB 2.0/Hi-Speed: 480 Mbps.

the 10€ ones only do 12Mbps


You can try a active (power) USB hub. And you can try to put Ferrites on the cables (especially the usb cable)
they work as tested here but mainly for high frequency.

RCA cable Magnetic interference measurement and comparison

Cables are cables and there is no difference?! I designed a test to see if they perform significantly different in therms of "magnetic shielding" or Leakage inductance. This test dose not test for ground loop currents but interference from close by magnetic in homogeneous magnetic files. The...

www.audiosciencereview.com

And this post goes in what might causes this problem:

To get ride of PC white noise, dac with balanced outputs or dac with optical input ?

I got the jbl 305p mkII recently and I've that white noise when connected to my PC, it was also present with my old Logitech speakers but not as much. Reading past threads here, it seems like a DAC with optical input would remove the noise : - Topping DX3 Pro+ for 220$ CAD (bonus : it comes...

www.audiosciencereview.com

 

[mansr]

Questions:

1. Will using some kind of a USB-isolator help?

Quite possibly.

1. AliExpress has various devices based on Adum4160 or 3160.
   If yes, which one is recommended?
   
   
2. Is there a good reason to splurge on Topping HS01 which looks about the same, but ~5x times the price?

The ADuM isolators support only low/full-speed USB, which limits audio to 96 kHz stereo. The Topping supports high-speed and thus all sample rates.

It's also worth trying a plain powered hub.

 

[Dilettante]

which limits audio to 96 kHz stereo

Fine with me. The question is will ID4 negotiate using this speed.

---

Update: Powered USB hub (USB-C, separate good quality 60W PSU) has zero effect. All artefacts can be reproduced exactly as in the 1st post:

To make sure it's not an interface itself - I have connected it to the battery-powered laptop and all issues are gone.
So it must be USB bus of my PC.

 

[AnalogSteph]

How exactly are you taking these measurements, and what else is connected to the interface?

 

[Dilettante]

How exactly are you taking these measurements, and what else is connected to the interface?

Dynamic microphone --[XLR cable]--> DBX 286 ( turned off ) --[TRS 1/4"]--> ID4 mk. II

Steps to reproduce:

1. Record when PC is idle. Noise is clearly present and "sharper" compared to the calm hum of the noise floor.
   
   
2. Run something to load the CPU. Specific instruction set appears to be irrelevant, any process will do.
   
   At this stage the noise floor is at its lowest, no issues apart from the faint regular clicks which can be seen if you squint here:
   
   
   
   
   
3. Run GPU-intensive process. This is where things go bad. In extreme cases (90%+ load) audio is dropping and distorted.
   This particular problem is generally reported across the web.
   
   
4. Stopping processes that load CPU & GPU returns everything to the initial state.

 

[anotherhobby]

My computer does this with every single DAC I've tried (1 schiit, 3 toppings) other than the miniDSP Flex. I don't know how the Flex is silent, but it is. The USB isolators have not worked for me. The only 100% fool proof fix is to use optical to decouple the grounds. The Topping D10s is perfect for this. USB > D10s > Optical > DAC

 

[Dilettante]

My computer does this with every single DAC I've tried (1 schiit, 3 toppings) other than the miniDSP Flex. I don't know how the Flex is silent, but it is. The USB isolators have not worked for me. The only 100% fool proof fix is to use optical to decouple the grounds. The Topping D10s is perfect for this. USB > D10s > Optical > DAC

Valuable, but I would absolutely try to avoid additional purchases if possible. Will keep it in mind though.

 

[anotherhobby]

Valuable, but I would absolutely try to avoid additional purchases if possible. Will keep it in mind though.

The only answer I know of to the CPU/GPU usb audio noise problem is to find a magical DAC thru trial and error that doesn't get your USB noise, or to decouple the grounds (like with optical). If you want to go cheaper, the Douk Audio USB/Optical interface will do it for only $56, but I like the extra functionality, form factor, and quality of the D10s for the price.

 

[Lambda]

audio is dropping and distorted.
This particular problem is generally reported across the web.

if it is completely drooping out it might be a software/firmware/diver chipset problem es.
In this cases i might be wrong!

But normally USB has error correction an it works or it don't.
If USB would have this problems with file transfer it would mess up all sorts files not only audio?

But maybe i'm wrong.

How high/ is the absolute level of the noise your getting?

input can be 21dbu

DBX can output 21dBu..

is it running at 21dBu?

 

[Dilettante]

the Douk Audio USB/Optical interface will do it for only $55, but I like the extra functionality, form factor, and quality of the D10s for the price.

Don't see how this will help. ID4 has no optical input and I need to record as well, not only listen.

 

[anotherhobby]

Don't see how this will help. ID4 has no optical input and I need to record as well, not only listen.

Sorry... depends on if your DAC accepts optical. My other DACs have had optical input. Computer > USB > Douk > Optical > DAC.

 

[Dilettante]

How high/ is the absolute level of the noise your getting?

ID4 mic. preamp is set to max gain, all other variables unchanged.

CPU idle: -48dB
CPU stressed: -53dB
GPU stressed: -39dB

Alternative visualisation

RAW, unmodified recording is attached to this post.

 

[Lambda]

ID4 mic. preamp is set to max gain, all other variables unchanged.

Why? it shuld be set to min. gain?!
waht input is used? mic, line or instrument

your gain range is 58dB so with proper gain staging you shuld get under -90dB what shuld be fine

 

[AnalogSteph]

Dynamic microphone --[XLR cable]--> DBX 286 ( turned off ) --[TRS 1/4"]--> ID4 mk. II

I suspect what's effectively a Pin 1 Problem that is uncovering the ground loop between the DBX (an IEC Class I device) and the PC. Notably, said loop includes both USB ground return and power supply mains PE.

Are you able to modify your TRS cables? If I'm right, the problem should go away once the shield connection is disconnected at one side. Likewise when adding a line isolator like a Behringer HD400.

It is not necessarily easy to determine which side is responsible, i.e. not conforming to AES48 guidelines - it takes a thorough examination of internal ground routing. They are very clear but it is not uncommon to see ease of assembly take priority.

I have encountered reports of such issues on the output side before, where monitors would be connected to an interface via balanced cabling yet unwanted noises would still occur.

 

[Dilettante]

Why? it shuld be set to min. gain?!
waht input is used? mic, line or instrument

Mic. There is no separate line input on ID4. I have tried again, with preamp at min. level (as far as I can discern it from "line in" mode which is exactly the same knob, but a 1 degree difference in rotation)... Nice monotonic hum. So the problem manifests itself when I drive the ID4 preamp higher than absolute minimum! Shall double-check with the battery-powered laptop tomorrow.

Recording at min. level attached.

 

[Dilettante]

Think I have a bit of a breakthrough here. Remember I've mentioned that the problem occurs with DBX turned-off? It was indeed turned-off, but plugged into the wall socket.

Now check this out:

Two suspects left:

1. DBX itself
2. Shoddy power network in the building

Make your bets. Call the perpetrator. Suggest an appropriate reprimand fix.