Reducing computer noise reaching USB audio interface

[Khron]

My Maya44 USBs are 3 years old, no idea when the model was last revised. I have two, though the noise problems are present when one, the other or both are in use.

I need a minimum of 3x2 outputs (stereo track 1, stereo track 2, stereo beds/sweepers), but ideally 4x2 (stereo track 1, stereo track 2, stereo bed/sweeper 1, stereo bed/sweeper 2) and 1x2 inputs (stereo mixed audio for streaming out)

I used to use the amazing M-Audio 1010LT, but then upgraded my PC to one without the old fashioned PCI slot. I use Linux which limits my range of interfaces a little bit.

If anyone can recommend an 8+ line output, 2+ line input, Linux compatible, PCIe or external audio interface for a reasonable price, I'd be most grateful.

If being bus-powered is not a requirement, that opens up your options a LOT. Especially if the RCA ins & outs are not mandatory either.

https://wiki.linuxaudio.org/wiki/hardware_support
That might not be a comprehensive or up-to-date list, but it's at least a starting point.

 

[57sputnik]

I've used the iDefender with several Focusrite 2i2 USB interfaces and it has helped. Your mileage may vary...

https://ifi-audio.com/products/idefender3-0/

 

[michaeljtbrooks]

The computer case has a whopping glass panel on one side, it wouldn't surprise me if this is part of the problem, though I cannot hear the noise on the analogue lines if the Maya44 USB is switched off (external USB hub has switchable ports). It certainly seems to be coming down the USB power lines rather than being induced in the cabling. Though as I've said it's relatively low noise (-80dB).

This is for FM broadcasting with online simulcast (inc high fidelity options), so stereo image and fidelity matter. (Before anyone raises concerns, the source audio is lossless wav or FLAC, and the stream back to the studio uses Opus codec @ 510kbps).

 

[cyrano]

The computer case has a whopping glass panel on one side, it wouldn't surprise me if this is part of the problem,

Aha!

Easy experiment: put a grounded shield plate inside, or even outside the glass panel?

 

[Newmarket]

Try this: Take an extension cord with an earth ground prong and plug it into the power strip that your Behringer mixer is plugged into. Then insert some fat wire or bit of metal into the ground prong hole at the other end of the cord. Obviously you want to be careful not to stick things into the other holes or you'll get shocked and die. But if you don't die, run your test, listen to the noise and then see if there's a difference when you touch the fat wire / bit of metal to metal parts on the USB interface like the 0V of one of the USB I/Os, the case, a screw or similar. If this reduces the noise, this is a clue that what I described above is at least part of the issue. Some gear has a thumb screw for grounding it in scenarios like this.

Please DON'T try it quite like this. Despite the warning it's just never a good idea to play around like that with a mains outlet.
It will (UK at least) open the safety shutters on L & N and you're not guaranteed to get a good Earth connection anyway.
Do it properly - wire a mains plug with just a clearly marked - Yellow/Green - Earth wire only. Now you have a good Earth connection and the L & N socket holes are covered. You could also remove the N pin and the L pin if your country's mains plug type allows for it.

 

[Bo Deadly]

If the Maya44 is physically close to the PC and far away from the mixer, try making it far from the PC and close to the mixer. Meaning use the digital cable (which is immune to noise) to create distance between the computer and the interface and make the analog cables short. It doesn't sound like ground potentials are a problem but it could be radiation from the PC in which case just putting a little distance between them could make a difference. And if the glass is only on one side where the other is metal, try re-arranging things so that the Maya44 and mixer are shielded by the metal side. A blingy gamer PC with lights and such may not be ideal for working with sensitive analog electronics.

 

[ccaudle]

Behringer DX2000

Is it actually the DX2000USB, or was there an older model DX2000 which did not have the USB interface included?

I use Linux which limits my range of interfaces a little bit

Any class compliant interface will work with linux. With audio class 1 devices that limits you to 16 bits, audio class 2 devices can utilize 24 bits. Any device which claims it is compatible with iOS is definitely class compliant, any device which claims compatibility with MacOS is almost certainly also class compliant.
If the device specs explicitly say USB audio class 2 compliant then it will definitely work. iOS only works with class compliant devices, which is why iOS compatibility is a good implicit way to tell even if the marketing docs don't use the words "class compliant" anywhere.

Is the computer earth grounded? Chances are, it is not.

A desktop computer is definitely grounded through the power cable. A laptop may or may not be depending on power level. A very low power laptop may not be, a high end laptop probably is. Obviously not when running from battery power, but from a quick check of the laptops I have all have 3 pin power cords except for a Chromebook.

As others have mentioned, an isolation transformer in the signal connection is probably the most effective and reliable. There are cheaper solutions, but these Jensen products work well:
Jensen two-channel RCA connector isolator

That is basically the same as a DI box with ground lift. If you have any passive DI boxes you could give it a try using some adapter cables to make sure it works well for you (presumably a DI box will have an XLR output, which is why you will likely need adapters).

You will have to check the financial math, you can get a new DX2000USB mixer with built in USB interface for $300, that is probably not much more than buying three or four stereo isolators, so at some point it becomes cheaper to just get new equipment than to try to band-aid what you have.

 

[michaeljtbrooks]

Is it actually the DX2000USB, or was there an older model DX2000 which did not have the USB interface included?

It is the DX2000USB, but the USB functionality is rather anaemic. You get one stereo input (appears on "Channel 3" of the mixer) and a copy of the record out stereo output.

However you did remind me to test the mixer's USB input to see if the noise is present there... and yep it's there and a little louder via the Behringer DX2000 USB interface compared to the Maya44 USB. Also of note the background white noise is louder too when the gain is adjusted to give equal signal level. The Maya is definitely a better interface!

So I really think the noise is coming up the USB power lines from the computer internals.

If the Maya44 is physically close to the PC and far away from the mixer, try making it far from the PC and close to the mixer.

Already is the case. Moving it around reduced things marginally and only when it was right by the computer, but putting it on that external hub made the biggest difference.

And if the glass is only on one side where the other is metal, try re-arranging things so that the Maya44 and mixer are shielded by the metal side. A blingy gamer PC with lights and such may not be ideal for working with sensitive analog electronics.

Yes this is also the arrangement I've gone for. The glass panel is against the wall, all other kit is in the direction of the metal sides.

A desktop computer is definitely grounded through the power cable.

Yes mine absolutely is grounded.


So, assuming this is common mode noise, it sounds like both a mixer upgrade and a USB interface upgrade to ones with balanced inputs/outputs is the way to go. However, I'm not sure dropping £500 is worth it to nudge down noise which is already only -80dB, as much as I'd like an excuse to buy an X32!

 

[Khron]

So I really think the noise is coming up the USB power lines from the computer internals.

... But the DX2000 is not powered from the USB, surely...

 

[cyrano]

With audio class 1 devices that limits you to 16 bits, audio class 2 devices can utilize 24 bits.

I doubt that as I still own an Emagic 6/2 and a 2/6 which both support 48 kHz 24 bit @ 8 channels (6 in, 2 out or vice versa) and even 96 kHz @ 4 channels.

They're also an example that "every class compliant device works on Linux" isn't necessarily true as the bootloader needed to upload the firmware has been dropped from the kernel several kernel versions ago. I guess Linus decided nobody uses these old devices anymore. I guess he's right as I've given up hacking the bootloader for Mac many years ago.

I also own an Apogee USB mic that is class compliant, works with Macos and iOS, but not on Windows 10 or Linux. The newer model does.

 

[Ivan K.]

Look up "USB Isolators". Amazon & elsewhere. Most use an Analog Devices chip. They are galvanically isolated and provide up to 2.5W for power. There are 3 levels: 2.5kV to 7.5kV
https://www.analog.com/en/product-category/usb-isolators.html

 

[Ivan K.]

Try a "USB Isolator", available at Amazon, etc. Most use Analog Devices, galvanically isolated with 2.5W of isolated power. Isolation is from 2.5kV to 7.5kV.

 

[ccaudle]

I still own an Emagic 6/2 and a 2/6 which both support 48 kHz 24 bit @ 8 channels

It was common for older devices to support more than basic USB audio class 1 features by using a proprietary driver rather than a generic audio class driver.

They're also an example that "every class compliant device works on Linux" isn't necessarily true as the bootloader needed to upload the firmware has been dropped from the kernel several kernel versions ago.

First, the devices are not class compliant, it had a proprietary driver even on Linux, and second, even if a device is class compliant, if it needs firmware loaded from the host then they aren't class compliant from power on, so you would still need a proprietary driver to get it into a compliant state.
I don't think quirky hardware that hasn't been produced in 18 years is particularly relevant in this particular situation, but if you wanted to try to resurrect your old hardware someone fixed the firmware loader about 12 years ago. No idea if it can still compile with the latest kernels, I suspect most of that generation of hardware has been retired by now.
eMagic firmware loader fix for 2.6.x kernels

 

[cyrano]

There's NO driver. Just a firmware loader. At least in the Mac version. I've never tried the Windows software.

One trick I used a while to get them to work was loading the firmware with an older system and then rebooting to a newer system. The newer system had no driver or any other software for the emagic devices. As long as USB power wasn't interrupted by the reboot, it worked. That worked for every Mac I ever tried it with, but not with some PC hardware. Then some kind person from Australia reversed the firmware loader and released not only the result, but also the information so I could try it with other software versions and newer Mac systems. Until Apple went with Intel processors. But by that time I was using an RME.

Besides, the point is/was that USB 1 class compliant devices aren't limited to 16 bits, or 48 kHz. I'll admit that the emagic devices are the only ones I know of to support 96 kHz while being audio class 1 compliant. And that only with the latest firmware version. That story ended when Apple bought emagic in 2002.

I know the USB audio class 1 specs says so, but it simply isn't true.

 

[Michael Tibes]

Steinberg URC22 here, about one and a half year old (2in/2out) into ESI active speakers -> no way to get rid of the hum / noise, even if I power the interface from an external power supply or even a battery. I had no soldering iron at the place, so I cut pin 3 in the XLR jack inside the speaker on the amp pcb. It is dead quiet since that, but I guess the noise will be back if I connect a grounded device to the input. So far I have only used a mic and a guitar. I had no problem with other speakers before. It seems the problem is a ground loop. So does a usb3 connection need ground?

Michael

 

[Khron]

Steinberg URC22 here, about one and a half year old (2in/2out) into ESI active speakers -> no way to get rid of the hum / noise, even if I power the interface from an external power supply or even a battery. I had no soldering iron at the place, so I cut pin 3 in the XLR jack inside the speaker on the amp pcb.

Why pin 3, and not pin 1?

So does a usb3 connection need ground?

If the device is bus-powered, definitely.

 

[ccaudle]

does a usb3 connection need ground?

Yes, there is no transformer coupling, and the input devices generally run from 3.3V, so limited common mode offset range. I suppose in principle you might be able to disconnect the ground at the receive end if both devices were connected to safety earth to limit common mode voltage difference, but the device will almost certainly no longer be within EMI emissions limits if the cable shield is disconnected. Dealing with 10Gb/s circuitry has significant differences to audio circuitry, some of the tricks you think you know from dealing with audio don't work the same at those frequencies.

 

[Crihnoss]

Hey I'm having the same problem
Here is my situation:

• When I use my active monitors + active subwoofer on my PC with integrated audio I get this horrible static noise.
• Same with an external DAC (USB powered. Same cable as data source).
• Monitors + PC, without the subwoofer. No noise.
• Monitors + subwoofer with my laptop/Switch/PS4. No noise.
• Monitors + subwoofer + external DAC powered with one of the PC USBs. Noise
• Monitors + subwoofer + external DAC powered with a wall socket USB(On PS4/Switch). No noise.

I tried removing the video card and using the CPU iGPU because the noise is the loudest when the GPU is on high load. No change.

So my guess is the Motherboard or the PSU.
The PSU is a BeQuiet! Pure Power 11 FM (1000w).
The Motherboard is an Asus ROG Strix 660-F Wifi, which has the dreaded alc 4080 chip, which I read all sorts of nasty stuff online (I get random total sound corrupting out of nowhere and the only way to stop it is resetting the bit stream settings. Also if the ARGB is on, there is like 1-second latency with windows sounds... Yeah, no idea).

I'm sending the motherboard back for a refund because I don't want to have to disable a feature to work around (Even though I despise RGB and I turn it off, I'd like to have a fully functional piece of hardware which I paid for) And going to try another brand, without the alc 4080.

If this fails I'll get a replacement on the PSU.

If any of this fails... I don't know what to do... I tried a mini-plug/RCA cable ground-loop noise filter. It solves the static noise but it destroys the lower-frequency sounds. So, not an option.

I also read about USB isolators. I could use that between the PC and the external DAC, but in the reviews on amazon, I read that they only support MONO sound?... Any model to recommend?

Sorry, I don't want to hijack the thread, but it's pretty old as is.

For the curious ones here's the rest of my hardware, in case you think it might be related:

• i7 13700k
• 16x2 Kingston Fury DDR 5 6000 CL36-38-38 1.35v XMP
• Gigabyte RTX 3080ti Vision OC
• Asus ROG Strix 660-F Wifi
• Kingston KC3000 SSD
• BeQuiet! Pure Power 11 FM (1000w) PSU.
• Corsair 4000D case
• Pioneer DJ DM-40 monitors
• Mackie CR8s XBT Subwoofer
• DAC Fiio BTA30pro

 

[Khron]

That sounds like some sort of ground loop. I'm assuming your active monitors get connected through the subwoofer?

And obviously (?), you're using the RCA inputs to the subwoofer, when you're using it? Are your monitors also earthed, or are they double-insulated?

 

[Crihnoss]

Yes, they are connected with RCA. The monitors don't seem to have a grounding pin in their plug, now that you mention it. The sub does though.
But why does it only happens with the PC as a source? With other devices, I have no issues.