8kHz USB noise: How do I get rid of it?

[oohlou]

TLDR:
PC as my source, Massdrop SDAC, no music playing, volume set to 100%, noise floor measurement, how do I get rid of this 8kHz spike?

Some history:
I bought my first external DAC, SMSL SU-8, this past summer and while it was a huge update over my existing equipment in terms of sound quality it also brought with it some audible noise during quite sections of music / no music playing. At the time I thought it was a ground loop problem so I bought a iFi iDefender3.0. It reduced the noise but not enough to my liking. So I returned the SU-8 and bought a "MASSDROP X GRACE DESIGN STANDARD DAC BALANCED". The noise was significantly better but after awhile I noticed the 8kHz noise.



How do I get rid of the 8kHz peak? What do I need?

A better power supply for the iFi and thus SDAC? I thought about this but the SU-8 had a nicer power circuit (directly AC powered with ground and a with torrid transformer) than the SDAC (just USB powered) and the noise was still there.

What about a better DAC? I've thought about the Topping D50s or even a SMSL M500 but I don't have a good feel if they will actually fix the problem. I notice the M500 doesn't have a grounded power cable. I wonder if that makes a difference?


Thoughts?

As a short aside:
Here is my noise floor without the iFi iDefender3.0 (note I'm not doing anything to minimize external noise, so there is non-trivial PC fan noise and cars driving by some in both this measurement and the above measurement). So you can see there is tons of noise coming from the PC that the iFi actually successfully filters out.

 

[pma]

Is your soundcard in a desktop PC, or is it a USB soundcard? Do you use a notebook? Best option is notebook supplied from its battery and USB soundcard, re clean noise floor. Or, an isolation transformer with low primary/secondary impedance. Class I desktop PC is usually a problem, also regarding its internal loops to chassis.

 

[Berwhale]

How are you measuring the noise?

 

[pma]

Looks like signal groundloop or two different DAC - ADC devices on the same PC - such setup would never work if balanced I/O are not used.

 

[oohlou]

I am measuring using REW RTA measurement with a UMIK-1 microphone. I'm just taking 10-20 averages holding the mic up near my ear. So far from scientific, but I think good enough to demonstrate the problem.

My PC is a big gaming PC, and it is connected to the DAC via USB. I am sure there is tons of RF / EM noise coupling into the USB signals and ground. However, the iDefender3.0 seems to be doing a good job of filtering that noise out at least to the point where I cannot hear it over other ambient noise. My problem now is that 8kHz which I've read is coming from the USB interface itself (USB microframe period is 125uS so due to poor linearity of the power supply, grounding ,and/or decoupling I can "hear" the frame rate. If you search for "USB 8kHz" you will find this discussed a number of places but I haven't found the defacto solution. Also by the way modern search engines and ASR SEO is amazing, this thread already shows up pretty high when you search for those terms).

What about the iPurifier3? I've drank the measurements matter koo-aid and I don't buy into the all the "it sound richer blah blah" reviews, I just want to get rid of this 8kHz noise. The iDefender definitely helped a lot but the iPurifier is supposed to remove USB noise as well. Thoughts?

 

[Berwhale]

So you're measuring your room, not your DAC. Why are you so convinced that the DAC is the source of the 8Khz spike rather than something else? What's the DAC plugged into? (amp, speakers, headphones?)

 

[oohlou]

Here is how I know:

Room with audio equipment off, PC still on:

Room with phone driving amp directly, DAC off, PC on:

I'm not saying the DAC is the source parse, the PC is the source. I do however want a DAC or piece of equipment between my PC and DAC which can eliminate the 8kHz noise.


The total setup:
PC as source -> iDefender3.0 -> USB cable -> SDAC -> 3.5mm to RCA cable -> JDS Atom -> RCA cable ->Parasound 2125v2 -> speaker wire ->
Klipsch Forte 1

(I've taken the Atom out of the chain to test but it did not make a difference which is why I didn't mention it before).

 

[mansr]

I am measuring using REW RTA measurement with a UMIK-1 microphone. I'm just taking 10-20 averages holding the mic up near my ear. So far from scientific, but I think good enough to demonstrate the problem.

Are you able to record the output from the DAC electrically without using a microphone?

If the 8 kHz spike is (which is likely) USB packet noise, the cause of this is probably within the DAC, and there won't be much you can do about it. As you note, USB data transmission occurs in bursts every 125 μs. With the arrival of each such packet, the receiver will draw more power while creating a burst of high-frequency noise. The increased power draw can/will result in a corresponding dip in voltage. Without sufficient regulation, this dip will be seen by the DAC chip and end create an 8 kHz ripple in the output. Increased ground currents can also create voltage fluctuations at the DAC chip with the same effect. Another possibility is that aforementioned noise from the USB receiver gets into the analogue circuits through parasitic coupling where it is demodulated, again creating an 8 kHz tone.

None of these causes of USB noise can be remedied externally to the DAC. That said, if the DAC offers the option of using a separate power supply rather than USB power, this may have better regulation and/or a different distribution path. Also, make sure the USB cable is a decent one with low impedance in the ground wire. Otherwise part of the return current for the USB signals may find its way through the audio interconnects where you do not want it. Try the shortest USB cable you have around.

Devices like the iDefender can actually make this worse, so make sure to try without any of them in the path.

 

[Berwhale]

So would switching to a DAC with it's own external power supply solve the issue?

 

[miero]

I had a similar issue with "Pro-ject Pre Box S2 digital". PC with ATX supply, USB connected to DAC and then RCA into XLR on Behringer active monitors (Class I). I returned the DAC on the next day. Later I found that active speakers were connected to an electric outlet with a different fuse. So there was a big ground loop (at least 20 meters) which I was not aware of that time.

 

[Cahudson42]

Pretty much a shot in the dark, but for $9 you could temporarily replace the SDAC with an Apple dongle. Also need a USB-C adapter for connecting to your PC most likely...

Thinking one by one component replacement/removal to track it down..

Edit:
Wasn't familiar with the SDAC. So took a quick look at Amir's review. IMO, didn't seem overly impressive (if I'm looking at the correct $70 device). Interesting peak and noise bump around 8khz.. Except for the relatively horrible 2nd harmonic peak, the Apple doesn't look much/any worse. No 'funny business' around 8khz either..

Also, don't see that anyone has mentioned it yet - but the Dark Side wins - how about an AudioQuest Jitterbug?

 

[oohlou]

Pretty much a shot in the dark, but for $9 you could temporarily replace the SDAC with an Apple dongle. Also need a USB-C adapter for connecting to your PC most likely...

Excellent, cheap, easy idea. Thank you! I picked up one today. Unfortunately it didn't solve the problem or at minimum exacerbated other problems. Here is the room measurement with the apple adapter on my PC to my Atom:

Note it is as bad as my second graph above (SDAC without iDefender). The 8kHz tone may not be there but the other noise is really bad.

Interesting enough since the Atom has both RCA in and 3.5mm in I could test with both the SDAC and the Apple adapter connected at the same time. This showed the noise is present when the Apple adapter is connected independent of what input I had selected. Once I disconnect the Apple adapter from the Atom the noise is gone. It also shows that the 8kHz doesn't seem as bad with the Apple. So I think this definitively confirms one of my problems is electric noise from my PC ground is getting into my audio equipment leading to audible noise from my speakers and the iDefender actually does a good job in filtering that out (by isolating the power and ground from the PC). The 8kHz problem does in fact seem to be related to the SDAC power. I tried a few different USB power supplies I had for the iDefender (since it is powering the SDAC as well), but I none of them make a difference.

I think I'm probably going to give the Topping DS50s a try. Even if it has the same USB problems I can just switch to optical which should address both of my problems.

Oh and

Are you able to record the output from the DAC electrically without using a microphone?

Hmmm. I have a mid level oscilloscope which does have FFT functionality. Is there a standard load used for testing? e.g. a standarized test jig? Of if the DAC is driving directly into the scope what impedance is needed?

 

[McFly]

Holding the mic up to your ear? First of all, get the mic on a stand or something inert. Your body could be exciting the frequency (kidding)

Does changing USB audio class work? Can you try toslink or coax spdif out from your soundcard? A lot of gaming motherboards can do spdif out

 

[Cahudson42]

Another thought before $250 (?) on the D50s. Eliminate the PC?

For $25 you can get:

TracFone LG Rebel 4 4G LTE Prepaid Smartphone (Locked) - Black - 16GB - Sim Card Included - CDMA https://www.amazon.com/dp/B07GC5QKTX/
Snapdragon 425 based, and likely wdc9xx DAC. Also works USB with OTG cable.

Throw away the sim and tracFone stuff. Runs fine on 2.4g wfi. Garden variety Android 8.1. Run whatever app player, Amazon Music HD, etc. Try 3.5mm, and USB via OTG cable.

If you like the result, consider a LG V20 with it's Quad ESS DAC . Maybe no need for a D50s. They are $120 NIB on ebay. ATT 910 (? Maybe h910, not remembering). I use one with a Liquid Spark and HE400i HP. Replaced a D10 with the V20.

 

[Berwhale]

Presumably the UMIK-1 is plugged into the same PC as the DACs, so are their grounds linked? I wonder if the measurement when using your phone as a source would change if you plugged it into your PC? (I guess the phone will run off it's battery, even when charging, so maybe not).

The D50s is certainly going to give you more options with connectivity and the ability to electrically isolate the DAC from the PC (via SPDIF or Bluetooth).

 

[oohlou]

@McFly Changing the USB to 1 indeed changes the tone. It is now 1kHz and harmonics of it as expected giving the slower frame rate. This is actually really cool to see. Red is USB 2, Green is USB 1

@Cahudson42 That would probably solve my problem but I want other PC sounds too such as game audio.

@Berwhale Yes, the UMIK-1 is plugged in to the same PC. I suspect the UMIK-1 has better power isolatation and filtering especially since it is a calibration microphone but that is just conjecture on my part.

 

[oohlou]

I got my Topping D50s already. I still need the iDefender to break the ground loop between my PC and amp but I'm happy with the results. Problem solved! Woot!

 

[majingotan]

I got my Topping D50s already. I still need the iDefender to break the ground loop between my PC and amp but I'm happy with the results. Problem solved! Woot!
View attachment 41514

At least the iDefender is not snake oil and objective measurements prove that case.
I also use the larger iFi iUSB 3.0 Micro to break the ground loop from my laptop to my DAC. No impact on performance based on my subjective opinion. It’s a nice to be used as a USB charger as well where I can turn off my laptop and still charge my Chord Mojo and iPod Touch 7G