So I was reading up on USB3 over the weekend and it’s well known that it emits quite a bit of RF in the 2.4ghz band and is known to effect wireless performance on nearby aps or devices. Not noticed this myself but never looked for it and this got me thinking could this RF also effect my audio side as Naim seems susceptible to RF.

Now my Roon core is located in my rack below my Atom. I use a USB SSD attached to it for storage of music and for a CD Drive for ripping. Both of which where connected to USB 3 ports and the SSD does use USB 3 to connect. So I decided to disconnect them and plug them into the USB2 ports instead on my Roon server.

Unexpectedly I noticed immediately the whole system sounds better as a result. It’s smoothed off the top end which was often quite harsh. I had put it down to my room and the new speakers I had bought. So to combat nasty bass modes my room has I had gone to using DSP and room correction which worked brilliantly. But I could not use it for analogue part of my system.
I have recently modded the room a bit with some bas traps and moved stuff about and have reduced the bass considerably. So was trying without DSP again. But it was still a tad harsh on the top end. So was thinking it must be my speakers. But ffter switching to USB2 it looks like to me it was the USB3 ports causing the issues and my Atom picked it up.

I used to use the USB2 as default so maybe when I rejigged things which was around the same I got the new speakers I did not think USB 3 would change anything . Either way I have found a tweak that’s done something positive. It might be expectation bias, who knows but it’s made more impact than any ethernet cable has. So maybe beware of USB 3. Or I’m just mad.

I recall a couple of comments from @Simon-in-Suffolk that USB3 might be less noisy than USB2, if I understood him correctly. That would seem to be at odds with your findings here. Not that I’m trying to prove you wrong, just an observation.

According to the web it’s very RF noisy in 2.4 GHz band and raises it’s noise floor which has caused issues in WiFi. But no mention of audio side effects. There is a white paper on it. But it seems to have changed my system swithing or usb 2 so god knows what it is but its helped.

Correct, by design all other things being equal, USB3 is less noisy than USB2 (by design and the way it functions). Obviously that requires the USB3 interface working at USB3 with a duplex (USB3) cable.
I suggest USB2 and delicate audio and RF applications are kept physically apart, and use ONLY USB3 for quality audio applications.

In layman’s terms USB2 creates pulses that will likely radiate for signalling by using high frequency unbalanced voltages. USB3 works like modern duplex Ethernet, and sends data both ways concurrently, and all signalling is embedded in the balanced data flows.

Well my current experience tells me otherwise and that using USB2 provides a much nice sound which I can only put down to reduced noise. This is not for a DAC connection just purely a SSD with my music on plugged into my server given there is a lot of evidence of it being noisy in 2.4ghz perhaps something from this is affecting my Atom somehow.

Sure… but USB2 will be electrically noisier in operation when transferring data, you can see for yourself with a spectrum Analyzer, or a band scope which is what I have used… and more suited for RF noise measuring.
I ensure I physically route my USB2 cables away from my RF cables… and I now don’t use USB at all for audio…

If its putting out RF though it seems to be counter productive to me. Also increased speeds are useless for audio.

Why? For the consumer applications and the relatively low bandwidth and simple data flows which USB2 was developed for, this was a compromise worth taking to simplify and lower the cost of cables… the EM radiation is relatively small and within consumer tolerances … however we are talking interactions with delicate Hi-Fi and audio components which are sensitive to far finer tolerances. (And I was also referring to sensitive RF applications in my case)… these would be considered special cases.

I have an Audiophilleo USB2 to spdif converter sitting on the back of my nDAC using a BNC to BNC connector as recommended by Audiophilleo. I keep the Innuos Zenith away from nDAC (well my pop records sit between).

So is the generation of the signal the biggest generator? Is the flow control return signaling as bad? Would the slack on the 2m cable need to be keep away from the nDAC as much as possible? Would the 552 PS on the left of the Innuos be susceptible?

With gratitude for your advice


It’s the flow control in USB2 that is the cause…
I think separation of two metres is more than ample…

1 Like

But if the Audiophilleo is on the back of the nDAC spdif wise. I could use a BNC cable to move the Audiophilleo but Audiophilleo suggest these cables can introduce signal echos.

The Audiophilleo chips are of course build into the DAC V1. I’ve been really pleased by the SQ of my nDAC solution.



not sure what is meant by signal echoes - unless reflections - and if reflections as long as the SPDIF cable and sockets are 75 ohm impedance then this should be minimal and probably irrelevant


Audio use of usb would be limited to 12 Mbps (full speed) at 24/192 I suppose? Does that qualify as RFI? Does the host choose the data rate dynamically or purely based on initial device identification?

Thanks for your help


Hi, the throughput is not really linked to the signalling techniques.

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.