HiFi USB cards in HQP computer; SoTM, Matrix, JCAT, among others

Placed an order today; Matrix Audio Element H PCIe internal USB Card. Looking for thoughts and experience from users of this family of USB cards. The reason for me is that going for the often recommended NAA end point, will still suffer from the problem with a computer USB output, a global and local onboard (DC/DC) SMPS, to mention a few.
Keeping my silent HQPe PC close to the rig, is not an issue for me and therefore I will try the alternative of inserting an USB card that allegedly with put out a far better bitstream and optionally can turn the USB 5 VDC off. The card I will try is not forced to be connected to the internal psu (SATA) as long as there is an alternate feed of 9-12 VDC from expernal psu. The channels between the card and PCIe port is proclaimed to be galvanically isolated and the signal reclocked by a precision femtoclock. Further is this card shielded and suppress the inside airborne noise from the extensive use of semiconductive applications.

Would you like to share your experience with your card, which ever you chose for your application?
Happy for any contribution. :slight_smile:

I do not understand what problem you are trying to cure.

1 Like

These cards do a good job of electrically isolating any influence from the PC itself but can’t do anything about separating, via physical distance, any stray RF or EMF from within the chassis.

I know people who have had very good luck with them but I’ve not tested any myself. I think they do provide a good benefit over the onboard MB USB ports. I still prefer a networked endpoint and keeping the server in another room.

But can you hear or measure anything that can be ascribed to “any stray RF or EMF from within the chassis” and which needs to be eliminated?

I’m not going to tell someone what they can or cannot hear :wink:

But, with the right test equipment I can show you these things exist on a screen.

These things matter only if they affect the analog domain. DACs should be more than capable of eliminating noise and jitter in the digital domain to an extent that makes them inaudible.

I am going to chime in though when the “problems” are just marketing fluff.

1 Like

That’s not what I asked.

I do not doubt that.

There are many types of DACs and many types of systems. These things depend on more than one device, so there are no simple answers either.

Some DACs have galvanic isolation for the USB, some don’t. Especially for those that don’t, it greatly matters what kind of grounding layout the overall system has.

And there’s more than one solution.

For example for certain single-ended measurements I use certain laptop because it gives cleanest results (least leakage currents over USB).


For zero leakage currents and AC RF, why not measure with laptop running off its battery only?

Or are you deliberately trying to see how well DAC can filter AC related garbage?

Oh, you meant me personally? Yes, I can hear RF but I’m also legally allowed to transmit within specific bands at 1500 watts in the USA so that’s pretty easy to hear if the equipment is too close to my antenna. Maybe that’s an unfair comparison? For this use case no, the little amount that PCs make I’m fairly sure I could not hear as long as it was high enough frequency to stay out of the audible bands. This is especially true of machines which are fanless. I still prefer my DACs in big heavy cases away from my computers.

1 Like

I thought all laptops were bad.

Not always, this Fujitsu from 2008 (made in Germany), even when running from the PSU is very clean.

Here’s example, same measurement (Topping D90SE), made with Intel Core i5 NUC:

And Fujitsu Core i3 laptop:

There’s a little bit of difference… (this is with “peak hold” mode)

Oh, and I’ve tested the Fujitsu with some USB cleaning gadgets. So far those have not improved the performance, but some make it worse.


Thanks, that does look cleaner, but why are we looking at MHz frequencies? Wouldn’t it be more interesting to see the audible band? Also, what test signal was used?

I look at everything. Here the discussion was about RF/EMI. Audio band has various interesting things too.

It is 0 - 22.05 kHz linear sweep, 2 minutes long, left running 3 cycles. So the plot is 1M FFT peak hold for 6 minutes. Input data is 352.8/32 PCM.

I can do 1M FFT up to 200 MHz bandwidth.

(Edit: corrected length after checking)


I agree it’s interesting to look at everything, but if it’s not audible, does RF/EMI really matters? I think we should focus on the delta in the audio band before concluding we need an “audiophile” digital chain.

1 Like

How’s the same measurement with your RME ADI-2 Pro with the same i5 NUC?

Why don’t you try, and then post your experiences. :wink:

Audibility always the usual exit excuse I see. Just like some people think that SINAD ranking matters, but when a problem is found claiming that it is not audible.

It matters because it can become audible in the noise floor through intermodulation for example in later amplifier stages. To focus on delta in audio band you need to first include the final equipment chain in the picture. Or at least you also need to test behavior of the amplifiers and other equipment when exposed to such high frequency inputs.

1 Like

I don’t think I have data stored for that. The old Fujitsu laptop is usually used because it gives cleanest results. Although I’m looking for another newer one for the same purpose. The low power Acer laptop I’m typing this on is not nearly as good.

But since ADI-2 doesn’t have galvanic isolation for the USB, one needs to be a bit careful with it. I think someone measured improvements with Intona USB isolator with it. I think I’ll get one too and check how it fares combined with the NUC using the D90SE since I have other cases stored for it already.

RME sent Intona their ADI-2 for the measurements on Intona website.

ADI-2 measurements improved with Intona but it can be system dependant.

For NAA, I use 5Vdc UpBoard Gateway powered by big powerbank for basically same end result - disconnection from AC related crap.

But if someone can’t be bothered with charging battery (understandable) Intona is certainly the thing to get. Forget the impersonators.