'Engineered by Audio Note' EVGA's Nu Audio sound card

[hollowman]

Asus Xonar (and, to some extent, Creative) PC soundcards seem to have held their ground despite lack of upgrades to their lines.

Now it's EVGA's turn. At $250.00, the card seems to offer quite a package (see links).

Darko wrote:

EVGA -- among audio circles they barely rate a mention but for those building PCs the Californian motherboards and graphics card manufacturer is much bigger news. For their latest product, EVGA teamed up with Audio Note UK for a most unlikely collaboration.

The NU Audio Card sports a multi-layer PCB with passive heatsinking and 'isolated dual ground planes', presumably to hold back (the negative influence of) electrical noise from infiltrating both analogue and digital sections. Audio Note UK's in-house developed capacitors and resistors feature in the card's fully-regulated power supply and analogue circuitry.

On D/A conversion, we see a pair of oscillators (one each for 44.1kHz and 48kHz families) clock an XMOS chip's output into an AKM AK4493 decoder chip (PCM up to 32bit/384kHz and DSD256) which in turn feeds the card's analogue output stage: an ADI AD8056 op-amp spills via a pair of RCA sockets and is fully rollable.

Rollable too is the ADI OPA275 dual op-amp that forms the backbone of the 6.4mm (variable) headphone output that offers an output impedance <1 Ohm. It's reportedly good for headphones whose rated between 16 Ohms and 600 Ohms.

On volume control, we are not necessarily forced into the digital domain (where DSD doesn't play). EVGA's own software interfaces with the NU Audio Card's Maxim DS1882 controller for fully analogue attenuation and, making hi-fi purists wince,'audio reactive' lighting.

A TOSLINK output is there, ready and waiting, for when we upgrade to a superior sounding outboard DAC.

For signals moving in the other direction --for those who like to do needledrops of their favourite records -- the NU Audio Card also features an AKM AK5572-powered ADC, fed via a 3.5mm socket. For gamers, the all-important mic-in socket rounds out the EVGA card's feature set.

And whilst EVGA is almost certainly aiming this audiophile-grade sound card at gamers and PC builders, hi-fi enthusiasts might wonder if high-end audio server/streamer manufacturers will see the NU Audio Card as a drop-in solution for their own products. Time will tell.

EVGA - Articles - EVGA Nu Audio
'Engineered by Audio Note': EVGA's Nu Audio sound card | Darko.Audio

An externally hosted image should be here but it was not working when we last tested it.

 

[seanp2k]

EVGA's own software interfaces with the NU Audio Card's Maxim DS1882 controller for fully analogue attenuation and, making hi-fi purists wince,'audio reactive' lighting.

I winced at the first part too. Please don’t try to sell analogue volume control as superior in 2019. Stepped attenuators, relay-controlled resistor networks, even LDRs...just regulate how much juice the D->A puts out into the amp chips, or even better, put it all into the same chip. I try to stay away from anything these days with a pot on the outputs. Putting one here where it’ll be on the back of the PC makes even less sense to me, unless that’s just a pad for the input or a rotary encoder that changes it in software. Then again, RME is doing this all-digital volume control on a $2000 ADI-2 Pro, whereas this is a $250 unit, so I guess I should attenuate my expectations accordingly.

 

[hollowman]

Hey DIYA mods...
Can you move this thread to "PC Based" child forum?
Thx!!!!

 

[canvas]

I like the idea of external power source and power re-generation from charge pumps or boosters.

 

[sgrossklass]

I winced at the first part too. Please don’t try to sell analogue volume control as superior in 2019. Stepped attenuators, relay-controlled resistor networks, even LDRs...just regulate how much juice the D->A puts out into the amp chips, or even better, put it all into the same chip. I try to stay away from anything these days with a pot on the outputs. Putting one here where it’ll be on the back of the PC makes even less sense to me, unless that’s just a pad for the input or a rotary encoder that changes it in software. Then again, RME is doing this all-digital volume control on a $2000 ADI-2 Pro, whereas this is a $250 unit, so I guess I should attenuate my expectations accordingly.

Joke's on you - have you researched the Auto Ref function on the RME? They are switching output stage gain settings (by relay) depending on reference level chosen. (If in doubt, consult the ADI-2 DAC manual, which might be the latest and most verbose in this respect.)

I have to assume that you do not have too much experience with headphone outputs. It is easy to end up with unreasonable demands on DAC performance if both ends of the sensitivity spectrum are to be accommodated well. If you need at least 6 Vrms of output for insensitive 600 ohm cans and 2 µV of noise for silence in 130 dB/V BA IEMs, that's almost 130 dB of dynamic range right there. (And that's not exotic performance - some amps that manage this performance level or higher include the NwAvGuy O2, the JDS Labs Atom, and the Massdrop THX AAA thingy.) Analog, that's about 3$ in parts, a good layout and a modest amount of power. In a DAC, it requires a fancy part and a first-rate implementation all around, with low tolerance for error. Remember how the CMI8788 and related chips would drop to a "mere" 108 dB of dynamic range in loopback upon using 44.1 kHz and multiples? People with sensitive headphones on e.g. the Xonar ST(X) could hear that.

Now, of those 130 dB, at least 20 dB are required just to accommodate different setups. We do not actually require more than 110 dB of instantaneous dynamic range to cover everything from silence to deafening levels. If, then, you manage to shift those 110 dB around as needed, the whole affair becomes substantially more tolerant to DAC implementation.

For an extreme example, the DAC in the AMS SoC used by the venerable Sandisk Sansa Clip+ offers a whopping 95 dB of dynamic range and the poor low-level nonlinearity of late-'80s multibit DACs. Yet, the addition of a PGA for output level control makes it a perfectly acceptable DAP at paltry power consumption levels. A FiiO X1 no doubt has substantially better converter hardware, but it enjoys of reputation of being slightly hissy, no doubt due to the lack of any form of analog volume control.

Lest we forget, even the converter ICs themselves owe no small part of their performance to highly-developed analog sections.

So definitely do not discount analog. Chances are you'll get a more elegant, robust and efficient solution by not trying to brute force a DAC into doing the heavy lifting. Brute force approaches have their merits but elegance tends not to be one of them.

 

[hollowman]

Analog volume control

So definitely do not discount analog.

Back in 2011, at RMAF, Martin Mallinson (CTO at ESS) gave a talk about digital vs. analog volume controls:
"An analog one is the best".
YouTube--RMAF, RMAF11: Digital Jitter and Volume Controls, Martin Mallison, CTO, ESS Technology (ESS)

 

[chris719]

What terrible op-amp selection. OP275 is kind of crap for driving headphones and AD8056 seems like a poor choice for audio. I guess that's the Audio Note "magic".

There is also no point in making a PCIe sound card when it is actually just a USB XMOS device with an Asmedia PCIe USB controller on it. You get none of the benefits of a real native PCI/PCIe interface.

 

[sgrossklass]

Back in 2011, at RMAF, Martin Mallinson (CTO at ESS) gave a talk about digital vs. analog volume controls:
"An analog one is the best".
YouTube--RMAF, RMAF11: Digital Jitter and Volume Controls, Martin Mallison, CTO, ESS Technology (ESS)

I'll have a look at that.

What terrible op-amp selection. OP275 is kind of crap for driving headphones and AD8056 seems like a poor choice for audio.

They do seem to have some buffers for the headphone out, I'm guessing LME49600. I'm not a huge fan of the OP275 either, but I could see it outperform a TI NE5532 under the right circumstances. AD8056 seems to be one of these LM6171-ish jobs - possibly not ulralow distortion in the audio band but generally unimpressed by high-frequency noise and as such probably not the worst part one could follow up a DAC with, assuming you can get the layout right.

There is also no point in making a PCIe sound card when it is actually just a USB XMOS device with an Asmedia PCIe USB controller on it. You get none of the benefits of a real native PCI/PCIe interface.

Recent Asus cards have been similar. This way of doing things does seem rather convoluted. I can only assume that it simplifies driver development or something, and judging from the number of past driver fiascos, consumer soundcard manufacturers can typically use all the help they can get in this department. (Or perhaps you can't get any off-the-shelf PCIe controller chips including the latest features like >192 kHz sample rates and native DSD. Also keep in mind the Xonar STX 2 vs. Ryzen board fiasco with the seemingly incompatible PCIe-PCI bridge chip... one would hope that USB controllers would be less problematic.) Ironically, I bet that no small number of Thunderbolt audio interfaces actually makes use of the PCIe bus, but they tend to be in a different league in terms of cost and support, too...

 

[H713]

The use of a PCIe USB controller may have more to do with wanting it to be a PCIe card to appeal to gamers. Something tells me that the RGB LEDs aren't there to appeal to the crowd on this forum.

In general, the Asus Xonar, Creative Soundblasters and now the EVGA seem to target the gaming crowd. I think EVGA knows they already have a name within the PC / gaming community, while they would have a huge uphill battle if they wanted to compete with the likes of Apogee or RME.

This still looks better than the majority of "gamer" sound cards, as it seems to have some amount of audio performance to back up the excessive bling.