Aurender and Computer Audio Design: Two Different Companies, Same Peculiar Idea

Audio shows are great for hearing new components and systems—of course they are. But they can also be good for encountering new ideas. Over two days at AXPONA, I encountered the same curious idea from two manufacturers.

First came the Aurender suite, with a passive display of many of the company’s server/streamers and server/streamer DACs, which range in price from $6000 (plus the cost of storage) for the A 200, one of three new A-series models (“A” denotes Aurender products with an onboard DAC) up to $24,000 for the N30SA, Aurender’s flagship server/streamer that comes with 8TB of onboard SSD storage plus an expansion slot if you need even more storage.

All the DACs in Aurender’s A-series products can fully decode MQA; in other respects, they vary as you go up the line. The DAC in the flagship A30, for instance, is dual mono and utilizes the AKM AK4497 chipset, with discrete linear power supplies. The A30 has a defeatable volume control, so you can use it with or without a preamplifier, and a built-in class-A headphone amplifier.

Audio servers are computers almost by definition, yet computers are far from ideal for high-end audio: They’re too electronically noisy. Aurender’s software engineer Eric Shim, who I spoke to during my visit, told me that Aurender comes at server design from a non-computer point of view. Aurender is very concerned about electronic noise and noise reduction, including in the power supply.

As I examined various Aurender devices on display through transparent plexiglass covers. Shim described one, the $24,000 W20SE, as a “legacy product”; it has been in the lineup in one form or another for a long time but remains popular with customers; it it was updated to “SE” status 3 or 4 years ago. One thing that sets it apart from the company’s more recent products is its battery-based power supply.

Sitting next to the W20SE on the static display was one of the company’s newer streamer/servers—I’m not sure which one it was, possibly the $6000 (plus storage) N200—in which the power supply looked completely different. Here, Aurender was using an A/C supply, reflecting the company’s latest thinking. Pointing to the battery supply in the W20SE, Shim said, “that’s what you want if you’re really into nuance.” Then, pointing to the newer, A/C supply, he said, “that’s the one for ultimate dynamics and slam.” Makes sense, I thought to myself; batteries, after all, are very quiet—low noise—but battery supplies generally can’t achieve the high rail voltage that an A/C supply can provide.

I had left the room and was walking down the hallway when it hit me: Wait a minute: Those were pure digital components. Their only output is a digital signal, whether it’s AES3 to USB: no analog output stage in sight. How could the power supply have that kind of impact on the sound of a purely digital device?
The following morning, before the opening bell, I met with the folks from CAD (Computer Audio Design). I remember quite clearly the time I spent in the CAD room at the 2019 AXPONA; I thought that room (with Trilogy electronics and Verity loudspeakers) was among the very best in show.

CAD was one of several companies I heard about that experienced shipping-related glitches. The Vivid Audio S12 loudspeakers that they intended to show arrived shattered, rattling around inside their boxes. Late Thursday before the Friday opening, much of the rest of their equipment—including the Aesthetix Mimas integrated amplifier—had not arrived. (Aesthetix’s Jim White carried a second sample with him on the plane. The original shipment arrived in time—barely—and they ended up with two.)

CAD is one of the more original companies on the hi-fi scene, run by CEO and chief designer Scott Berry, an American expatriate living in England, with his wife Isabel. One of CAD’s main products is the TDA1543 Mk.II DAC ($12,000), which uses 16-bit, 1543 DAC chips—16 of them. (Those knowledgeable about digital-audio history will recognize “1543” as the designation of a DAC chip introduced in 1989, the Philips TDA-1543.) A Mk.III version of the DAC is forthcoming; look for a review in Stereophile in the next few months.

CAD also offers a server—the CAT, which stands for CAD Audio Transport ($17,500)—which includes SSD storage, automates CD ripping, and outputs data over USB only.

Where some server companies are employing more computing power for more upsampling and DSP capabilities, CAD has gone in the other direction, aiming to create a server that does as little computing as possible in order to reduce the amount of noise radiated and transmitted over electrical connections, including, especially, the ground line. Which is why CAD’s other products are a specially grounded USB cable (Berry was showing a new model, price not yet determined but probably around $1500 per meter) and several sizes of “Ground Controls), which are passive devices for absorbing noise on the ground line. Ground Controls range in price from $2250 to $29,500. See Michael Fremer’s treatment of the Ground Controls here.

To return to the theme of this post: As I chatted with Scott and Isabel, something he said struck a chord. He was discussing the effect of the new linear power supply CAD has just introduced, now supplied with the CAT server and available as an upgrade to previous buyers, on the sound the server produces. The new power supply is manufactured in the same facility that produces Constellation Audio amplifiers. “It’s amazing how much difference it makes,” Isabel said. “It always surprises me,” added Scott, “how in a completely digital product does very similar things as it would do in an analog device—pace, rhythm, timing—even though there’s only USB coming out of that; there’s no analog.”

So there it is: The same point made independently by two people from different companies (Aurender and CAD) over the course of 24 hours: A power supply upgrade on a purely digital component can achieve much what it would in an analog device.

Cynics, of course, will be cynical. Those with open minds will go out and listen for themselves.