Linear Tube Audio Aero D/A processor

This paragraph from Linear Tube Audio's website description of their new Aero DAC sets the tone for the story I'm about to tell. "After trying various options, we chose the Analog Devices AD1865 R2R DAC chip, which is sometimes called the 'vinyl DAC,' for its organic sound. It is a non-oversampling DAC, with no digital filters. The AD1865 is much-loved by audiophiles and is used by at least one hi-fi company in a flagship DAC costing over $150,000."

Check the forums and you find that the AD1865 chip is also a heavy DIY favorite. Home brewers are attracted to this discontinued, "obsolete" 18-bit chip for its easy implementation and unprocessed, music-friendly sound.

Manufacturers of luxury-priced DACs love it, too. They don't mind that it's discontinued; they've already bought thousands of chips. The main thing they care about is that it sounds conspicuously musical and engaging. I've listened at length to at least three well-known, very expensive DACs that use this chip, and I would describe their pleasures as the type favored by long-playing record collectors. I'm friends with the designer of one of these "luxury-priced obsolete-chip DACs," and when I asked him why he chose the AD1865, he told me, "We tried all the others, and this one sounded the most right."

"After trying various options," LTA owner-founder Mark Schneider and his Aero design team—Will Schneider-White, Tony N. Nguyen, and Jacob Knibb—apparently came to the same conclusion, but I think they thought they could do "most right" better at a lower cost than those luxury DAC sellers.

Description
I asked Nicholas Tolson, LTA's head of sales and marketing, how LTA's Aero DAC project got started. He replied, "The idea of the first Linear Tube Audio DAC spent years hovering in the background as a blurry sketch, being redrawn regularly. We knew that if we made a DAC, it needed to be deliberate and unique. Two years ago, the pieces started to come together, and we saw that we finally had something that could be the LTA DAC. Now the Aero enters production as a generational step forward for the company. It is the first ZOTL circuit in production that was not an existing Berning design. It is the first single-ended LTA circuit, the first balanced output stage in an LTA product, and the first LTA product without inter-stage coupling capacitors. Along with the Velo [headphone amplifier], the Aero takes a significant step forward in utilizing higher performance SMD components and PCB layout."

Of course I asked about that "obsolete" AD1865 chip. "There is an undeniable kinship between out-of-production non-oversampling DAC chips and vacuum tube amplifiers. The best of these chips scale up in sound quality alongside the right implementation. They also do not add anything that then needs to be filtered or removed." The Aero offers no digital filters, no oversampling. Likewise, there's no DSD or MQA support and no LCD display.

LTA's Aero DAC caught my attention with its twin GE 12SN7 tubes sticking out of its hood. The 12SN7 is the 12V heater variation of the venerable 6SN7 medium mu triode, a tube I have 40 years of experience with. I regard the GE version as one of the best-sounding, longest lasting voltage amplifier tubes ever made. Here, it is used to buffer the output of the digital processer and will probably outlast the DAC and its owners.

According to Schneider-White and Nguyen, "The Aero's analog section begins with a discrete current mirror with a balanced output. This active I/V stage allows the DAC chip to work less hard; and it sounded better than passive circuits during our testing.

"The signal from the I/V feeds the four single-ended amplification stages that make up the balanced ZOTL output stage."

When I installed the Aero, I noticed that its brushed-aluminum back panel is the same thickness and finish as the front and side panels, and its lettering is as nicely engraved. It was obviously designed to be admired as much as the front panel, or as little.

"The casework for the Aero was designed as a nod to classic wooden tube amplifier chassis, re-interpreted in machined aluminum," Nicholas said. "As fewer people use standard equipment racks, especially for headphone listening, we moved away from the basic box-with-a-faceplate formula and created something that can stand alone as an object intended to be seen from all sides. We also wanted to make the tubes more visible than our previous products and allow users to change their tubes without removing the cover."

Don't laugh, but the instant I turned the Aero on, I smiled wide in appreciation for three tiny cream-colored LEDs that look soothing rather than aggressive. I cringe when I see aggressive—non-soothing—chassis designs.

The left brass button and its associated LED handled Power On; the second button and LED were for the Input Selector, choosing between OPT (TosLink), USB, and RCA (S/PDIF). The third LED scoots back and forth in a line just off-center, indicating sampling rates for the incoming signal: 44.1, 48, 88.2, 96, 176.4, and 192. All very classy and understated.

I got so excited when I spied those sweet, warm LEDs that I texted Nicholas asking who made that decision.

"Quite a bit of effort was put into the indicator LEDs. We use the same warm, natural color that you would want for a reading lamp, along with a carefully chosen diffuser-plus-filter arrangement to create the right glow without glare." This is when I began noticing a pattern: LTA was basing design decisions around the quality of the user's whole experience.

Next, I asked Nicholas and LTA CEO Mark Schneider's team for some information about the power supply, which I already knew was a hybrid switching/linear design. He told me he was glad I asked, because "combined, there is over 300,000µF of storage capacitance in the Aero, which is a ridiculous amount for a line-level source component. Power amps often brag about having 1/3 that amount."

I can't swear to the exact sonic effect of this much capacitance, but during my auditions the Aero flashed surprising amounts of drive, dynamism, and clarity—traits I associate with well-engineered power supplies and more expensive DACs.

Perhaps this is a good time to remind readers that lab numbers don't always line up with the character of a component's listening experience, and that sometimes, if we listen to music with an open enough mind, we might discover how many of the things we thought were important turn out to be less important or maybe not important at all.

Starting up
I inserted two American-made NOS GE 12SN7 GTA tubes (supplied), connected an AudioQuest Cinnamon USB cable (connected to my Roon Nucleus+), a Kimber Kable D60 coax, and a pair of balanced Cardas Clear Beyond interconnects; then pushed the Power button and used the second button to select the USB input on the Aero's front panel. My Aero DAC setup was done.

Qobuz recognized the "LTA Aero" immediately. Ten seconds later, I was listening to French violinist Ginette Neveu's The Complete Recordings (24/96 FLAC, Warner Classics/Qobuz).

That's plug'n'play digital. My preferred kind.

COMPANY INFO
Linear Tube Audio
7316 Carroll Ave.
Takoma Park
MD 20912
hifi@lineartubeaudio.com
(301) 448-1534
ARTICLE CONTENTS

COMMENTS
cognoscente's picture

I always find it pleasant and good in a review when the product is compared to other similar devices. And then particular only in the same price range. But also with devices that are 2 or 3 or 4 times as expensive can be interesting (to confirm the strong degree of flattening of the price-quality curve - we only talk about little nuances). In this respect a pleasant and good article. Also because my dac is used in it (the Holo Spring 3 3).

Something that I find annoying, however, is the mention that the chip used in this 4k dac is also used in a dac that costs 150k. That price difference is so absurd that something must be wrong. At least that is what my healthy and yes also suspicious mind tells me. Then there is something wrong with that 150K dac, which makes mentioning it immediately pointless. Mentioning it in this review now feels like an advitorial.

I only have one question, otherwise everything is completely clear to me, will this be in the A+ category next time in the Stereophile Recommended Components?

Jewbacca's picture

I ended up mainly curious what the DAC was. Most of the DACs at the $150K level tend to use FPGA chips (which are basically generic chips that you can program to do whatever - I use some of the same FPGA chips used in several $100+K DACS in specialized industrial equipment), rather than a (what I gather to be, I know little about dedicated DAC chipsets) dedicated, non-programable, DAC chip. (They do this because of low volume; producing a dedicated chipset for a lot of 500 would be insanely expensive.)

My assumption (just due to process of elimination of the FPGA users) was WADAX, but the WADEX is dual mono, and this chip is stereo. Although, I suppose you could either (1) fudge in what you call "dual mono") or (2) use two chips in a rather inefficient manner.

And I only say WADAX because I couldn't find what chipset it uses and it's the right price.

cognoscente's picture

Oh okay, thanks

The question is, how does a 15k DAC sound opposite a 1.5k dac? Not 10x better. At most 15%. Or something about it. So you have to pay 10x more for only 15% improvement! And the question is how does a 150k dac sound opposite a 15k dac? Not 10x better. At most 5%. Or something about it. So you have to pay 10x more for only 5% improvement!

And what technology they use, or materials, or how they measure (poorly an objectification of something that is subjective, enjoying music) does not matter to me. It's about how it sounds at the end of the day. And from where the price increase is no longer in proportion to quality improvement. But this naturally depends on the available personal budget and interest.

Jewbacca's picture

I understand your point, and fully agree that it’s not a “15X improvement”. I’d posit that almost nothing works like that.

Housing, after safe and dry, functioning plumbing, climate control, a workable kitchen and internet connection in a small efficiency compared to a mansion hundreds times its price and size does not “proportionately improve” things.

And my daughter’s Jeep JK that I just bought for high school is not 1/3 the vehicle of my Range Rover, despite being 1/3 the price. Does the Jeep have a 500+ HP engine, massage seats, maneuver surprisingly well, all while being whisper quiet? Nope. But it gets to point A and B well enough, is slow (a feature, not a bug for a kid), and can go down the dirt road to her horses.

There most definitely is a law of diminishing returns, on everything.

The only thing I worry about (especially when over my skis on something like a DAC, that I only kind of know how works, despite working with similar hardware in other contexts) is snake oil, which exists aplenty in the audio world.

I have zero issue overpaying for improvements, if they are improvements.

All that said, does anyone know the $150K DAC that uses this chipset? I’m assuming WADAX, probably two chips, one for each channel. And no, for clarity, I am not opining that the WADAX, is ripping anyone off. A dedicated chip has a number of advantages over FPGA chips in other circumstances with which I am familiar. For example, programmable chips are noisy, inefficient, often hot, and subject to programming corruption. All of this can be fixed with careful layout and isolation. And they offer tremendous benefits of not being locked into data processing that is locked in 10 years or more behind the curve.

Further, there is much more to the process of getting the ones and zeros into audio than the chipset. Most notably, the actual amplification of the faint converted signal into something that can be used properly by the next component. The opportunities to introduce corruption and noise into the signal abound.

My assumption is that this mystery $150k supplier determined the above-mentioned issues with a FPGA chipset were not worth the improvements that can be made with a cutting edge firmware in a FPGA. It’s not a slam, it’s an observation, and I’m curious why they went that direction.

cognoscente's picture

you wrote "I have zero issue overpaying"

That is probably the difference between us, it will have to do with upbringing and background. My grandfather always said "you don't stay rich by spending". I was raised with restraint and modesty. And responsibility.

Sometimes I curse my upbringing and background. Always finding that balance between (responsibly) enjoying and not squandering.

So here certainly no 150K dac, not even 15K but 1/4 of that.

Jewbacca's picture

I need to correct myself. I do have issues "overpaying." I don't, however, mind paying a lot for small improvements in things I care about. I am in place where not only are my financial needs met, but unborn grandchildren's financial needs are met.

My background, btw, is having two parents who survived the Shoa and growing up dirt poor. I am hardly wasteful. But being overly frugal wastes life.

dpetr's picture

I'd bet its an Audio Note Fifth Element

Dave

kafo's picture

I think it might be Audio Note DAC 5. I know they use AD1865 DAC chips.

DunaAudio's picture

It is the Aries Cerat. The name of their reference DAC is Ianus Ithaka. A special model for more than $150k is the Homerus.

https://aries-cerat.com/ianus-series-ithaka-digital-to-audio-converter/

DunaAudio's picture

After reading the information about the Linear Tube Audio DAC on the manufacturer's website, I have the feeling that their device was strongly inspired by the company Aries Cerat.

I would be very pleased if I could see a review and JA measurements on the Stereophile website, at least of their cheapest Helene model. This one is much cheaper than many of the DACs tested at Stereophile.

In Europe, ARIES is currently very popular in the High End sector. It can be said that their DACs are considered Reference Class A+++ far above the flagship DACs of other companies. Stereophile could find out if it's legitimate :-)

Glotz's picture

I'm surprised there wasn't a more flexible way to develop something similar without these resolution limitations, esp. for $4k. (I read the reasons- okay.)

While the 24/96 chip limitations give me pause, LTA has always impressed at the shows, so yes 'deliberate and unique' is acceptable to me. I also trust in HR's listening skills. Great review either way 'cause Herb is speaking his heart and mind.

Even with tubes inline, the measurements seem strange.

'Processor', not er.

supamark's picture

Those are some of the worst measurements I've ever seen for a modern DAC. The jitter is straight out of the 1980's, and the undithered output looks like the scribblings of a madman. Check out Holo Audio's May (KTE) DAC measurements to see what can be done with R-2R.

That AD1865 DAC chip should be taken out back and shot for crimes against music. If someone uses it in a $150k DAC, I have to question the ability of anyone in that company to hear well at all. You can build a literal R-2R network like Holo Audio, or Denefrips (who *does* over/up sample, even in NOS mode) that crushes it in performance for pennies on that $150k.

Glotz's picture

Disappointing coming from such an esteemed mfg.

Ortofan's picture

... for the poor measured performance. Instead, the blame should go to the particular implementation developed by LTA.

The application circuits provided by Analog Devices (for this 30+ year-old chip) show it being used with a separate over/up-sampling chip plus an output reconstruction filter.

In contrast to some ultra-expensive DACs which used this chip, back in the early 1990s the AD1865 was incorporated into CD players from Nakamichi and Parasound, either of which sold for under $400.

teched58's picture

...is so that analog purists can point to the stairsteps and say, "See, I told you so."

supamark's picture

would affect the unditered output mess or the jitter fail... which is what I was talking about. You continue to be tiresome and boring.

Glotz's picture

So glad you are here to sift the bullshit out.

Ortofan's picture

... by an oversampling filter chip not be able to reduce jitter?

https://fastercapital.com/content/Jitter-Reduction-in-DACs--Enhancing-Digital-Audio-Accuracy.html

ok's picture

is actually the audio note uk fifth element/force dac which I have admittedly never heard.
https://www.dagogo.com/audio-note-fifth-force-fifth-element-dac/

ok's picture

to agree.

Jewbacca's picture

I am not really familiar with their products, but my snake oil detector was going off reading their website.

For clarity to the casual reader: WE ARE NOT TALKING ABOUT Linear Tube Audio (the product being reviewed) but rather a curious English DAC that sells for $150K.

ok's picture

products typically have a thick syrupy sound that gets gradually refined as price goes up; their digital expertise is disputed at best; they mostly rely on their transformer/tube analog stages for achieving their goals.

jond's picture

Your characterization of the AN sound is absolutely wrong and they make excellent sounding digital gear at a variety of price points. The presence of a tubed analog stage is a design choice not a panacea.

ok's picture

..wadax reference dac uses a TI ΔΣ chip (probably the 1794A) for conversion purposes only. AD1865 is out of the question because it does not support DSD.

Valter's picture

What produces better quality is related to:
the quality and quantity of the power supplies, better if separated starting from the transformers.
The quality of the output buffer circuit.
The quality of all the components used.
the circuit layout.
Keeping the jitter low.
Focusing only on the conversion chip is shortsighted.

John Atkinson's picture
I wrote in the Measurements sidebar that "without a reconstruction filter, the LTA outputs a broken-looking sinewave (fig.3) because each sample presented to the DAC chip results in a DC output voltage that is sustained until the next sample."

Something I had not considered but was recently reminded of is that while a DAC without a reconstruction filter has a time-perfect impulse response, the actual time resolution is limited to the 11us quantizing interval of the 44.1kHz sample rate.

John Atkinson
Technical Editor, Stereophile

X