This review page is supported in part by the sponsors whose ad banners are displayed below

Q: Tell us about your choice of Sabre and Tenor. I assume you auditioned a whole slew of chips before you arrived at this particular combination?


A: We were looking for the best DAC chip regardless of cost. After an extensive search we narrowed it down to two, the Sabre 32/ESS9018 and the Wolfson WM8741. After evaluating each for a few months, we selected the much more expensive Sabre32 chip for its higher analytical capacity and relative stability over the Wolfson. Regarding the selection of TE8802, we have worked with Tenor for over two years. Their support has always been incredible and consistent. The TE8802 offers a good mix of performance and reliability.


Q: What exactly was the design brief over the HA-160D - what did you want improved and how did you go about it?
A: We wanted to improve every single aspect of the 3-in-1 design. As you can see, we left no stone unturned. Design objectives included three areas: 1/ select the best DAC chip and realise its full potential with a tailored discrete network. 2/ since the introduction of the HA-160D in 2010, new headphones had flooded the market. We wanted a new headphone amp that can drive them all including the low-efficiency planarmagnetics. 3/ since 2010 the audiophile desktop community has grown. They are no longer confined to headphone listening. Many desktop audio systems now include desktop amplifiers and nearfield monitors. Therefore we wanted to refine the Conductor so that it would functionally suit any audio system. This included an improvement to our preamp and an improvement to the connectivity of the Conductor.


Q: How did you go about optimizing the new output stage? Your website references a fully differential input stage yet the outputs are single-ended. And do we assume that you run the 8-channel Sabre with 4 channels per side for a multi-parallel configuration?

A: During the R&D process we tried all eight channels in multi-parallel. We found the sound to be overly colored and unnatural. Very ‘digital’ if I may say so. Hence we abandoned the 8-channel configuration and used the main 2 channels, which in our opinion sounds more organic and balanced. More 'Burson' if you know what I mean. The signal output from our Sabre32/ESS9018 is fully balanced. It then goes through a passive resistor-based I/V conversion stage feeding a pure class-A FET differential gain stage which amplifies it to line level. Our real pride lies in this tailored gain stage. Other Sabre32 implementations before us used a common design approach to amplify the DAC's very weak signal to line level. First they implement an output stage consisting  of two IC opamps, one per channel. This is followed by a buffer stage of another two IC opamps or input tubes such as 12AU7. While the IC opamp + tube approach inevitably colors the sound beyond acceptable levels, the IC opamp + IC opamp approach deviates even further from the ideal. 
 
As you know a typical IC opamp has over 50 micro components inside it. Using two such integrated opamps means placing over 100 inferior components into the critical signal path. Each part blocks and distorts the signal so less is definitely more. Our tailored FET gain stage on the other hand achieves line-level amplification with only 25 discrete components directly in the signal path.  Achieving our kind of magnitude of amplification with so simple a circuit is our greatest accomplishment yet. This simple—but not simpler than effective—FET output stage then gets powered by a dedicated 35W transformer and a 5-stage filtering power supply. Only then were we satisfied with the results.


Burson's two DAC finalists were interesting in that Gordon Rankin of Wavelength Audio arrived at the same conclusion when he split the very same vote between his Numerator and Denumerator modules. Daniel Weiss had to pick just one and went with the Sabre. By disregarding six of its channels the Aussies once again swim against the stream however. Where other Sabre champions wave at us improved S/N specs and hard measurements from their multi-paralleled outputs—which for once doesn't require multiple chips—Burson claims it's all the more chipper when just two do the tango. "More organic and balanced, more Burson" is how they put it again. They know exactly what sound they're after even if, perhaps, it sacrifices a percentage point or two in the specs. Listen with your ears, not test gear seems the message.