Robert Watts of Chord Electronics Ltd. was in Singapore, the home base of Porta-Fi.
Our Editor, Beng Yeow was able to catch up with him to find out more about the legend himself, Chord and its FPGA offerings.
Interview with Robert Watts of Chord Electronics
[BY] Hi Rob, welcome to Singapore. Knowing that it is not the first visit to the island state, maybe you can share with us your favourite local dish?
[RW] Singapore is fantastic as you can get excellent food of all types. But what I can remember is chilli crab!
[BY] Could you tell us more about yourself? Your passion for audio electronics and how the advancement of technologies have influenced your design work?
[RW] I first started assembling electronics when I was a boy; I made a matchbox radio so that I could do my paper round listening to Radio 1 (BBC pop station). Then 40 years later I found myself designing the portable Hugo DAC/amp from the bottom up!
Things have changed remarkably in the last 35 years, and nobody predicted the huge changes that we have seen. Indeed, in the 80’s when digital recordings became available and coupled with my understanding of sampling theory from university, I knew that digital needed 1,000,000 tap filter just to guarantee 16-bit performance from the interpolation filter; and that was inconceivable that it would be ever possible to be able to do this. And today, we have 1M taps with the M scaler.
[BY] When did you know you had a passion for audio electronics?
[RW] I was always interested in electronics and audio and knew from an early age that I wanted to be an electronics designer and to do as much as possible. But my interest in high-end audio was ignited when I heard my first system at the age of 16; my electronics teacher was a Hi-Fi enthusiast, and he demoed a Sheffield Lab direct cut, and I was just knocked out by how real it sounded! That got me hooked.
[BY] Coming from someone who has experienced the transition of analogue to digital electronics, can you share with us what was it like during the initial years? And if today, digital technologies have truly made analogue better?
[RW] Design in the 80’s was simply arduous work. I started as an analogue design engineer and had designed thick film hybrid op-amps for audio – it was 0.5 GHz gain-bandwidth product, unmeasurable (in those days) distortion. But it took months of detailed calculations using pen and paper and calculators. Today, it can be done more accurately with distortion and noise analysis in a day using SPICE modelling.
Digital technologies have not made analogue better – excepting for the design tools and modelling we have today – so designing is very much easier and faster, but the actual design has not changed that much.
[BY] Chord is possibly one of the most known implementers of FPGA for DAC, why FPGA and how different is it from implementing off-the-shelf DAC chips?
[RW] FPGA, as one would understand, is a device which allows an individual to design their own digital circuits as such with FPGA, you will have control over the hardware. Therefore, designing a DAC with FPGA and discrete analogue components allows the ultimate flexibility to maximise performance away from the severe subjective and technical limitations by off-the-shelf DAC chips.
Timing is very important when designing DACs and so are so are taps; especially the number of taps. These are more fully known as FIR taps – put simply, the number of taps is an indication of the amount of memory required to implement the filter and the amount of filtering the filter is capable of. Unfortunately, off-the-shelf DACs contain only a limited number of taps. A filter has a long delay line featuring a bunch of taps along it. If you want to reconstruct the original waveform, you need lots of taps. There’s a theory (which has been proven) that says that, if you want to perfectly recreate the original recording, you must have an infinite tap-length filter. An FIR filter matched to our brain’s processing power would require 1,000,000 filter taps which were realised last year in the form of the new WTA M scaler technology offering an incredible 1,015,808 taps.
Using FPGA also allows us to implement the analogue components discretely as the FPGA is entirely digital and cannot have analogue outputs.
[BY] Is there any weakness when implementing FPGA? How does Chord address these weaknesses?
[RW] I would like to think of it as challenges, when I first started, the Chord DAC 64 (1999) had 1,024 taps, we then progressed to Hugo (2013) 26,000 taps; DAVE (2015) 164,000 taps and Blu MKII (2017) 1,015,808 taps, each successive increase in tap-length coupled with the continuous improvements to the WTA (Watts Transient Alignment) algorithm allowed me to continuously set new benchmarks for better sound quality.
[BY] How different is the FPGA implementation in DAVE and Hugo 2?
[RW] 90% of the technologies implemented on Hugo 2 was referenced or adopted from DAVE. How? We used Verilog, the software to configure or design the digital hardware for FPGA. It consists of defined modules, each with its own specific function; once a Verilog module has been designed, it is then simulated and verified functionally, and checked for timing closure; then it will get listened too and tested in hardware; once a Verilog module has passed listening tests, functional and simulation tests, then it is added to my library, whereupon it may be used for other designs. This re-usability is crucial, as some modules take many years (or even decades) to fine tune and get right. So in the case of Hugo 2, modules that were written for Dave was simply added to the Hugo 2 code, thereby giving an identical digital internal performance.
[BY] Next, we have the questions from our readers. Why did Chord opt to create Poly to allow Mojo to sync up with a mobile device wireless? Why not just incorporate wireless industry standards such as the aptX-HD or LDAC and release a Mojo Wireless?
[RW] I was not involved in the Poly project – but the rationale is clear; we wanted 768 kHz operation, and you can only do that via Wi-Fi – hence the Poly development.
[BY] With the Mojo, Hugo and Qutest, Chord is fast becoming a specialist in electronics miniaturization, would you one day say that the portable variants offered Chord would one day surpass that of the desktop counterparts (the reference range)? As it is today, how much how far in performance (deficit) is the portable products like the Qutest or the Hugo in relation to the reference range?
[RW] While it is fair to say that the portable models have been recognised as top performers when compared with the competition. It is my belief that the desktop models will remain to be better in performance when compared to the portable offerings, factors not limited only to the FPGA implementation but other components which can be better implemented on the desktop model. If we were to just look solely at taps, the Hugo 2 and the Qutest offers 49,152 taps versus the 1,015,808 taps on the Blu MKII which translates to a 20 times capacity advantage.
[BY] Balanced outputs are not a common feature of Chord’s portable products, why is this so? Will Chord ever offer balanced outputs on its portable range?
[RW] At this juncture, there are no plans to work on a balanced design as there is no need to do so. Transparency is for me the most important, component count is very important for transparency and doubling the number of parts in the direct signal path does degrade depth perception and detail resolution. The requirements of power-hungry headphones can be easily addressed on the single-ended connection, and yes, you would need a balanced connection to drive some loudspeakers but not on headphones at the current juncture.
[BY] There has been a lot of talk about the capacity of the FPGA implemented, for example, the DAVE has 10 times the capacity of Hugo in that aspect, what does that really translate to when it comes to DAC?
[RW] As shared earlier, it is my belief that an FIR filter matched to our brain’s processing power would require 1,000,000 filter taps, we have most recently achieved 1,015,808 taps with Blu MKII and the increase in tap length provides a substantial improvement in sound quality.
[BY] In your years of designing products, what was your most fulfilling project and was there ever a project that you would wish you never started?
[RW] This is a tricky question to answer – and it would come down to Mojo, Dave or the M scaler. Dave because it was a research exercise with unlimited time to get the best possible performance, and that had huge knowledge gains that have permeated through current designs. The M scaler because it was a lifetime dream to get 1M taps – and then I got a huge surprise when I eventually heard it! But Mojo is used by an enormous number of people (you could not fit all of the Mojo owners into a UK premier league football stadium now) and is capable (in my opinion) of a level of musicality unmatched by any other non-Chord DAC that I have ever heard. And designing volume products that give high-end performance is an enormous challenge. So, I guess my vote goes to Mojo. But – having said all the above – the project that most changed my listening experience was Hugo – I achieved a level of sheer musical enjoyment that I had never heard before, and my appetite for music changed profoundly.
I have never regretted starting a project, but there have been several projects that never made it to production, and sometimes after several prototypes! But I am not prepared to release a design unless I am 100% happy with it. And the benchmark is not the competition – I have terminated projects that I knew would beat the competition – but my own internal mindset of knowing what is possible. So, some projects were cancelled because I thought I could do better. But my motivation is to learn new things and to get better performance; so, a failed project (that is something that does not make it into production) is not really a failure, as knowledge would have been gained. And my hunger for better sound, learning new things, and coming up with innovative ideas has not diminished with time or with Chord’s huge commercial success.
[BY] Rob, on behalf of our readers at Porta-Fi, I would like to thank you for taking time off your busy schedule to share such insightful details with us and I look forward to catching up with you soon.