Q&A With Wavelength Audio's Gordon Rankin Page 2

How did you first get the idea for an Asynchronous USB DAC?
One of the engineers who worked for me at ICC/DCA was doing some work and wanted a custom audio device. This was like 2002 and what he wanted wasn't really available. So we talked to fab company making USB Audio converters and I looked at the films and said we could laser out this, that and some other stuff and that chip would do the job. It worked well but not great and was adaptive. I took some of the chips and made the first Cosecants and Bricks with them. They were good for the time period but not great.

Wavelength Crimson HS Directly Heated Triode all tube USB DAC

I was talking to Steve Plaskin on a regular basis and he was pushing for something better and so I figured that Asynchronous was really the way to go. The problem was that none of the Computer OS's were really supporting it because it had not really been done by anyone. The first Crimson's were Adaptive TAS1020B based and the software was on the DAC modules so if I did crack the Asynchronous code then we could just change out the board or even the code and it would work.

I linked up with Apple and Microsoft and a couple other companies and then something silly happened. A large company had made a product that had to run asynchronous because their PLL did not work to make it Adaptive. They had a bunch of these made so we all hashed out the details and everyone was pretty much ready to go.

The problem was the TAS1020B was not so easy to get working Asynchronous. Texas Instruments still does not know how to make this chip work Asynchronous. I spent a ton on development hardware and compilers and worked on it for about 3 months before it started to work. At the same time Charlie Hansen (Ayre) had contacted me and wanted to use my code. We really hit it off as we were both saying the same thing: The audio oscillators need to be at the DAC chip and then routed back to the USB controller for the I2S creation.

"Then something happened in my sleep. When I woke up I realized Streamlength™. I sent Ayre and Steve Plaskin new firmware and both of them are like, what the heck did you do!"

So I released a NOS version of the Crimson with Asynchronous Code that only worked at 16 bit stereo/44.1KHz. Next was 24 bit stuff which was a lot easier to do and so that worked and then the QB-9 was running. Then something happened in my sleep. When I woke up I realized Streamlength™. I sent Ayre and Steve Plaskin new firmware and both of them are like what the heck did you do! It was really forward thinking and something that 25 years of protocol design helped out.

"Reception. This desk is great for testing behind it are 8 custom Linux machines which will probably never see the light of day, DLNA/UpNP, latest Airport Express, Sonos, NAS drives, 2 MAC Mini's and two Monitor and Keyboard sets with switches to operate any of the above. The 2 MAC Mini's of course are boot camped with Vista and 7."

Some people claim that USB is an inferior method of connectivity as compared to S/PDIF. Are they are wrong?
Yea!!!!! I hate SPDIF more than anything Digital.

Long story, short answer... SPDIF was created by Philips to do closed box testing of their CD players. Basically the tech would attach the SPDIF cable and the right and left analog cable and then put in a test disk. SPDIF was never meant to do or go where it has.

"But the biggest problem is that SPDIF is not bi-directional and therefore the receiver always has to track the sender by changing the frequency by small amounts which leads to the receiver having a PLL to lock onto the clock and recreate that."

First off let's just talk silly stuff... Networks are simple, they have an impedance and you are suppose to match to that. The connector determines a lot of this in what is called DOD, or diameter over diameter. The RCA cannot be 75 ohms, heck look at a 75 ohm BNC connector and the DOD and then look at the RCA. Then what about an XLR cable, come on 110 ohms at 5vac... it would take a miracle for that to be true. But the biggest problem is that SPDIF is not bi-directional and therefore the receiver always has to track the sender by changing the frequency by small amounts which leads to the receiver having a PLL to lock onto the clock and recreate that. Well as we all know that never works well and is the topic for jitter reduction since day one of SPDIF.

So my point... Why FIX jitter?

johnnya's picture

michael, thanks for a fantastic q/a with gordon rankin; it cuts through the iusual written under brush, with concise, comprehensible information. now, if i could only afford one of his signature products :)


Michael Lavorgna's picture

And thanks for the kind words johnnya.

Steven Plaskin's picture

These were my first impressions of Gordon's asynchronous USB interface for the Crimson I posted at the Audio Asylum on 11/28/2007:

Wavelength Audio has just released Asynchronous USB function for the Crimson USB DAC. The Crimson can be easily updated as the code is on a removal EPROM. Jitter is reduced in this mode. I think that Gordon is the first to have successfully executed this USB function. I will leave it to him to explain the technical aspects of Async Iso USB mode.

The improvement in sound is quite noticeable. Detail, focus, and control are the most obvious changes. For example; listening to James Taylor's new album, One Man Band, JT is now in better focus. Previously, his image was more diffuse and less defined in space. The guitar strings have greater detail. Again, with the previous code, everything was softer sounding-less focused. This is not to say that the former sound was anything less than excellent, but the improvements I am hearing now are quite striking. I also noticed an improvement in the bass; better defined and tighter.

I'll be doing some more listening this weekend and report any other improvements I hear. For now, the Async code is really something else!

earwaxxer's picture

A very exciting time for DACs!

Charles Hansen's picture

Several years ago at Ayre we were interested in doing something with computer audio. There were a number of different approaches to take -- servers (with everything built into one box, including a simplified computer dedicated to audio playback), Ethernet (which the UK companies have focused on but require complex dedicated software, and a strong understanding of computer networking), and USB (which can use any music player software you prefer, and is trivial to hook up and use).

A mutual friend introduced me to Gordon when he was first starting work on the asynchronous USB code. Ayre had never built a separate DAC, because the only way to do it before (and maintain compatibility with other brands) was using S/PDIF, which inevitably adds a significant amount of jitter. When he told me what he was working on, the first words out of my mouth were, "What do we have to do to license your code?"

We hit it off very well, as we really have the same interests and goals -- it's just that he does it with tubes and we do it with solid-state. It took him over a year to work out all the wrinkles, but there was absolutely nobody else in the business who could have done it. Gordon has a super-strong background with the deep-down internals of computers, and he has a real high-end sensibility. Basically my role was to keep egging him on -- when he would get stuck, I would just tell him, "Gordon, you are the smartest guy I know. If you keep working on this, I know that you will figure it out."

And his asynchronous code changed the entire high-end audio industry. For the first time, a computer-based system could sound better than any other digital source. It meant the end of the format wars between DVD-Audio and SACD. Any computer built in the last five or ten years can do 192/24, and now with DoP (DSD over PCM), playback of DSD files are possible with a compatible software player and USB DAC. And as Gordon has since proven with the DragonFly he designed for AudioQuest, great sound doesn't have to cost a boat-load of money.

I've got a few inklings of what Gordon is working on now, and I think it has the potential to be even bigger than asynchronous USB. No wonder he has trouble sleeping at night! Hats off to Gordon for moving the entire industry forward, and thanks to Michael Lavorgna for a great interview. Exciting times, indeed...

Charles Hansen, Ayre Acoustics, Inc.

labjr's picture

I've had Wavelength gear for 20 years. People are still wowed when they hear it. Looks cool too! What impresses me about Gordon is that he goes all out to try new technologies and ideas to make products. Doesn't seem to be like a typical company with budget restraints etc.& trendy ideas.

Should be interesting to see (hear) what Gordon does with the Thunderbolt license. The optical cables would seem to be preferable for their obvious advantages like long length and galvanic isolation.