VTL Reference D/A processor Page 2

The VTL converter uses two 6201 dual triodes per channel, which is a military version of the 12AT7WA. The four triode elements per channel provide gain after the DAC, perform active de-emphasis, and form an ultra-low–impedance totem-pole cathode-follower output driver. One of the 6201s is also used in the fourth-order Chebyshev anti-imaging filter, of which two poles are passive, two active.

Gain is required after the DAC because of the custom nature of the UltraAnalog parts used in the VTL converter. A stock UltraAnalog DAC D20400 (as used in the Stax DAC-X1t) puts out 5V at full level, provided by an op-amp internal to the D20400. Rather than use an op-amp for signal gain, VTL designer David Manley specified that the op-amp be removed so he could provide gain with a tube instead (footnote 2). The 0.5V output of the customized DAC is then amplified by one of the 6201 triodes. In addition to removing the op-amp, other changes were made to the D20400 that are exclusive and proprietary to VTL. These include making the part single-channel rather than dual-channel, and using the additional space inside the module to use larger, upgraded passive components. This modified version of the D20400 carries the VTL name and logo.

An interesting aspect of the VTL's design is the attention paid to reducing jitter in the recovered clock signal. Since the S/PDIF interface (a transport or DAT machine digital output signal) incorporates left- and right-channel audio data along with the clock signal in the same conductor, the outboard digital processor must generate a new clock with a Phase Locked Loop (PLL) based on the incoming clock frequency. In a standard implementation of the Yamaha YM3623B S/PDIF receiver chip, between 4 and 5 nanoseconds of clock jitter are introduced, increasing the noise floor and adding enharmonic information in the form of sidebands as well as intermodulation products of those sidebands. The VTL D/A uses a proprietary jitter-reduction circuit to reduce clock jitter and thus improve the unit's sonic performance. For a further discussion of jitter's effects, see JA's sidebar to his Meridian 208 review elsewhere in this issue.

The entire design is dual-mono, from power supply to analog output (footnote 3). Passive component quality is quite high, with polystyrene caps used throughout and many metal-film resistors. Four optocouplers, two per channel, switch in the de-emphasis circuit. It should be noted that these optocouplers are not in series with the signal path. The polarity inversion is performed in the digital domain by inverting the bitstream's polarity. No output muting relay is used, resulting in a very unpleasant burst of noise under the following conditions: when the input selector switch is thrown; if the unit loses lock with the incoming data (as would happen when disconnecting or turning off the digital source); or if the VTL's power is turned off.

Digital converters typically have several relays (output muting, de-emphasis switching), but David Manley feels they compromise the sonic quality as well as long-term reliability. I found, however, that the lack of a muting relay was a serious shortcoming when more than one digital input was connected and it was necessary to switch between them. When comparing transports or alternating between DAT and CD sources, I avoided the problem by selecting a different input on the passive Stepped Attenuator. This liability was mitigated after I got into the habit of de-selecting the VTL's output, but the learning curve's beginning was quite unpleasant.

Overall, the VTL Digital to Analogue Converter is well-built and typically VTL: an emphasis on the primary design aspects that affect sound quality rather than peripheral or esoteric aspects. These latter considerations, espoused by some highly regarded digital designers, include massive ground planes, Teflon pcbs, digital circuit shielding, no signal-carrying pcb traces longer than 2", consideration of air convection currents within the chassis, sonic effects of voltage-regulator style (can or TO-200 package), necessity of having a chassis machined from a solid block of aluminum, and effects of AC line-cord copper purity, to name a few. By contrast, VTL put its design effort into the heart of any D/A converter: the DAC and subsequent analog electronics. This is reflected in the extensive work that went into upgrading the UltraAnalog D20400 (already an exceptional device), to the point that it is became proprietary to VTL, as well as the analog section realized entirely with vacuum tubes.

I was particularly eager to audition the VTL converter for several reasons. First, I wanted to hear the UltraAnalog DACs in a design other than the extraordinary and expensive Stax DAC-X1t. The DAC-X1t provided a qualitatively different musical experience from digital playback, and I wondered if its superb performance was due primarily to the previously unheard UltraAnalog DACs, or to its no-compromise build. Second, I am quite enthusiastic about other VTL products, especially my reference amplifiers, the 225W Deluxe monoblocks. Could the VTL converter equal or better the $5000-more-expensive Stax?

I had not heard the Stax DAC-X1t in my system for several months after completing the review. It found a temporary home in Larry Archibald's system, LA feeling that it was the only converter worth listening to digital through. (Ah, the advantages of being the magazine's publisher.) At any rate, I had become used to listening to the Theta DSPro Basic as my reference digital playback. As I heard the first few seconds of music through the VTL D/A converter, I immediately felt a shock of recognition: the VTL sounded very close to what I remembered the Stax sounding like. These two converters are significantly different in their musical presentations from any other digital processors I've heard, and bear many remarkable sonic resemblances.

First, the presentation through the VTL was very "un-digital." No glare, no graininess, no hard textures, yet significantly, no "tubey" sound that achieves smoothness at the expense of musical detail. The presentation took on a very analog-like ease and liquidity that was like climbing into a hot tub on a cold day. The music flowed with a relaxation and comfort not often associated with digital playback. Instrumental textures were smooth and velvety yet finely woven and highly detailed, creating the sense of ease noted, while simultaneously riveting the listener to the performance with the abundance of detail and musical nuance. It is just this ability to deliver these two often mutually exclusive qualities that puts the VTL and the Stax in a league by themselves. More on this later.

Footnote 2: Comment, from June 1991 (Vol.14 No.6): The VTL Reference D/A converter review in Vol.13 No.12 contained some factual errors concerning the unit's custom UltraAnalog DAC. Based on incorrect information supplied by the manufacturer, I reported that the D-20400's internal op-amp had been removed so that gain could be performed with a tube. In fact, the op-amp was never removed: its gain was merely reduced to provide 0.5V output rather than the standard 5V. In addition, although the DAC is custom-made exclusively for VTL, the extent of the modifications is less than first reported.—Robert Harley

Footnote 3: Does the single line-cord disqualify the VTL as a dual-mono unit?

Vacuum Tube Logic of America
4774 Murrietta, Suites 9 & 10
Chino, CA 91710
(909) 627-5944