Benchmark Media Systems DAC1

Jun 1, 2003 12:00 PM, By Jim Donahue

When Benchmark Media Systems introduced its latest model digital audio converter, the DAC1, in October of last year, I requested a unit for the station to evaluate.

The half-rack unit has a milled-aluminum front panel with curves, extra-bright LEDs, two headphone jacks, a single volume control for adjusting both headphones and the output levels of the balanced and unbalanced analog outputs by engaging a back-panel switch.

A three-position switch on the rear panel selects a fixed output level, which is adjustable via trim pots, or a variable level that is controlled by the front-panel knob. The middle position on this switch mutes all but the headphone outputs. There are three digital inputs: XLR, optical and BNC, which are selected by a front-panel-mounted switch. The DAC1 accepts AES-3 or S/PDIF formats on any of the input connectors. A BNC-to-RCA adapter is included with the unit because S/PDIF signals are usually carried through RCA connectors.

Internal jumpers can be used to bypass the input selector switch, remove the 75ohm terminator on at the BNC input or add a 10dB, 20dB or 30dB pad to the balanced output for level matching.

The small size of the DAC1 makes it portable. The extra-bright LEDs make it easy to see the level indicators. In Boston's Symphony Hall, we used the DAC1 to monitor the 24-bit/96kHz feed of the Boston Symphony broadcast. The multiple inputs made it easy to switch between monitoring the air signal, the 24-bit/96kHz fiber link and the reference CD.

During a recording session with the Handel and Haydn Society, we used every input and output on the DAC1. The XLR and BNC digital inputs were used to monitor two mixes, while the optical input contained a feed from the conductor's snoop mic (a mic used to permit the producer to listen in on a conductors conversations with the musicians). The balanced output, which was set to a fixed level, was connected via a Krell pre-amplifier and amplifier to B&W 802 speakers. The unbalanced output was connected to a phase/correlation indicator, and the headphone amp was in constant use. We used 64ohm Sony MDR 7506 and 300ohm Sennheiser HD 580 headphones without a glimmer of distortion, or for that matter, any desire for more gain. It's an exceptionally clean high-current headphone amplifier with a peak output level of 23dBu. The headphone amplifier, with a full-scale digital signal, has a distortion level of 0.0005 percent (-106dB) measured at the input of the headphones. The internal analog gain structures are set so that with a full-scale digital input it is not possible to clip anywhere within the analog chain. It is an incredibly loud, incredibly clean sound.

The reason for the unit's clean sound is its electronic design. The oversized toroidal power supply is robust and well isolated, and the secondary supply voltages are heavily regulated. The circuit board is heavily shielded by the top and bottom layers of the multilayer PC board, which are grounded. This also keeps digital-source crosstalk from contaminating the analog signals.

The gremlin of any digital system is jitter. Jitter can degrade an audio signal because of the shift in timing errors it creates. These timing errors create phase-modulated sidebands from the original audio and render digital filters ineffective. These harmonics cannot be eliminated by the cutoff filter, resulting in crosstalk and aliasing. This will lead to increased intermodulation distortion and other unpleasant artifacts. In most professional recording equipment, jitter is controlled by a multistage phase-lock-loop. However, broadcast-grade converters seldom have this circuitry. Rather, they derive their clock source directly from the AES receiver, which has little jitter-reduction capability. That type of design allows jitter to pass directly to the digital-to-analog converter chip itself, with all the resulting nasty effects. Benchmark uses its proprietary technology called Ultralock to control jitter. While the exact workings of Ultralock are not made public, it seems to do the job well. Benchmark claims that a change in performance between using 1,000ft. of digital interconnect cable vs. 2ft. of cable at the test bench is unmeasurable, even with a measurement capability extending to 150dB below full scale. The company also claims a maximum amplitude from jitter sources of nearly 141dB below full-scale digital. During our recording session, we used between 2ft. and 20ft. of cable from our various digital sources and could not discern a difference in performance when switching between the sources.

The manual is extensive, offering graphs of performance. For example, frequency response is virtually flat at ± 0.1dB from 10Hz to well beyond 20kHz. At 43kHz the response is about 1dB. The broadband noise floor hardly rises above the DAC 1 idle noise, which measures 114dB below full-scale digital. A narrow-band FFT of the idle noise shows a noise floor 145dB below FSD with the highest spurious peak at -128dB from the power supply. This is based on curves provided by Benchmark.

An unexpected added benefit is that it isolates the master tapes from accidental static discharge. Especially on cold, dry winter days, removing your headphones or simply standing up while wearing headphones can send what feels like kilovolts of static charge heading back into the recorders and onto tape. The DAC1 does a good job isolating headphone-induced static from the master tapes.