Company Address: 5740 Green Circle Dr., Minnetonka, Minn. 55343, USA.
Audio Research has been a top high-end preamplifier manufacturer for many years. The brand name has been synonymous with advances in the design of tube and phono preamps. While most other firms focused on transistors, Audio Research helped bring tubes their finest hour.
Although Audio Research's top of-the-line products are still tube units, the company has worked with transistor designs for many years. It also has been one of the first firms to focus on simplifying the signal path and to discover that reducing the number of active and passive components in the signal path, and improving the components' quality, has a perceptible impact on sound quality in a high-resolution system.
The LS3 B line stage preamp is a logical extension of that experience. It is a transistor unit, and as its name indicates, it does not have a phono gain stage. Audio Research does, however, make two separate phono preamplifiers that can be used with the unit. The LS3 B sells for $1,995 with a balanced output and $1,495 without. Interestingly, Audio Research shows its tube heritage by making a higher priced hybrid tube/transistor unit with very similar features, the LS2, which costs $2,995 with a balanced output and $2,495 without.
The LS3 B emphasizes simplicity, eschewing the host of switches and features common on many mid-fi preamps and receivers. The main controls are for gain (in 31 steps), balance, and mono/stereo, with an additional knob allowing selection of five high-level inputs. There are for balance, mode, and even input selection, and it activates (via a relay) a pair of "Direct" inputs on the rear panel that enter the circuit just before the volume control. This provides a way of getting a signal with the least possible coloration from a CD player or D/A converter hooked up to that input pair.
I question the practical value of comparing most preamp specifications. The frequency response and distortion specifications of today's good preamps are well within the area where any differences between units are psycho-acoustically irrelevant. The LS3 B is specified as having less than 0.01% distortion at an output of 2 V rms from 2 Hz to 100 kHz, and less than 0.005% in the mid-band. This is impressive, but no more so than for most preamps, and it is the same specification given for the LS2 B.
What may be more relevant is that the rated frequency response is a wide 2 Hz to 100 kHz, ±0.5 dB. Gain is 18 dB, with an optional reduction of 10 dB. There is no gain at the tape output and 24-dB gain at the balanced output. Noise is rated at more than 100 dB below a 2-V output signal with regular outputs and 4 V with the balanced outputs.
Input impedance is 50,000 ohms for most inputs, with 100,000 ohms for the direct input. Output impedance is 250 ohms, allowing the use of long interconnects.
The basic circuit is pure Class A, with a high-voltage supply. No push-pull circuits are used. There are three sections to the active gain stage. The first is a differential input using I-FETs. The feedback drives the negative input of the differential pair, which allows the feedback to include the output capacitor and improves the sound quality. The second stage has a J-FET follower driving a MOS-FET in a common gate, with feedback to limit its gain to 14 dB. The third stage is a source follower with a constant source load. The gain is limited by feedback, and the output capacitor is again included in the feedback loop.
The circuit design emphasizes short, simple signal paths and has d.c.-coupled inputs and a tightly regulated power supply. The circuit hoard uses relay switching and has board-mounted switches. The input-selector knob on the front panel is actually connected, by a long shaft, to a switch at the rear of the circuit hoard, near the input jacks. This input switch configuration, and the fact that other front-panel switches actually control internal relays, allow the LS3 B to eliminate any wires in the signal path except for the oxygen-free copper jumpers from the connectors to the circuit board.
Audio Research emphasizes the quality of the power supply and claims that the electronic line regulation is better than 0.01%. Construction, the layout of the p.c. board, and component quality of this unit are excellent.
This kind of preamp is obviously designed for the audiophile who is creating a stereo system with the best possible CD player (or CD transport and D/A converter), power amplifiers, and loudspeakers. In all frankness, the LS3 B's superior sound will not show up in mid-fi systems using mediocre CD players and colored speakers or in systems where speaker placement and the overall setup do not focus on accuracy in imaging and soundstage or provide a clear path between the speakers and the listening position. Only in a more demanding system does the LS3 B reveal its outstanding performance.
Before I began my listening tests with music, I tested the LS3 B's noise and coloration by inserting it into the tape monitor loop in my reference preamp. I compared the sound of a setup in which I put the LS3 B in and out of the signal path between a Theta Digital Balanced DS Pro Generation III D/A converter and my reference power amplifiers. The LS3 B did virtually nothing to alter sound quality. I noticed only a tiny increase in noise when I listened at high volumes with my ear close to the speaker, and there was an almost imperceptible loss of detail and reduction in dynamics. No active or passive device is ever totally neutral, but this unit's neutrality in these two tests ranks with some of the most expensive line stage preamps I have reviewed.
I also compared the LS3 B to a totally passive custom line stage preamp that has somewhat similar switching features. As has been my experience in the past, the purely passive preamp was slightly more silent and less colored if all impedances were carefully maintained, if I selected a CD or other digital player with the right output level for the power amp's input, if I used special low-capacitance and low-impedance interconnects, if I did not connect a tape recorder, and if I carefully set the level control in the passive unit to the "sweet spot" that produced the best possible sound.
I regard these conditions as totally impractical for most real-world systems and listening conditions, and the LS3 B outperformed such a jury-rigged approach to listening under virtually all conditions. It was neutral enough so that its added gain and buffering provided far better overall sound quality than the passive device. The LS3 B also worked well with a variety of different front-end devices, power amplifiers, and interconnects, and I did not encounter any interface problems in the course of my testing.
It is difficult to review the individual aspects of the sound quality of a unit this good. The LS3 B's colorations are very slight and often masked by the interactions involved in using it with specific types of interconnects, front-end devices, and amplifiers. The following comments should, therefore, be taken as more of a guide of what to listen for than for predictable nuances you will hear at an equipment dealer or in your own system.
Overall timbre: Flat, with just a slight hint of reduced energy in the lower midrange that was still further reduced by using the "Direct" signal path.
Bass: Excellent. The only qualification I can think of is that some d.c.-coupled units may provide slightly more energy in the lowest part of the lowest octave with the few speakers and subwoofers that can resolve such differences, but such added bass energy may not be more accurate.
Upper bass and lower midrange: Also excellent. At the nit-picking level, perhaps just a slight hint of leanness.
Midrange and upper midrange: Very neutral, with excellent transient life, detail, and dynamic energy. There was just a slight touch of added dryness or constraint of subtle harmonics and transient detail, but no competing transistor preamplifier exists at any price that is not without similar levels of coloration.
Treble: Clean and detailed. No exaggeration or edge.
Dynamics: Excellent and very similar to the best tube units.
Transparency and overall transient detail: As noted, even the best equipment is not perfect.
Soundstage: Very neutral, just a tiny reduction in depth.
Imaging: Very good right-to-left imaging and very revealing of different layers of depth.
Compatibility: No interface or special placement problems. Perfectly reliable and no unusual operating problems.
I did, incidentally, experiment with using the LS3 B in both the balanced and unbalanced modes. As has been my previous experience, the difference in sound quality was very minor provided I used top-quality cables for both the balanced and unbalanced connections and kept the interconnects and equipment free of potential hum fields.
Under these conditions, the real-world benefits of the balanced output were then very slight and might well not be worth an additional $500 to an audiophile on a budget. At the same time, the balanced outputs did provide a minor reduction in noise levels but more audible improvement under less optimal conditions. Like many high end audiophiles, I now use balanced connections as standard practice. Indeed, many of the more expensive line stage preamplifiers have balanced inputs as well as balanced outputs.
In summary, the Audio Research LS3 B is an excellent piece of high-end equipment that can be used reliably in a very demanding system. The only question that I had about the unit after prolonged listening had nothing to do with its sound quality.
My question was whether its hybrid counterpart, the LS2 B, could really provide $1,000 more in sound quality. This is a test you may wish to conduct for yourself.
Given Audio Research's history, you can be sure that both the LS3 B and LS2 B line preamplifiers will be excellent units.
-Anthony H. Cordesman
[adapted from AUDIO magazine/Mar. 1993]
Also see:
Audio Research D300 amp and LS2B Mark II preamp (Auricle, Jan. 1995)