KEF Model 105/3 Speaker System Review

The KEF Model 105/3, a floor-standing speaker system embodying several advanced and unconventional design features, is the newest member of the company’s Reference Series. A fairly large column-shaped speaker, the 105/3 weighs more than 92 pounds. Its walnut-veneered cabinet, which stands 43-1/2 inches tall and measures 11 inches wide and 16 inches deep, is handsomely finished on all visible surfaces. The cabinet is constructed of 22-millimeter (about 7/8-inch) particleboard and is fitted with four spiked feet with removable plastic caps to protect finished floors. For use on carpeted floors, KEF recommends that the caps be removed once the speakers have been placed.

The KEF 105/3 is a four-way system rated for very high sensitivity, high acoustic-output capability, and exceptional power-handling capacity. Its bass section consists of two upward-facing 8-inch drivers; they are mounted within the cabinet in the double coupled-cavity configuration first introduced by KEF in its top-of-the-line Model 107 speaker system. The drivers’ magnetic structures and frames are linked with a rigid force-canceling rod (about a half-inch in diameter) that is said to reduce acoustic distortion and the transfer of energy from the woofers to the cabinet walls. The combined output of the two woofers is radiated through a smoothly contoured molded duct, about 6 inches in diameter, that acts as an air diaphragm of very low mass.

The other drivers are housed in a separate module, about 10 inches wide and 22-1/2 inches high, that is machined from a solid piece of 3-inch-thick fiber material. This section of the speaker is mounted on the front panel of the cabinet, extending from the top of the bass port to the top of the cabinet. Like most of the front of the speaker, it is hidden from view by the black cloth grille, which is retained by magnetic fasteners and is easily removable. The grille structure weighs a surprising 4-3/4 pounds.

The lower-midrange frequencies are handled by two 6-1/2-inch cone drivers near the top and bottom of the front-mounted module. Between them is a KEF Uni-Q driver, with a concentric 6-1/2-inch cone and 1-inch dome tweeter. The tweeter’s tiny, powerful neodymium magnet allows it to fit within the voice coil and magnet structure of the cone driver, placing the middle- and high-frequency sound sources in the same plane. The Uni-Q design is said to minimize crossover discontinuities and provide similar directional coverage in the crossover region for the outputs of the two concentric drivers.

The symmetrical, vertical placement of the lower-midrange speakers around the Uni-Q mid/high-frequency driver are said to give the Model 105/3 system the highly desirable properties of a broad, uniform horizontal radiation together with a narrower vertical radiation pattern in the midrange, where the acoustic sources of all four drivers appear at the same point in space (the center of the middle driver). This point is about 31 inches from the floor, roughly on a level with the ears of a seated listener.

Although KEF does not include the crossover frequencies in the system’s specifications, the manual points out that the Conjugate Load Matching design of the system presents the driving amplifier with a nearly constant load, almost purely resistive. This permits an amplifier of any type to deliver its best performance, especially when handling complex program material at high volume levels.

The instruction booklet supplied with the KEF 105/3 system provides important information about how to unpack the speakers (which could be a very difficult task if not done in the suggested manner) as well as advice on placement. KEF recommends that the speakers be located at least 20 inches from the back wall and 3 feet from the side walls. The rear panel of each cabinet contains separate gold-plated binding posts for the bass and the mid/treble frequencies, joined by gold-plated metal straps. The binding posts accept bare wires, spade lugs, and single or dual banana plugs. By removing the jumper straps, the system can be operated either in a bi-wired mode, with separate cables for the high and low frequencies fed from a common amplifier, or in a bi-amplified mode, with separate amplifiers for the upper and lower frequency ranges.

The specifications of the KEF 105/3 include a frequency response of 49 to 20,000 Hz ±2.5 dB at a 2-meter distance on the reference axis and within 2 dB from 50 to 17,000 Hz up to 30 degrees off the axis in any direction. The system’s sensitivity under anechoic conditions is rated for a sound-pressure level (SPL) of 93 dB at 1 meter on the reference axis with a pink-noise input of 2.83 volts rms. The impedance is rated at 4 ohms (resistive) from 20 to 20,000 Hz, and the system is recommended for use with amplifiers capable of delivering 50 to 300 watts into a 4-ohm load. Its maximum acoustic output on program peaks is rated at 115 dB SPL under typical listening conditions.

Our test speakers were furnished with an active external equalizer, the KEF Kube 200, designed for use with the Model 105/3’s. The Kube 200 should be installed in a low-level signal path such as a tape-recording loop or between a preamplifier and power amplifier. Pushbuttons on its panel can replace the tape-monitoring function (there are tape-in/out jacks on its rear apron) and bypass its equalizing circuits. It has separate low- and high-frequency control knobs, with center detents, and a small, separate power supply.

The Kube 200 is 6 inches wide, 8-1/2 inches deep, and 2-1/2 inches high. Price: KEF 105/3 speakers, $3,500 a pair; Kube 200 equalizer, $390.

Lab Tests

We installed the KEF 105/3 speakers as recommended, although we left the protective caps on the spiked feet to simplify moving the speakers around on the carpeted floor. We measured their frequency response without equalization. The frequency response of the Kube 200 equalizer was measured separately and added to the measured acoustic response of the speaker.

The room response of the Model 105/3 was very smooth and flat, about ±2.5 dB from 400 to 20,000 Hz. The close-miked woofer response at the front-panel port was ±2.5 dB from 47 to 140 Hz, and the midrange response was ±2.5 dB from 140 to 400 Hz. Combining these curves to obtain a composite frequency response produced a curve flat within ±2.5 dB from 47 to 20,000 Hz.

The Kube 200’s response, with its control knobs set to their detented positions, was flat from 800 to 20,000 Hz. It dipped to —1.5 dB from 100 to 70 Hz, returned to reference level at 56 Hz, and rose to + 16 dB at 20 Hz. Adding this to the measured speaker response resulted in a system response flat within ±2.5 dB from 23 to 20,000 Hz. About 5 dB of additional bass boost between 100 and 20 Hz was available at the control’s maximum setting. At full boost the treble equalization of the equalizer created a gently sloping rise of 4 dB from 500 to 20,000 Hz, and at full cut there was a drop from 500 to 5,000 Hz and a uniform — 6-dB output from 5,000 to 20,000 Hz.

The system’s impedance remained between 3 and 5 ohms from 20 to 20,000 Hz (between 3 and 4 ohms from 20 to 1,000 Hz). The phase angle of the impedance did not exceed 25 degrees over the full audio range, confirming its essentially resistive character.

The FFT response curves of the Model 105/3 were essentially alike at horizontal angles of 0 and 45 degrees off the reference axis, differing in level by 6 dB or less over most of the range up to 10,000 Hz. The two curves did not diverge significantly at frequencies under 13,000 Hz. The phase linearity of the system was very good, with an overall group-delay variation of 0.3 millisecond from 1,000 to 20,000 Hz.

The measured system sensitivity was 94 dB SPL, slightly better than the rated 93 dB (we were not measuring under anechoic conditions). Bass distortion was measured at a drive level equivalent to a 90-dB SPL in the sensitivity measurement. We measured distortion as a function of frequency with the Kube 200 equalizer bypassed, with its middle (detent) bass setting, and with its maximum low-frequency boost. (When the equalizer was used, the drive input to it was kept constant.) Without equalization, the distortion was under 1 percent down to 36 Hz, reaching 4.8 percent at 20 Hz. The middle equalizer setting increased the distortion below 60 Hz, where the equalization became significant. Distortion reached 2 percent at 40 Hz and 5 percent at 36 Hz. Maximum boost, as expected, increased the distortion substantially at low frequencies. With that setting, distortion slowly climbed from 0.4 percent at 100 Hz to 1 percent at 76 Hz and 2 percent at 54 Hz, then rose steeply to 5 percent at 37 Hz and 9.5 percent at 30 Hz.

Pulse-power tests confirmed the huge power-handling ability of the KEF 105/3. We were unable to induce significant (audible) distortion within the power range of the Carver Mono-Block amplifier we use for this test. In each case, the amplifier clipped first, at outputs ranging from 1,835 watts at 1,000 Hz to 1,550 watts at 100 Hz. The latter figure was a record in our experience, since very few woofers will withstand a single-cycle pulse of 100 Hz at power levels of even a few hundred watts without exceeding their suspension limits with a loud rasp. The KEF system sounded clean until the amplifier’s output waveform clipped.

Comments

The KEF 105/3 handily lived up to the claims made for it. Its frequency-response range and flatness were exceptional, with a seamless integration of the bass, midrange, and treble drivers. Although the bass extension of the speaker itself—without the Kube 200 equalizer—was not at all unusual, especially for a speaker of its size and weight, the uniform output down to nearly 50 Hz actually includes almost the entire effective bass content of the vast majority of recordings. In fact, most of the time we found no audible effect from bypassing the equalization. And, without equalization, the Model 105/3 has what may be the lowest deep-bass distortion of any currently available full-range speaker of its size and price.

The impedance characteristic of this system was also the closest to a pure (and constant) resistance that we have seen. Although it is a rather low resistance, because of the speaker’s high sensitivity, practically any amplifier worth being used in a home music system should drive it with ease.

Merely from a measurement standpoint, this combination of qualities—high sensitivity, flat and wide response, and low distortion— would be impressive in any speaker, and especially in one of moderate size. One would naturally expect such a speaker to sound very good indeed, and the KEF 105/3 certainly lived up to those expectations.

At first listening, there seemed to be nothing unusual about the sound of the 105/3—but that, in itself, is unusual. No two speaker models sound alike, and most of the differences between them can be attributed to frequency-response and directional (dispersion) irregularities that color certain portions of the audible spectrum. There are, of course, many other factors involved in this process, especially where the subtleties of sound quality are concerned, but their specifics are the subject of wide disagreement.

To put it as directly as possible, the KEF 105/3 sounded strikingly neutral, without apparent emphasis or lack in any portion of the audible spectrum. There are many speakers that can produce surprising, and sometimes pleasing, effects by virtue of their emphasized bass or extreme treble extension or emphasis. Usually, such effects are program dependent, and these speakers are likely to produce unnatural or unpleasant sounds when playing other material.

Heard through the KEF 105/3, male voices were not unnaturally chesty, strings did not sizzle and scream, the low bass did not boom, and the overall excitement of the listening experience was no greater, but also no less, than that of the music itself. In short, the Model 105/3 was an accurate reproducer. Not surprising in view of its response smoothness and wide dispersion, its stereo imaging qualities were also excellent.

In addition to its impressive technical qualifications, the KEF 105/3 was thoroughly enjoyable to listen to, never requiring the listener to accept any distortions or unpleasant qualities that were not in the program material. It would probably be at its best in a large room, where it could produce a natural listening level; in our 15 x 20-foot room, we were never able to tolerate a peak drive level greater than a few watts without a feeling of being physically assaulted!

Until recently, we used the Model 105/3’s predecessor, the KEF 105.2, as our reference speaker for a number of years. It was (and still is) an excellent speaker, but the contrast between it and the Model 105/3 is a striking demonstration of the advance in speaker technology during the past ten or twelve years. The 105/3 is well worth its fairly considerable cost.

Read all

KEF reviews

Read all