Integrated Amplifier (1982-84)
The AU-D33 Super Feed Forward DC Integrated Amplifier is a medium-power amplifier that has what most competing models don’t – low distortion sound, thanks to Sansui’s remarkable Super Feed Forward System.
Super-Feed Forward DC power amp. Push button input selectors with indicators. Independent tape-record selector. Phono equalizer for moving-magnet and moving-coil cartridges. High filter. 2 tape monitors with bi-directional dubbing.
Power output: 50 watts per channel into 8Ω (stereo)
Frequency response: 20Hz to 20kHz
Total harmonic distortion: 0.004%
Signal to noise ratio: 65dB (MC), 84dB (MM), 105dB (line)
Dimensions: 430 x 112 x 334mm
The Sansui AU-D33 integrated amplifier features the “Super Feedforward” circuitry first introduced in the company’s higher-priced amplifiers a couple of years ago. “Super Feedforward” is a combination of a relatively small amount of negative feedback with “feedforward” circuits. Feedforward is also a distortion-reducing technique that has certain theoretical advantages over negative feedback (among them being its ability-in theory-to eliminate distortion completely instead of merely reducing it).
The AU-D33 is rated for very low distortion-less than 0.004 per cent for all power outputs up to its rated 50 watts per channel into 8-ohm loads between 20 and 20,000 Hz. It uses direct-coupled circuits from the high-level inputs to the speaker outputs. The program source is selected by pressing one of a group of vertical bars in the upper center of the panel (for tape1, tape 2, aux, tuner, and phono); each bar has a light above it to show that it has been pressed. Volume is controlled by a large knob to the right of the input selectors, and there is a balance control immediately below it.
Smaller knobs across the lower part of the panel operate the bass and treble tone controls, which can be inserted into the signal path by pressing a small button to their right (a light above the knobs shows that the tone controls are active). Below the input selectors is the rec selector, which selects the program going to the tape-recorder outputs (independently of what is being played through the speakers). It has positions for aux, tuner, phono, and off, in addition to two dubbing positions for copying tapes from either of two tape decks to the other.
Small pushbuttons control the two sets of speaker outputs (lights on the panel show their status), the high filter, loudness compensation, and muting (a 20-dB gain reduction). The phono-preamp gain and input impedance can be switched for either MM (moving-magnet) or MC (moving-coil) cartridges by a pair of buttons on the panel. Power is controlled by a larger button above which is a red protector light that serves as a pilot light and blinks when the amplifier’s protection system has been activated (which happens for a few seconds each time the amplifier is turned on, until the operating voltages have stabilized and the speakers have been connected by the protection relay). There is a headphone jack below the power switch.
On the rear apron of the Sansui AU-D33 are the various input and output signal jacks, two a.c. outlets (one switched), and two pairs of speaker outputs. These resemble plastic insulated binding posts, with holes into which the stripped ends of the wires are inserted. When the post is rotated about 120 degrees clockwise, the electrical connection is firmly made. The Sansui AU-D33 is finished entirely in black. It is 17 inches wide, 13-1/4 inches deep, and 4-1/2 inches high, and it weighs 16 pounds. The suggested retail price is $350.
The top of the Sansui AU-D33 (above the internal heat sinks) became quite hot during preconditioning, although it never became more than mildly warm in normal operation. With both channels driving 8-ohm loads at 1,000 Hz, the outputs clipped at 63 watts per channel, corresponding to an IHF clip-ping-headroom rating of about 1 dB. The amplifier is not rated for driving any impedance other than 8 ohms, and we found that the clipping output into 4-ohm loads was 38 watts per channel, and into 2 ohms it was only 18 watts per channel. In dynamic-power measurements, the maximum output into impedances of 8, 4, and 2 ohms was 77, 35, and 18 watts, respectively, so that the amplifier’s IHF dynamic-headroom rating was 1.85 dB.
Harmonic distortion was very low, typically under 0.002 per cent at 1,000 Hz for all power outputs under the clipping point. Even operation with lower load impedances did not increase the distortion. With 8-ohm loads, the distortion at or below rated power was between 0.002 and 0.003 per cent from 20 Hz to 2 or 3 kHz, rising to a maximum of 0.014 per cent at 20 kHz and rated power. Excellent high-frequency linearity was demonstrated by the two-tone intermodulation test of the EIA RS-490 standard. We used 18- and 19-kHz signals of equal amplitude whose peak level was equal to that of a 50-watt sine wave. The second-order (difference-tone) distortion at 1,000 Hz was – 96 dB (the residual of the test equipment), and third-order distortion at 17 kHz was not detectable down to our – 100-dB measurement floor (relative to the 50-watt reference power output).
The amplifier tone controls had conventional characteristics, with the treble curves hinged at about 1,500 Hz and the bass turnover frequency shifting between about 100 and 400 Hz as the control setting was varied. The loudness compensation boosted both low and high frequencies, and the full effect of the circuit came into play by the time the volume had been reduced 20 dB from maximum, with no further change at lower settings. The high-cut filter had an ineffective 6-dB-per-octave slope, and it was 3 dB down at 12 kHz.
The RIAA phono equalization was accurate within 0.5 dB overall from 40 to 20,000 Hz and was down 1.5 dB at 20 Hz. It was not affected significantly by cartridge inductance. The MC-input setting produced a slight rolloff at the lowest frequencies (beginning at about 60 Hz), with a maximum change of about 3 dB at 20 Hz compared to the MM response.
The amplifier sensitivity for a reference output of 1 watt was 30 millivolts (mV) through the aux input, 0.37 mV through phono (MM), and 0.04 mV through phono (MC). The MM phono input overloaded at about 160 mV at 1,000 Hz and lower frequencies and at the equivalent of 136 mV at 20,000 Hz. The A-weighted noise level, referred to 1 watt, was -82 dB for aux and – 79 dB for phono (MM). The phono (MM) input impedance was 51,000 ohms in parallel with 150 picofarads.
The amplifier slew factor was greater than 25, and its rise time was about 6 microseconds through the aux input. The amplifier was stable driving reactive simulated-loudspeaker loads and had an IHF reactive-load rating of 0.83 dB at 63 Hz. The protective circuit, which shut off the outputs when the amplifier was presented with a potentially dangerous load, worked perfectly, so that our initial concern about overdriving the AU-D33 at ultrasonic frequencies or using too low a load impedance was soon alleviated. Although the amplifier always shut off before damage could occur in abnormal operation, it was not “touchy” and never turned off in normal operation, no matter how hard we drove it.
The most obvious benefits claimed for the Sansui “Super Feedforward” design are extremely low distortion and very wide bandwidth. The low distortion was very much in evidence in our tests, although the measured bandwidth was not unusual for an amplifier of this caliber.
The tendency for many Japanese amplifier designers to limit the output-current capabilities of their products is clearly shown in the performance of the Sansui AU-D33, which appears to behave like a “constant-current” source instead of the “constant-voltage” operation usually associated with transistor amplifiers. With a low load impedance, either from 4-ohm speakers or two pairs of 8-ohm speakers, its power-output capability falls appreciably. However, when driving 8-ohm loads, it is all one could wish for in a moderate-priced amplifier. It is a very handsome, high-performance, medium-powered product, with sound quality to match.