Classe Audio DR-3 Amp (Equip. Profile, Jul. 1986)

Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting


Departments | Features | ADs | Equipment | Music/Recordings | History

Manufacturer's Specifications

Power Output: 25 watts channel into 8-ohm loads, 50 watts/channel into 4 ohms, 100 watts/channel into 2 ohms: bridged mode, 100 watts into 8-ohm load, 200 watts into 4 ohms, 300 watts into 2 ohms.

Output Current: 15 amperes rms.

Frequency Response: At rated power, 0.4 Hz to 80 kHz, +0,-3 dB

Distortion: 0.1% THD and IM at rated power.

S N: 85 dB. unweighted, wide-band.

Input Impedance: Stereo, 100 kilohms bridged, 25 kilohms.

Input Sensitivity: 0.8 V for rated output.

Power Consumption: Approximately 500 watts at idle.

Dimensions: 19 in. W x 10 3/4 in. H x 14 1/4 in. D (48.3 cm x 27.3 cm x 36.2 cm).

Weight: 70 lbs. (31.8 kg).

Price: $2,895.

Company Address: 227-G Brunswick Blvd., Pointe Claire, Que., Canada H9R 4X5.

The Canadian-built Classe Audio DR-3 is a solid-state stereo power amplifier rated at 25 watts per channel into 8-ohm loads. Physically, the unit employs exotic construction.

Larger than most 200-watt amplifiers, the DR-3 features huge heat-sinks, massive speaker connectors, and Class-A output circuitry.

The DR-3's 25-watt, 8-ohm power rating; Class-A operating mode, and solid, massive metalwork are reminiscent of the Levinson ML-2, although the DR-3 is a stereo amplifier and costs far less than the Levinson (which now lists at $8,848 per pair and must be custom ordered). Like the ML-2, the DR-3 sports an engraved black faceplate, over sized handles, and sculpted, finned radiators that give it a striking appearance.

The DR-3, Classe Audio's latest "pure" Class-A design, features heavier gauge internal wiring, floating-ground power-supply modules, polyfilm capacitors to couple and by pass key circuit areas, shield plates to reduce hum in the input circuitry, increased input impedance in the bridged mode to interface better with tube preamps, and better speaker connectors than their previous amp.

Construction

The DR-3's construction suggests the very high quality of expensive test instruments. It is built the way amplifiers were usually built in the tube days--on and inside a box chassis. The welded, 1/8-inch-thick aluminum chassis is 12 x 17 x 3 inches and has a steel bottom cover attached by eight Allen-head machine screws. A 10 1/2-inch-high, 1/4-inch-thick, notched rack-mount faceplate is bolted to the front of the chassis and to the front heat radiators. Classe Audio says the amp can be rack-mounted if ventilation room is provided.


--------With 1,695 square inches of heat-sink radiating surface, the DR-3 runs reasonably cool without a fan in Class A. The large central box holds the power transformers.


-----------The DR-3's components nestle within a welded chassis of 1/8-inch-thick aluminum. Note the four 20,000-uF capacitors in each channel's power supply.

The DR-3's front panel bears only a single LED pilot light and a toggle switch for power. The rear apron has silver plated signal input connectors and extremely heavy, silver plated speaker binding posts. Six 8-ampere fuses (accessible from the rear), one a.c. line fuse, and two rail fuses per channel are the amplifier's only protection. The rear of the DR-3 is also equipped with an IEC power socket and a detachable line cord.

The DR-3 runs reasonably cool in Class A without an internal fan because of four exposed heat-sink assemblies, mounted on top of the chassis, which provide 1,695 square inches of radiating surface. A welded aluminum box. mounted on rubber bushings to minimize vibration, runs from chassis front to back and contains the amplifier's two power transformers, one for each channel's power supply. Each 350-VA power transformer is wound with wire that has a square cross-section; this is said to yield higher flux density and greater efficiency.

Underneath, the amplifier's dual-mono circuitry resides on a total of nine p.c. boards, including one a.c. line board, four heat-sink module boards, two power-supply boards, and two main-drive circuitry boards. Four 20,000-uF, 50-V capacitors are used in the power supply for each channel.

Two are input capacitors which feed second-stage filters through 0.4-ohm, wire-wound resistors. The four capacitors' common terminals are bolted to a thick silver-plated strap, the center of which serves as a ground reference point. These p.c. boards are liberally bypassed with large film capacitors.

One of the two audio circuit boards contains additional components for bridged operation. The bridging switch features three DPDT sections which completely disconnect these components from the rest of the circuit and the power supply when in two-channel mode.

For each channel, the DR-3 output stage employs four high-speed, high-power bipolar devices. Claimed total out put-device capability per channel is 800 watts and 120 amperes. The four heat-sinks are drilled, not punched, to help maintain flatness for tight mechanical and thermal contact with the output transistors.

Four-conductor, audiophile-grade cables lead from each RCA input connector to the circuit board for each channel.

The right-channel drive circuit board also contains the circuitry and three switches for converting to bridged mono operation. Circuit-board finish and parts quality are among the best we have seen in home audio components, and are reminiscent of some older military equipment. Trimpots are multi-turn and sealed, capacitors are molded polyfilm types, and small resistors are all 1% film. Three high-quality electrolytic capacitors are also used in each channel.

Large, modular, molded-nylon connectors are used to carry high-current signals to the output transistors. While these are undoubtedly up to the task, we would have liked soldered connections here. Each p.c. board also is terminated through a 12-pin, gold-plated connector said to be able to carry four times the current of the conventional card-edge contact.

The DR-3's massive, silver-plated, brass speaker terminals deserve special mention. Speaker wires are attached by tightening down a 1/4-inch hex-head bolt on the terminal's solid base. If the speaker cable is terminated in sufficiently large spade lugs, this arrangement provides an excellent mechanical and electrical connection between amplifier and speaker wire. This speaker terminal is the best we have encountered in an audio amplifier.

Dave Reich, Classe Audio's president and designer, says each DR-3 undergoes strenuous burn-in and quality-control testing at the factory after running with no input signal for 24 hours. The amplifier's input is then driven with a square wave, logarithmically swept from 5 Hz to 50 kHz, at full power into 4 ohms. Stereo sweep and bridged sweep are carried out for 24 and 12 hours, respectively. A detailed test sheet for each unit is kept on file.

The DR-3 owner's manual contains no standard test-bench ratings. Reich believes that standard power ratings and distortion figures do not describe the DR-3's sonic qualities. For this review, manufacturer's specs were obtained by means of not-so-subtle editorial pressure. Reich confirmed that both the company's former amplifier, the DR-2, and the DR-3 have identical power output and frequency response ratings into 8 ohms (25 watts/channel in stereo mode, 100 watts bridged), into 4 ohms (50 watts/ channel in stereo mode), and into 2 ohms (100 watts/ channel in stereo mode).

Circuit Description

The DR-3 is a straightforward Class-A amplifier design. Such circuitry keeps output transistors conducting through out the output waveform and requires high continuous bias current in the output stage. Because the current delivered to the load is diverted from the bias, the peak available load current is about equal to the static bias current. The designer must therefore estimate a maximum instantaneous requirement, based on speaker impedance, and use that as the steady bias current. Designer Reich notes that his "pure" Class-A amplifiers do indeed remain in Class-A operation only so long as their output transistor current is not required to exceed that of the static bias. However, if speaker loads lower than 8 ohms are used, most Class-A amplifiers will supply additional peak current, a condition which Reich says is a shit to Class-AB operation. Shifting out of Class-A operation is not the intent of Classe Audio, particularly when driving the low-impedance loads of some electro static or panel speakers favored by some audiophiles. The DR-3 was therefore designed with a relatively low power rating of 25 watts, so it can maintain its full rating in pure Class-A operation while keeping its size and heat dissipation realistic.

The DR-3's circuitry uses nine small, plastic-case transistors on each board, with one extra on the right-channel board for bridging the two channels. The power supply employs two 35-amp bridge rectifiers followed by the two-stage RC filter network; this provides a high power reserve with low ripple. Total power-supply capacitance is 160,000 uF, which is equivalent to 250 joules of energy storage.

Silver-plated, solid copper buss bars; stiff power transformers, and capacitors with low equivalent series resistance all contribute to a circuit with extremely low power-supply impedance.

Discrete bipolar transistor circuitry is used throughout the DR-3. The differential input stage is current-sourced and cascoded. Reich says that the cascoding topology in creases the input impedance and effectively "elongates" the linear operating region of the devices. The driver transistors are also cascoded, and their current-source diodes monitor the output-stage bias current. All the devices are matched for gain and checked for leakage. Small resistors are metal film, 1% tolerance; emitter resistors and other large resistors are custom-made wire-wounds rated at 10 watts. Each fully discrete amplifier circuit is mounted on a glass epoxy p.c. board and soldered with silver-content solder. The p.c. board manufacture is computer-executed, and the plating pattern is sealed with an epoxy solder resist.

No intermediate-size driver transistors are used, because the Darlington output devices contain their own drivers. Also unusual is the virtual absence of thermal feedback from the heat-sinks to circuit parts. This caused a slow bias shift of more than 50% over a two-hour warm-up period from turn-on to "full Class-A" operation.

Measurements


Fig. 1--Power output vs. frequency at 0.3% THD + N, stereo and bridged operation, for 8-, 4-, and 2-ohm loads.


Fig. 2--Power output vs. THD + N at 1 kHz for 8-, 4-, 2-, and 1-ohm loads.


Fig. 3--Power output vs. THD + N at 20 kHz for 8-, 4-, 2-, and 1-ohm loads.


Fig. 4--Clipping characteristics (top trace) for 20-kHz input (bottom trace). Note the absence of "sticking" as the trace leaves the flattened peak area. Input is 2-V sine wave. output is 50 watts into 4-ohm load. Scales: Horizontal, 100 uS/div.; vertical, 10 V div. (top trace) and 2 V div. (bottom trace).

The DR-3 was first run for one hour at 33% of rated power, about 8.3 watts per channel into 8-ohm loads with a 1-kHz test signal. The usual FTC specification of one-third power does not strain a Class-A design the way it can a Class-AB amplifier. Because a Class-A amplifier is fully on all the time, power is either radiated as heat or is delivered to the load. Thus, the Classe can run cooler in operation than at idle.

The output-stage bias of the fully warmed-up amplifier, measured as a voltage drop across an output-stage emitter resistor, was found to be 2.67 amperes. When first turned on, the emitter-resistor current flow measured only 1.56 amperes, which means that only a narrow Class-A power range is available when the amplifier is cold. From cold turn-on, the amplifier heated up noticeably for 10 minutes and required two hours to fully stabilize.

Class-A operation limits were studied by noting the power-output level at which the a.c. line-current draw began to increase. Theoretically, the amplifier's line-current draw should remain constant as long as it is operating in Class A.

As expected, the DR-3's current draw did not vary notice ably with an 8-ohm load up to the 40-watt clipping point. For other load impedances, the draw on the a.c. line current shifts as the amplifier transfers into Class-AB operation. The DR-3 delivered approximately 30 watts (11 V) into 4-ohm loads before line-current draw increased. At lower impedances, not often encountered in loudspeakers, the DR-3 settled into Class AB at progressively lower powers, including 16 watts (5.7 V) into 2-ohm loads and 7.8 watts (2.3 V) into 1-ohm loads.

Power output was measured from 20 Hz to 20 kHz into a variety of load conditions, as shown in Fig. 1. From 20 Hz to 20 kHz, at less than 0.3% THD + N, the DR-3 delivers a minimum of 35 watts per channel into 8 ohms. 36 watts into 4 ohms, and 23 watts into 2 ohms. However, as Fig. 1 shows, power output into 4-ohm and 2-ohm loads is substantially higher at frequencies up to 2 kHz. Driving 1-ohm loads, the amplifier pumps out an amazing 192 watts per channel at mid-band frequencies, although at much higher distortion levels.


Fig. 5--Large-signal square-wave response at clipping for 20-kHz input (top trace) and 20-kHz input signal (bottom trace). Output is 25 watts into 8 ohms. Scales: Horizontal, 10 µS/div.; vertical, 10 V/div. (top trace) and 1 V div. (bottom trace).

In bridged (mono) operation, the DR-3's continuous power output at mid-band measures 158 watts into 8 ohms, 256 watts into 4 ohms, and 40 watts into 2 ohms, all with THD + N of less than 0.3%. At very high frequencies, however, power drops considerably (Fig. 1).

Running the amp at its rated 25-watt output into 8-ohm loads, THD + N was steady from 20 Hz to 2 kHz (at 0.020% to 0.022% in the left channel and 0.032% to 0.034% in the right). Above 2 kHz, THD + N began rising, reaching a maximum of 0.12% at 20 kHz; this rise suggests increasing nonlinearities or reduced negative feedback in the high-frequency range. Measured THD + N ran 0.36% for 50 watts output into 4-ohm loads, and 0.74% at 100 watts into 2 ohms. Although these distortion figures are somewhat higher than usual for amplifiers in this price category, measurements at lower power levels revealed much better performance. Figures 2 and 3 show the DR-3's distortion versus power output, with four impedance loads, for 1- and 20-kHz signals, respectively. Note that at 1 kHz, distortion stays below a respectable 0.05% until the power is raised to the point where the DR-3 leaves Class-A operation for that load impedance.

When brought to clipping level at 20 kHz, the waveform flattened on top and bottom, as expected. This happens in all amplifiers when the output transistors have pulled the load up or down to he power-supply or "rail" voltages and more power is asked for. The DR-3's clipping (Fig. 4) is very clean, with no "sticking" and no power-line frequency ripple even at 6-dB overdrive. Many other amplifiers, once brought to the rail voltages, tend to stick there for a few microseconds, generating excess distortion before jumping back to the proper waveform.

Voltage gain was measured and found to be 25.2 dB. The IHF sensitivity, for 1 watt into 8-ohm loads at 1 kHz, was 150 mV. The IHF signal-to-noise ratio, which is A-weighted noise referred to 1 watt output into 8 ohms, measured-83.0 dBA for the right channel and-77.8 dBA for the left. Crosstalk versus frequency was measured by driving one channel and measuring the leakage into the other, with the unused input terminated by a 1-kilohm resistor. Worst case, cross talk was better than-60.0 dB from 20 Hz to 20 kHz.

Figure 5 illustrates the DR-3's response to a square wave at rated power of 25 watts per channel into 8 ohms. Rise-time is 4µS. The slew rate measured 11 V/uS, up or down.

This modest-appearing figure is actually very good, considering the amplifier's low power rating. The IHF slew factor into 8 ohms measured "infinite." Adding a 2.0-µF capacitor causes the usual ringing of the output network, with a 0.2-dB increase in sine-wave output at 20 kHz, but no instability.

For 8 ohms, the low-frequency damping factor was measured at 320 and the wide-band damping factor was measured at 15.7. The damping factor versus frequency is shown graphically in Fig. 6. Classe Audio gives no rating for this specification.

The 1-watt frequency response into 8 ohms (Fig. 7) showed the amplifier to be within ±0.1 dB from 5 Hz to 20 kHz. The-3 dB frequencies were 0.28 Hz and 90 kHz.

Input impedance was somewhat frequency-dependent, measuring 140 kilohms at 1 kHz and 62 kilohms at 20 kHz.

The IHF dynamic headroom measured 2.3 dB (18.4 V, 42.3 watts) relative to 25 watts rated power into 8 ohms. The 4-ohm IHF headroom (using the manufacturer's claim of 50 watts into 4-ohm loads) was 2.0 dB (17.9 V, 80.1 watts). The bridged 8-ohm IHF headroom (relative to the 25-watt stereo rating) reached 5.0 dB into 8-ohm loads (35.8 V, 160 watts) and 4.6 dB (for the 50-watt rating mentioned above) into 4-ohm loads (34.0 V, 289 watts). Into lower impedances, the DR-3's available IHF dynamic headroom was far greater, with the amplifier putting out 432-watt peaks (29.4 V) into 2-ohm loads.

Our standard test of peak output current utilizes a 20-mS pulse (repeated at a 0.5-S rate) driving one channel of the amplifier into a 0.1-ohm load. Under these conditions. the DR-3 delivered pulses of 15.0 amperes rms from both channels without apparent strain. The amplifier was not pushed to overload. because it employs no current limiting and we wished to listen to it. not break it! Fifteen amperes is well over the maximum conceivable requirement from a 25-watt amplifier. Even so, the amplifier's rugged construction probably would have withstood an all-out test of current delivery.

but we were satisfied with the measurements we took. There are few other amplifiers rated below 100 watts, let alone 25 watts. that can deliver this much pulsed current output.


Fig. 6-Damping factor vs. frequency for 8-ohm load.


Fig. 7--Frequency response into 8-, 2-, and 1-ohm loads.

Use and Listening Tests

Equipment used by coauthor Greenhill to evaluate the DR-3 included a Linn Sondek turntable, a Magnepan Unitrac 1 tonearm, a Shure V15 Type V-MR cartridge, a Magnavox CD player, a newly modified Mark Levinson ML-7 reference preamp, and Snell Type A-III and Apogee Scintilla speakers. Coauthor Clark's system included a Sony CDP 650ESD CD player. a Tandberg 3008A preamplifier, and Fried Studio IV and Magnepan MG-IIIa speakers.

Greenhill ran four controlled, double-blind tests with a single DR-3 gain-matched to an Onkyo M-510 amplifier rated at 300 watts. The Onkyo's meters, which are quite accurate peak-reading analog types, allowed the author to set the M-510's maximum power to equal the DR-3's. Green hill obtained 13 out of 32, 18 out of 32, 16 out of 32, and 30 out of 64 correct identifications. We don't consider these near 50:50 scores to have statistical significance, which means to us that Greenhill's ability to distinguish these two amplifiers in a double-blind test was no better than would be expected from chance alone.

For the subjective listening sessions, a DR-3 was auditioned by Greenhill on the Snell dynamic loudspeaker systems and later on Apogee Scintillas. The first impression was the change in the DR-3's sonics as it warmed up.

Initially the amplifier produced a thin sound, somewhat distant in perspective and shallow in depth. Within 15 minutes there was a dramatic increase in dynamic range and, for want of better terms, in richness and sweetness. These qualities also produced a startlingly valid sonic replica of the human singing voice. The resonance of strings and woodwinds in orchestral music and the pitch of the human singing voice had a new and different tonality not previously heard on the Snell Type A-III loudspeakers. Center-fill be tween the speakers was seamless, and depth rendition dramatic with two DR-3s. The effect of this biamplified mode was to expand the sound field's width and depth, as well as to increase the resolution of sonic detail.

The biggest subjective surprise was the amplifier's dynamic range. The rated power of 25 watts invited comparison with more powerful amplifiers, with the expectation that the DR-3 would clip audibly while the others sailed on. Not so. The DR-3 played to moderately loud levels with no audible signs of clipping such as static, strain or distortion.

With the 1-ohm load of the Apogee Scintillas, the DR-3 sounded as loud as a 100-watt amplifier (8-ohm rating, not rated for 1-ohm loads) in a non-matched, open comparison.

This is a tribute to the Classe Audio power-supply design, which enables the OR-3 to deliver almost 200 watts per channel into that load. Thus, the DR-3 amplifier "sounds louder" driving low-impedance speaker loads than would be predicted by its power ratings alone.

In summary, we find the Classe Audio DR-3 to represent good quality, reliability, and conservative engineering de sign. The DR-3's high-current power supply allows the amplifier to cope with loads as low as 1 ohm, delivering much more power than its 25-watt, 8-ohm rating would imply.

We'd like to see the manual itself include power ratings at various load impedances. However one views the DR-3's low power ratings and moderate distortion performance on the test bench. it has many intriguing qualities. Clark found the amplifier displayed exemplary clipping characteristics and did not misbehave when overdriven. Greenhill was also impressed with the DR-3, in particular by the high quality of its parts, its rugged construction, dynamic range, and ability to portray instrumental resonances.

-Laurence L. Greenhill and David L. Clark

(Audio magazine, Jul. 1986)

Also see:

Classic Audio CA260 Dual Mono Amplifier (Equip. Profile, Nov. 1989)

Conrad-Johnson Premier Seven-A Preamp and Evolution 2000 Amp (Jun. 1992)

David Berning EA-2101 Amp & TF-12 Preamp (Dec. 1991)

Counterpoint SA-220 Power Amp (Jul. 1990)

= = = =

Prev. | Next

Top of Page  All Related Articles  Home

Updated: Monday, 2019-04-15 16:09 PST