Manufacturer's Specifications:
Power Output: 145 watts per channel, continuous, both channels driven into 8-ohm loads. 20 Hz to 20 kHz; 200 watts into 4 ohms. 20 Hz to 20 kHz; 400 watts, bridged mono, into 8 ohms.
Rated THD: 0.05% at 8 ohms; 0.1% at 4 ohms; 0.1%, bridged mono. at 8 ohms.
SMPTE IM: 0.04% from 1 watt to 145 watts at 8 ohms.
Phase Shift: Less than 0.25°, 20 Hz to 20 kHz.
Damping Factor: 300 re: 8 ohms.
Input Sensitivity: 1.6 V rms for 145 watts at 8 ohms.
Slew Rate: 75 V/µS (using 10-kHz, 60-V peak-to-peak square wave).
Rise Time: 0.7 µS (10% to 90%, using 10-kHz, 60-V peak-to-peak square wave).
S/N Ratio: Greater than 100 dB.
Frequency Response: 10 Hz to 50 kHz. ±0.1 dB.
Power Consumption: 150 VA at idling: 1,200 VA maximum.
Dimensions: 17 in. W x 5 1/8 in. H x 10 1/2 in. D (43.2 cm x 13 cm x 26.7 cm).
Weight: 27 lbs. (12.3 kg).
Price: $600 assembled; $525 in kit form; XL-10 test unit. $50.
Company Address: 5910 Crescent Blvd., Pennsauken. N.J. 08109.
David Hafler is one of audio's true pioneers. Like Avery Fisher, Herman H. Scott, Sidney Harman, and a few other outstanding innovators, he was seeking better sound reproduction back when audio was still the hobby of a limited few. Amazingly, Mr. Hafler and his engineers continue to innovate. Taking theft cue from an often-stated goal of audio aficionados, they have come up with the XL-280, an amplifier that comes as close to being the proverbial "straight wire with gain" as any I have yet tested or listened to. Further more, Hafler has devised a simple but clever way of "tweaking" the amplifier so that it continues to act like a "straight wire with gain" even when it is driving real-world speaker loads instead of the resistive loads used in testing. In order to make this adjustment, it is necessary to use a small device that Hafler calls the XL-10. Since the adjustment need be performed only when the amplifier is first installed, you may not want to spend the extra $50 for it, so check with your Hafler dealer. He might very well lend you an XL-10. A readjustment would be necessary only if you changed loud speakers, and even then, the amount of readjustment required might very well be minimal or even inconsequential.
The principle involved in using the XL-10 is easily under stood by looking at the schematic of Fig. 1. One channel of the stereo amplifier to be adjusted is used as a driving amp, while the other channel is subjected to the test and adjustment procedure. First one sets a slide switch to position B and connects a suitable signal (such as interstation noise on an FM tuner, or even a music signal) to the input of the driving amplifier. Then volume is adjusted to a comfortable level as one listens through headphones or a monitor speaker. Next, the slide switch is moved to position D. and the coarse and fine controls are adjusted until a distinct null is achieved. In this switch setting, the direct signal from the driving amplifier is being subtracted from the same signal after it has been fed through the test amplifier channel. The purpose of the fine and coarse controls is to adjust the signal level to the input of the test amplifier so that an overall gain of unity is achieved. The deeper the "null" created as these controls are adjusted, the closer the amplifier under test is to being a "straight wire with gain." It does not matter if the driving amplifier alters the signal in any way, since the output of that driving amp is used to feed the speakers or phones directly and the same output is used to drive the amplifier under test.
In addition to the test for a null, it is possible to conduct a very accurate A/B comparison between the "straight wire" sound (the driving amp's direct output) and the sound of the test amplifier playing the same signal. To do this, you simply alternate between the A and B switch settings. (For a more complete discussion of this type of test, see "Nulling Out Amp Distortion" in the February 1987 issue.) Since these tests measure how an amplifier operates with a specific load, the XL-280 has adjustments in each channel which "tweak" the amplifier to match whatever speaker loads are connected to it. Using these adjustments, you can make the test null so deep that you hear hardly any sound at all when the XL-10's switch is in its D position. I will have more to say about how this adjustment actually worked during my use and listening tests, but first let's take a look at the amplifier itself.
Control Layout and Circuit Highlights
Fig. 1--Schematic of Hafler's XL-10 test fixture.
A large, rocker-type power on/off switch at the lower right of the front panel is the amplifier's only control. An optional rack-mounting kit is available. On the rear apron, I found a mono/stereo bridging switch, gold-plated input jacks for each channel, five-way speaker-cable binding posts, and separate fuse-holders. As received, the fuses in the holders had a rating of only 2 amperes. A quick calculation revealed that these fuses would allow me to measure continuous power levels no higher than 32 watts into 8-ohm loads and no more than 16 watts into 4-ohm loads. The owner's manual points out that these fuses do pass the amplifier's full power on normal musical peaks while still protecting what ever speakers you might be using. However, the manual also indicates that, as far as the amplifier is concerned, one can substitute fuses rated at 7 amperes. Even under continuous-power tests, this would allow me to reach maximum power levels of 392 watts per channel into 8-ohms and close to 200 watts per channel into 4 ohms. I therefore made the substitutions for my bench tests.
The circuit of the XL-280 represents something of a departure from previous Hafler amplifiers. The input jacks feed four J-FETs in a double-differential push-pull cascoded in put stage. The second stage current-mirrors the input. The usual input capacitor and output choke found on most amplifiers have been eliminated in this design to reduce phase shift at frequency extremes. The circuit is direct coupled throughout, and power-supply capacitors are by passed. Completely separate power supplies beyond the common power transformer reduce interchannel interaction, and total power-supply capacitance amounts to over 31,000 µF. The output stage employs MOS-FET devices.
Measurements
Switch position A is for listening to the output of the test amplifier, position B for bypassing to hear the driving amp's direct ("straight wire") output, and position D for the differential null test. All switch sections are ganged. See text.
Fig. 2--THD + N vs. power and frequency, 8-ohm load. Note that THD
scale is logarithmic.
Fig. 3--Same as Fig. 2 but with 4-ohm load.
Figure 2 shows how THD + N varied as a function of power output level and frequency when the amplifier was connected to 8-ohm loads. At the rated power level of 145 watts per channel, THD was only 0.0085% at mid-frequencies, 0.012% at 20 Hz, and 0.035% at 20 kHz. The 0.1% distortion level was not reached until the amp was delivering 174 watts at 1 kHz, 167 watts at 20 kHz, and 162 watts at 20 Hz. Dynamic headroom at 8 ohms was a bit more than 1.0 dB, while damping factor measured 111 at 50 Hz re: 8-ohm loads.
A complete set of measurements was also made with the amplifier connected to 4-ohm loads; THD + N versus frequency and power output level is shown in Fig. 3. At the rated output level of 200 watts per channel, THD measured 0.016% at 1 kHz, 0.057% at 20 kHz, and 0.017% at 20 Hz.
The amplifier delivered 225 watts per channel at 20 Hz before THD rose to 0.1%; for the same distortion level, it delivered 252 watts per channel at mid-frequencies and 240 watts per channel at 20 kHz.
Dynamic headroom when driving 4-ohm loads measured more than 2.0 dB, which means that even considerably higher short-term power levels could be delivered during music reproduction. The SMPTE-IM distortion was 0.03% for the rated power level of 145 watts into 8-ohm loads and 0.063% for the rated 200 watts per channel into 4-ohm loads. The CCIF-IM distortion was only 0.0015% at rated power when driving 8-ohm loads, rising very slightly to 0.005% when 4-ohm loads were connected.
The XL-280 exhibited one of the widest bandwidths of any amplifier I have ever measured. Its frequency response extended from 5 Hz to 600 kHz for the-1 dB points and from 2 Hz to 650 kHz for the-3 dB roll-off points. Input sensitivity for 1 watt output was 125 mV. It was not possible to measure signal-to-noise ratio in strict accordance with the EIA/IHF amplifier measurement Standard since this amplifier does not have an input-level control. As an alternative, measured S/N with respect to 1 watt output and read an A-weighted figure of 96 dB. Translated to rated output, this would correspond to an S/N of 117.6 dB.
Use and Listening Tests
The Hafler XL-280 is, first and foremost, a very clean-sounding amplifier. Power level delivered to my reference loudspeakers under actual listening conditions (4 ohms nominal impedance) was more than I was ever likely to need, even considering the fact that my reference speakers are fairly low in efficiency. I also connected the amplifier to three other sets of loudspeakers that were currently in my lab; in each case, the Hafler had no trouble driving them to more than adequate levels, and there were no problems caused by any of the speakers' impedance characteristics.
I had hoped that at least one of the speaker systems I used in my listening tests would require a drastically different setting of the internal "tweaking" controls. (These controls, by the way, are accessible from the top of the amplifier; one simply removes a couple of small plug buttons and uses a long screwdriver to make the adjustments.) The fact of the matter is that the setting which produced the most complete "null" when I connected Hafler's little XL-10 A/B/D switchbox to my reference speaker systems was pretty much the same as when the other speakers were connect ed. Furthermore, the musicality of the amplifier did not audibly change when I departed slightly from the ideal setting. Perhaps the speakers that I used did not present as great a problem to the amplifier as other models might have. It is entirely possible that, under some circumstances, an audible improvement might be detected when adjustments are made.
From my own point of view, I was more than pleased with the sound of the XL-280 even before I "tweaked" the internal potentiometers. The amplifier's construction and audible performance are fully consistent with what I have come to expect from David Hafler and his organization. I did not see the XL-280 in kit form. However, judging from other Hafler units offered as kits, I would guess that almost anyone who has used a soldering iron and the simplest of electronic hand tools would have no trouble assembling it and having it work perfectly the first time. Purchased fully assembled, the Hafler XL-280 is a real bargain. As a kit, it would be even more so!
-Leonard Feldman
(Adapted from Audio magazine, Nov. 1987)
Also see:
Hafler DH-330 FM Tuner (Nov. 1986)
Hafler XL600 Power Amplifier (Feb. 1989)
Heathkit AA-1800 Stereo Power Amp (Sept. 1982)
= = = =