"I'm a recording engineer, so I value accuracy," said a panelist in a discussion—titled "How to Read Between the Lines of Audio Advertising"—at last October's Rocky Mountain Audio Fest. I, too, was on the panel, which was moderated by Brent Butterworth, a writer for the SoundStage! Network of online audio magazines.
"Accuracy is overrated," I interjected from the other end of the dais. "Accurate to what? To your sonic tastes? To what you hear on your preferred loudspeakers? Other than one's personal preferences, I'm not sure the term accuracy has much meaning."
I then qualified those statements somewhat, saying that gross tonal colorations—like the ones that afflicted many of the demo systems at RMAF—were unacceptable.
In audio, this debate has always and will always amount to a sonic free-for-all. That's part of both the fun and the frustration of the hobby. I've heard absolutely thrilling sound from a tiny vintage tube amplifier putting out just a few watts to drive EJ Jordan Designs "full-range" speaker arrays, supplemented with carefully dialed-in subwoofers. The action, so to speak, appeared between the speakers, a few feet in front of the single seat, on a miniature soundstage on which appeared the most living, breathing Ella Fitzgerald I'd ever heard. She was, within the constraints of the system, brought to life.
Yes, it was a tiny visitation. Yes, dynamics and visceral bass were MIA. But what was there—able to be enjoyed by only the person seated in that sweetest of sweet spots and by no one else in the room—was, again, thrilling. It became even more so when the system's owner swapped out his "everyday" tubes for a pair of ultra-rare vintage tubes saved for special occasions.
Did that system produce "accuracy"? Had you heard what I heard, you'd have to say "yes"—but accurate to a specific aspect of sound reproduction, to the exclusion of other such factors. I have a few friends who prefer that kind of sound, and who look at me with poorly disguised derision, or the kind of concern usually reserved for someone with a life-threatening illness, when I remind them what my version of accuracy requires: big speakers capable of producing realistic SPLs, wide dynamic swings, big soundstages, and gut-socking bass—all to the exclusion, to one degree or another, of the aspects of sonic realism that they prefer.
Some audiophiles prefer looking into the sound; others want it projected outward toward them. There's room for all at the audio table, but I don't expect to convert a Quad ESL 57 enthusiast into a Wilson or Magico or YG Acoustics fanatic. Nor should the EJ Jordan guy expect me to head his way any time soon. However, understanding each other's preferences in sound can go a long way toward creating a more tolerant world—in audio, at least.
Description
The Boulder Amplifiers 2150 is a large, extremely heavy amplifier that weighs 220 lbs each and costs $99,000/pair. It's an update of the 2050 mono amplifier, introduced by Boulder more than 17 years ago—a long run for any audio product. Boulder's promotional statements almost make it seem as if the update was done grudgingly, more to "meet the demands of the market" than because the original's sound quality could be improved. The changes, more evolutionary than revolutionary, include the use of Boulder's proprietary 99H2 gain stage—a fully discrete, surface-mount, modular op-amp that, in 2013, replaced their 993 module. The H stands for High voltage.
The differentially balanced 2150 is said to be class-A biased to its full claimed continuous power output—1000W into 8, 4, or 2 ohms—by means of a circuit that continually adjusts the bias current based on voltage output, current draw, and load. Boulder claims that this operation is performed more quickly than the speed of the audio signal itself, which means that the circuit can maintain class-A operation when confronted with a taxing musical transient, after which it gently lowers the bias until it senses another peak (footnote 1). The 2150 can thus provide the positives of full class-A operation that's free of crossover distortion, without the negatives of massive power consumption, excessive generation of heat, or, Boulder claims, the audible "steps" produced by other kinds of active bias-management systems.
Each 2150 has two power supplies. The massive primary supply, for its input and output circuits, includes two toroidal transformers, one for each phase of the output, that are custom-wound in the US to Boulder's specs. A smaller, independent, regulated supply powers the microprocessor control sections. Mains connection is via an enormous 32-amp IEC jack—the kind used in shipyards and factories. On the other end of this cord is a standard AC plug: no need to call the electrician. The 32A IEC jack isn't just for show: the 2150 can draw nearly 30 amps of peak current, for high power outputs into low impedances. In my room, with my sensitive speakers, however, the Boulder was never asked to deliver anything close to that.
The 2150's input stage is a fully balanced, differential, high-impedance, servo-controlled, direct-coupled circuit using bipolar devices. Its 26dB voltage gain stage uses the encapsulated 99H2 op-amps in two stages, each 99H2 providing input buffering and voltage gain with a high slew rate, wide bandwidth, high current output, low distortion, and low output impedance.
Boulder proudly uses feedback where others dare not, claiming that appropriate levels of carefully implemented feedback can achieve ideal operating parameters and constant group delay across the entire audioband, for maximizing linear phase response. In fact, Boulder has a take-no-prisoners approach to feedback, claiming that those who don't understand its proper implementation—and so don't use it—have a philosophy that "still exists in a number of hobbyist workshops today." Ouch!
The 2150's output section has a specified nonreactive output impedance of 0 ohm (!), meaning it can drive any speaker to realistic playback levels, regardless of the speaker's sensitivity or impedance. The output section features 80 (count 'em—80!) bipolar devices, 40 each for the waveform's positive and negative halves. Rather than being screwed in, these output devices are clamped, by a CNC-machined bar, to a nonresonant heatsink cut from an 80-lb billet of solid aluminum. Boulder claims that this increases reliability and reduces mechanical resonances—two things that the company takes very seriously.
Boulder says it avoids standalone heatsink fins, which can resonate and ring. (Speaking of which, if your amp has separate fins, try listening to it, then wrapping the fins in adhesive tape and listening to it again.) Of course, the 2150's protection circuit is as heroically designed and built as everything else about it.
Boulder boldly states that "at no point is the 2150 'voiced' or tuned for a specific sound or type of loudspeaker." At this point, the low-watt, single-ended-triode guys (those few who've read this far) have their hair on fire—all of their belief systems have been challenged, crushed, or at least dented, despite the stiff resistance they've put up getting this far into the review.
Inside and out, the 2150's build quality meets and in most ways surpasses that of every other amplifier I've reviewed. Boulder is one of the few electronics manufacturers that does its own CNC machining in-house. I visited their previous facility, and it was mighty impressive. The new one is said to be even more so. Other than the toroidal transformers and individual components on the boards, Boulder designs, manufactures, and assembles everything in-house, including the circuit boards. All metal parts are fabricated in-house from solid stock—no sheet metal is used anywhere.
Hernia-Inducing Setup, Ball-Breaking Sound
Although Boulder recommends that four people move and install each 2150, it was just two diminutive guys—me and Rich Maez, Boulder's director of sales and marketing. We managed.
I ran the amps balanced. Boulder's wingnut-type speaker terminals are the best I've used, especially since the two pairs per amp are located on an "open field" rear panel—only spade lugs need apply. I began with Boulder's stock power cords. Then, to hear if power cords could affect the sound of an amplifier overbuilt to such demanding specs, I asked AudioQuest if they'd send over two of their 6' Hurricane cords ($1495 each), with Boulder supplying the amplifier-end plugs to match the 2150's 32A IEC jacks.
How does a measurement-derived amplifier sound? Essentially, especially in terms of tonality, it doesn't. That may sound like a reviewer cop-out, but the Boulder 2150 didn't really "sound," tonally. It was about as tonally neutral a piece of electronics as I've heard here, neither warm nor cool—unlike, say, the SAE HP2, which I thought was on the cool side.
Overall, my first listen immediately produced a reaction similar to when I first heard Boulder's 2110 preamplifier, which I reviewed two years ago—but to an even greater degree. The 2150s gripped the woofers of my Wilson Audio Specialties Alexandria XLFs (and, later, the Wilson Alexxes; review in the works) with greater clamping force than did my reference darTZeel NHB-458 monoblocks. I can't say the bass extension went further down or that bass definition increased (the NHB-458s perform really well there). But the punch and speed, the starting and stopping power, were definitely accelerated.
As with that 2110 preamp, everything, from top to bottom, tightened up through the 2150s—another good thing, because now the system spoke from one sonic perspective: with greater speed, transparency, and, to a lesser degree, resolution of detail than with my reference darTZeels.
Footnote 1: Unlike an amplifier with a traditional class-A output stage, the Boulder 2150 runs relatively cool.—John Atkinson
Some audiophiles prefer looking into the sound; others want it projected outward toward them. There's room for all at the audio table, but I don't expect to convert a Quad ESL 57 enthusiast into a Wilson or Magico or YG Acoustics fanatic. Nor should the EJ Jordan guy expect me to head his way any time soon. However, understanding each other's preferences in sound can go a long way toward creating a more tolerant world—in audio, at least.
DescriptionThe Boulder Amplifiers 2150 is a large, extremely heavy amplifier that weighs 220 lbs each and costs $99,000/pair. It's an update of the 2050 mono amplifier, introduced by Boulder more than 17 years ago—a long run for any audio product. Boulder's promotional statements almost make it seem as if the update was done grudgingly, more to "meet the demands of the market" than because the original's sound quality could be improved. The changes, more evolutionary than revolutionary, include the use of Boulder's proprietary 99H2 gain stage—a fully discrete, surface-mount, modular op-amp that, in 2013, replaced their 993 module. The H stands for High voltage.
The differentially balanced 2150 is said to be class-A biased to its full claimed continuous power output—1000W into 8, 4, or 2 ohms—by means of a circuit that continually adjusts the bias current based on voltage output, current draw, and load. Boulder claims that this operation is performed more quickly than the speed of the audio signal itself, which means that the circuit can maintain class-A operation when confronted with a taxing musical transient, after which it gently lowers the bias until it senses another peak (footnote 1). The 2150 can thus provide the positives of full class-A operation that's free of crossover distortion, without the negatives of massive power consumption, excessive generation of heat, or, Boulder claims, the audible "steps" produced by other kinds of active bias-management systems.
Each 2150 has two power supplies. The massive primary supply, for its input and output circuits, includes two toroidal transformers, one for each phase of the output, that are custom-wound in the US to Boulder's specs. A smaller, independent, regulated supply powers the microprocessor control sections. Mains connection is via an enormous 32-amp IEC jack—the kind used in shipyards and factories. On the other end of this cord is a standard AC plug: no need to call the electrician. The 32A IEC jack isn't just for show: the 2150 can draw nearly 30 amps of peak current, for high power outputs into low impedances. In my room, with my sensitive speakers, however, the Boulder was never asked to deliver anything close to that.
The 2150's input stage is a fully balanced, differential, high-impedance, servo-controlled, direct-coupled circuit using bipolar devices. Its 26dB voltage gain stage uses the encapsulated 99H2 op-amps in two stages, each 99H2 providing input buffering and voltage gain with a high slew rate, wide bandwidth, high current output, low distortion, and low output impedance.
Boulder boldly states that "at no point is the 2150 'voiced' or tuned for a specific sound or type of loudspeaker." At this point, the low-watt, single-ended-triode guys (those few who've read this far) have their hair on fire—all of their belief systems have been challenged, crushed, or at least dented, despite the stiff resistance they've put up getting this far into the review.
Inside and out, the 2150's build quality meets and in most ways surpasses that of every other amplifier I've reviewed. Boulder is one of the few electronics manufacturers that does its own CNC machining in-house. I visited their previous facility, and it was mighty impressive. The new one is said to be even more so. Other than the toroidal transformers and individual components on the boards, Boulder designs, manufactures, and assembles everything in-house, including the circuit boards. All metal parts are fabricated in-house from solid stock—no sheet metal is used anywhere.
Hernia-Inducing Setup, Ball-Breaking SoundAlthough Boulder recommends that four people move and install each 2150, it was just two diminutive guys—me and Rich Maez, Boulder's director of sales and marketing. We managed.
How does a measurement-derived amplifier sound? Essentially, especially in terms of tonality, it doesn't. That may sound like a reviewer cop-out, but the Boulder 2150 didn't really "sound," tonally. It was about as tonally neutral a piece of electronics as I've heard here, neither warm nor cool—unlike, say, the SAE HP2, which I thought was on the cool side.
Overall, my first listen immediately produced a reaction similar to when I first heard Boulder's 2110 preamplifier, which I reviewed two years ago—but to an even greater degree. The 2150s gripped the woofers of my Wilson Audio Specialties Alexandria XLFs (and, later, the Wilson Alexxes; review in the works) with greater clamping force than did my reference darTZeel NHB-458 monoblocks. I can't say the bass extension went further down or that bass definition increased (the NHB-458s perform really well there). But the punch and speed, the starting and stopping power, were definitely accelerated.
Footnote 1: Unlike an amplifier with a traditional class-A output stage, the Boulder 2150 runs relatively cool.—John Atkinson
