Solid State Power Amp Reviews

It takes a while for audio-related technologies to mature. Tubed amplifiers were invented by Lee de Forest in the nineteen-teens, but while there are still some adherents of early high-distortion triode designs, the age of mainstream high-fidelity amplification dawned with higher-power/lower-distortion amplifiers developed by Williamson and McIntosh followed by the Ultralinear take on the Williamson concept. That was 30+ years down the technology-evolution timeline after de Forest.

And when it comes to solid state amplifiers—the usual kind—does anyone prefer the state of the (germanium) art circa early 1960s to modern silicon class-AB designs? I doubt it.

Now, decades into its own development, class-D amplification seems to have sea legs, even in the audiophile world. Interesting fact: The class-D amplifier was invented and named in the 1950s by the man who had already invented pulse-code modulation for signal transmission, Englishman Alec Reeves. The first commercial class-D amp was a 2.5W kit from Sinclair Radionics of England, introduced in 1965 and followed the next year with a 20W second-generation model. Sinclair was eventually better known for pocket calculators.

The modern class-D amplifier, built into an integrated circuit, debuted in 1996 with the Tripath "Class-T" chips. Since then, there has been a steady march forward. Today, most "civilian" amplifiers in things like TV soundbars, "smart speakers," Bluetooth-connected portable speakers, non-fancy automobile sound systems, flat-screen TVs, and the like are commodity-priced class-D "amplifier bricks." They are small, efficient, cool-running, and cheap. Class-D amplifiers started to appear in audiophile-grade components about 10 years ago, although there were earlier, outlier examples. In the pro-audio world, they have been around a bit longer, often but not always inside powered monitor speakers. Class-D amplification is ubiquitous in the modern sound-reinforcement world, that is, for live-concert sound.

I have observed this evolution, and in fact have used a cheapo Tripath-based amplifier for 10+ years in a little garage sound system fed by my phone or iPod when I'm out there doing work. If I pried open the cases of the various TVs, Bluetooth speakers, and soundbars around the house, I'm sure I'd find class-D amp chips aplenty. But until very recently, I was not interested in a class-D amplifier to drive my listening-room speakers. Modern class-AB amps do that trick very well.

One of my audio-hobby buddies, Mike, is a fearless DIY builder of amplifiers and speakers. For several years, he has been acquiring the latest, greatest class-D "bricks"—self-contained amplifier circuit boards, sometimes including an input buffer stage—mating them to appropriate switch-mode power supplies, and installing them in professional-looking metal cases. Last fall, he was eager to show off his latest masterpiece, made with the new Purifi modules designed by Bruno Putzeys. We arranged to shoot out several of Mike's class-D amps plus a professional class-D audio amp I own that uses modules that were the newest thing about 10 years ago. Our friend Farrukh, who joins our semi-regular lunch-and-listening sessions, said he had a new class-D amp that utilized a commodity-priced chip from Texas Instruments that he wanted us to hear. He said this amp was a hot topic on interweb audio boards. Being a guy who often swaps components in and out of his system, he couldn't resist buying one to try out. He told me he was impressed.

On the day of our listening comparisons, I heard clearly what I hadn't liked about those earlier class-D amps. My older amp, and Mike's older builds based on prior generations of class-D modules, varied sonically in some respects, but they shared a "glaze" over the top end, which I find annoying. The best way I can describe it is as jagged treble, too harsh to be pleasing with most music. It's not what I'd call tizzy or splashy, but it's noticeable and not to my liking.

Two amps in the shootout did not exhibit this "glaze": Mike's amp made with the Purifi modules and the new amp that Farrukh brought: the Living Sounds Audio (LSA) Discovery Warp 1 ($1499). I knew right away that I wanted to live with this amplifier for a while.

Warp 1 backgrounder
The designer of the LSA Warp 1, 54-year-old, Virginia-based Viet Nguyen, works days for NASA, project-managing the next generation of weather satellites, which the space agency will build and put into space for NOAA, the government agency in charge of weather prediction and research. An electrical engineer by education and avocation, Nguyen undertook DIY'ing a pair of Bluetooth-connected speakers back in 2010—thus began an audio odyssey that led him to design horn speakers made of dual-layer foam-core board, then headphone amplifiers that he built and sold out of his home, and finally power amplifiers to drive his ever-more-ambitious speaker projects. In short, he DIY'd himself a second full-time gig.

A couple of years ago, on an online forum, he posted a design for a power amp using the commodity-priced Texas Instruments TPA3255 class-D amp chip (footnote 1) in a low-impedance, current-drive circuit backed up by an overspecified, high-current switching power supply with a solid state, electronically balanced buffer stage ahead of it. This caught the attention of Mark Schifter, who runs product development and manufacturing for Walter Liederman's mini-empire, Underwood HiFi (see Sidebar 1). The two connected, and Nguyen agreed to work with Underwood to commercialize his design.

Nguyen picks up the story: "Up to that point, I had never worked on things that were manufactured, just small runs of headphone amps I built by hand." Schifter connected Nguyen to an extensive network of engineers, parts suppliers, and contract manufacturers in China. "I learned how to be a project manager and make a product," Nguyen said. He has since designed several speakers and amplifiers for Liederman's LSA and Emerald Physics brands.

Warp 1 design
According to Nguyen, the strength of his design revolves around driving the TI amp in its comfort zone and using an overbuilt power supply to provide as much current as needed as quickly as any speaker will demand it. "These amplifiers are designed to work best with a balanced signal," he said—so the input driver board is a key part of the equation. He knew a power supply designer, and they worked together to come up with a 600W switch-mode supply that uses such "overbuilt" touches as "massive" pure-copper inductors from Coilcraft and special high-current wire (spec'd for up to 100 amps), designed to carry the battery power in high-performance drones. The goal, he explained, is "a low-impedance path for current from power supply to speakers." Low impedance means that the amp is fast and firm in its control of the speaker engines. "It should provide really slamming bass," Nguyen said. He noted that the damping factor is "around 500" although it's spec'd as ">300."

Nguyen says the input-buffer stage uses Texas Instruments OPA1637 "balanced drivers capable of low-distortion 52V [peak to peak] output swing with balanced inputs." He has designed the Warp 1's gain staging to never come close to the buffer's full output. That means tons of low-noise headroom.

Those over-specified coils are in the low-pass filter, which removes the class-D switching noise, which is centered at 400kHz. Removal of switching noise is the big challenge with class-D amps. The filter has to do the job thoroughly but not ring in such a way as to clash with parts of the audible frequency spectrum and create odd harmonics or other "glaze" sound components. To my ears, Nguyen's filter succeeds. Hopefully JA's measurements won't make me the fool!

Warp 1 details
Rated at 150Wpc into 8 ohms and 250Wpc into 4 ohms, the Warp 1 Underwood HiFi sent to me for review arrived double-boxed and securely packed, with a power cable and thin, laser-printed, stapled-together user manual. The amp doesn't weigh much and doesn't generate much heat. The TI amplifier modules are mounted on the bottom of the main circuit board, with a silicone heat-transfer pad between them and the metal bottom panel. In my time using the amp, the top and sides never went beyond room temperature and the bottom felt only slightly warm.

On the front panel is a pushbutton power switch, the outline of which glows blue when the amp is on and red when it's in standby mode. To the right are three LEDs indicating Power supply fault, amplifier fault, and "Over Temp"; the amp will shut down if this stays on too long. On the back panel is a reset button (required if the amp shuts down due to a fault), input connectors, speaker banana/screw-terminal binding posts, an IEC socket, and a master power switch, which cuts all power to all the innards. The input jacks are gold-plated RCA for single-ended and Neutrik combo connectors for balanced, so either XLR or pro-style ¼" TRS connectors may be used. The user manual notes that the amp is designed for balanced inputs by default, and balanced cables must be disconnected before single-ended sources are connected. The amp should be turned off before disconnecting and reconnecting input cables.

The sturdy metal case is free of sharp edges or other telltale signs of crude machining. It is held together with Allen screws. Inside are three circuit boards: the power supply, the input driver, and the power amplifier. All are securely attached to the chassis, and connecting wires are securely bundled with zip ties. Low-level signal wires are twisted pairs. Heavier-gauge, high-current wire is used to connect the power supply to the amplifier board and the amplifier output to the speaker terminals.

Here's something the user manual doesn't tell you: The driver board has DIP switches (one bank of switches for each channel), allowing the user to adjust its gain to best match the components feeding it (footnote 2). Nguyen says the class-D amps provide 22dB of gain. The driver board is factory-set for 6dB of gain, for 28dB total, "which is an average gain for most sources." Other switch combinations produce 0, 14, and 20dB gain, but Nguyen says "I would not ever use the 20dB setting, as that would easily clip the amp." Some preamps—for instance, my Benchmark LA4—produce high enough output to justify using the Warp 1 with 0dB gain from the driver board, which results in "lower noise performance." The 14dB setting may be appropriate "if all you have is a cellphone that can't even put out 1V RMS as a source," Nguyen said, but that setting usually provides too much gain and isn't generally recommended. I tried the amp with both 6dB and 0dB gain from the driver board. While I can't say I noticed a difference in its sound, I ended up preferring 0dB because I like to use my Benchmark LA4 with less attenuation (footnote 3).

---

Footnote 1: See ti.com/product/TPA3255#description.

Footnote 2: There are two DIP switches, one for each channel, located on the top of the driver board; both sets of switches need to be the same. Since it's not in the manual, here are the settings for various gain levels: all switches OFF = 0dB; 1 and 4 ON (others OFF) = 6dB; 2 and 5 ON (others OFF) = 14dB; 3 and 6 ON (others OFF) = 20dB.

Footnote 3: Benchmark would approve: In an Application Note on their AHB2 amplifier, the intended partner for Tom's LA4 preamp, Benchmark writes, "Most power amplifiers have far too much gain, and this degrades noise performance of the overall system." The highest gain available from the AHB2 is 23dB. See benchmarkmedia.com/blogs/application_ notes/14680625-the-ahb2-a-radical-approach-to-audio-power-amplification.—Jim Austin