Last month, an interesting Class AB audio amplifier was in the spotlight, a monoblock-configured pair of which I’m actually listening to (Widespread Panic’s Light Fuse Get Away, to be precise) as I type these follow-on words:

This time, it’s Class D’s turn in the spotlight. Here’s a summary link to EDN’s voluminous coverage of the technology over the past few decades, and a Wikipedia’s summary:

A class-D amplifier, or switching amplifier, is an electronic amplifier in which the amplifying devices (transistors, usually MOSFETs) operate as electronic switches, and not as linear gain devices as in other amplifiers. They operate by rapidly switching back and forth between the supply rails, using pulse-width modulation, pulse-density modulation, or related techniques to produce a pulse train output. A simple low-pass filter may be used to attenuate their high-frequency content to provide analog output current and voltage. Little energy is dissipated in the amplifying transistors because they are always either fully on or fully off, so efficiency can exceed 90%.

 Tripath, which branded the technology as “Class T”, was the first mainstream-volume supplier of Class D technology. My first exposure to Class D stretches back nearly 30 years, to 2007 when I first met with (and auditioned prototype gear from) D2Audio and its always entertaining (meant as a compliment), not to mention well-informed, Chief Technology Officer Skip Taylor at CES. I subsequently sic’d Skip on my then-colleague, EDN then-Analog Editor Joshua Israelsohn, who had a number of enjoyable mind-melding (maybe also melting?) meetings with Skip and his colleagues. Intersil bought D2Audio a year later, preceded by Texas Instruments’ 2000 acquisition of Burr-Brown and followed by Infineon’s 2018 purchase Merus Audio and Analog Devices’ 2021 Maxim buy (all as trend examples, not intended to be a comprehensive list)…and Class D technology was off to the races.

Class D vs Class AB

Here’s another take on Class D versus the Class AB (and A, for that matter) precursors, from Paul McGowan, the co-founder and CEO of high-end audio equipment supplier PS Audio, whose “Ask Paul” ongoing video series is both entertaining and educational, therefore highly recommended:

That this video comes from Paul (and PS Audio, located just “up the road” from me in Boulder, CO, for that matter) is highly revealing, I think. Audiophiles, the nexus of PS Audio’s customer base, are generally speaking both change-adverse and perversely picky when it comes to perceived quality. That they, who D2Audio was specifically targeting with its demos way back in 2007, were among the first to adopt Class D amplifier technology tells me a few things:

• With all due respect to Schiit co-founder Jason Stoddard, his diatribe about Class D’s tendency to “hiss like a demon cat, drilling slowly into your synapses and draining your soul” is dated and overstated, IMHO at least. As I noted last month, “he might have been right about Class D a few years ago, especially in the near-field configurations he’s specifically advocating for Rekkr, but no longer.”
• More generally, the technology has for a while now been good enough for audiophiles, so I’d wager it’s also good enough for the masses.
• And why was it appealing to audiophiles? Cost for them is a secondary-at-most concern, right? Well, their listening rooms tend to contain massive speakers, requiring formidable amounts of power to drive them. And, as Wikipedia notes, “The major advantage of a class-D amplifier is that it can be much more efficient than a linear amplifier, dissipating less power as heat in the active devices…also, given that large heat sinks are not required, class-D amplifiers are much lighter weight than class-A, -B, or -AB amplifiers.”

Schiit’s Class-AB-based Rekkr

Lighter…and much smaller, too. In the spirit of “a picture paints a thousand words,” here are some examples. First off, here again is the Class AB-based Rekkr (Internet Archive cache link), which is sound-spec’d as follows:

• Stereo, 8 Ohms: 2W RMS per channel
• Stereo, 4 Ohms: 3W RMS per channel
• Mono, 8 Ohms: 4W RMS

with the following form factor-related specs:

• Size: 5” x 3.5” x 1.25”
• Weight: 1 lbs

Schiit’s Class-AB-based Gjallarhorn

Next up is its Gjallarhorn “big brother”, also Class AB-based, also mentioned (but not shown) in last month’s piece, and sound-spec’d as follows:

• Stereo, 8 Ohms: 10W RMS per channel
• Stereo, 4 Ohms: 15W RMS per channel
• Mono, 8 ohms: 30W RMS

with the following form factor-related specs:

• Size: 9” x 6” x 2.5”
• Weight: 8 lbs

Rekkr vs Gjallarhorn size

Here’s a visual comparison of their respective form factors:

Foreground, then background, focus:

Ditto:

Keen-eyed readers may have already noticed that the comparison’s not entirely fair, since the Gjallarhorn integrates the AC/DC conversion circuitry that’s alternatively placed (at least partly) in the Rekkr’s external AC/AC “wall wart”. But as you can see, the Rekkr PSU is pretty tiny, so…:

Enter Class D competitors

Now for the Class D competitors (and, I’d generally argue, successors). Earlier this year, ahead of the looming tariffs, I acquired three different “Chi-Fi” manufacturer/model combinations…not counting the Class D circuitry inside the powered speakers I already owned…or my latest network audio streamer…or my latest sound bar (from Yamaha, replacing the Hisense unit I complained about in May)…all of which you’ll hear more about in other blog posts to come…

The first was a monoblock-only unit, the Fosi Audio V3 Mono:

Its (again, mono in this case) output specs, along with those of the other two devices I’ll be showcasing today, vary depending on the capabilities of the power supply connected to it:

• Rated Power Output : 48V/5A–240W@4Ω ; 32V/5A –100W@4Ω

And here are its form factor details:

• 6 x 4.2 x 1.6 inches (142mm x 105mm x 35mm)
• 1.06 lbs (0.48 kg)

The second is Douk Audio’s A5, a dual-channel (stereo) Class D amp:

Again, its output specs are power supply voltage-and-current, as well as speaker impedance, dependent. Douk Audio provides more granular detail on its website than Fosi Audio does even in the V3 Mono user manual, unfortunately. But the general trend is similar in practice, given that they’re both based on Texas Instruments’ TPA3255 chipset:

And its form factor details? Here you go:

• Dimensions (W*D*H): 95*92*50 mm/3.74*3.62*1.97 in
• Net weight: 506 g/1.12 lb

Last, but definitely not least, is a more recent Douk Audio device upgrade, the A5 Pro, adding both a Bluetooth receiver and a separate headphone output amplifier:

Same TI TPA3255-based audio power amplifier subsystem as with the base A5, so same output specs as shown earlier. The form factor is tweaked a bit (but only a bit, and likely mostly-to-completely to just make room for more knobs-and-such on the front panel), however:

• Dimensions (W*D*H): 130*112*33 mm/5.12*4.41*1.30 in
• Net weight: 525 g/1.16 lb

Schiit Rekkr (AB) vs Fosi Audio V3 (D)

That’s the last of the stock shots, at least for a while; now for some more “real life” ones. First off, here’s how the Schiit Rekkr stacks up (literally) against the Fosi Audio V3 Mono, which is capable of up to (speaker impedance- and power supply-dependent) more than 50x the mono output power at comparable distortion:

Schiit Gjallarhorn (AB) vs Fosi Audio V3 (D)

How about the Schiit Gjallarhorn versus the Fosi Audio V3 Mono? Glad you asked. Again, as a reminder, the latter has ~8X the mono output power in this case, with its beefiest power supply option and when driving the same impedance and at similar (inaudible) distortion levels. Not to mention being half the price (or even less, depending on where it’s sourced and how it’s kitted):

Again, a two-photo foreground-then-background focus shift:

Once again, the detail-oriented among you will point out that the Gjallarhorn chassis also encompasses AC/DC conversion circuitry, external to (and not shown in) the Fosi Audio case. You’re right, although there was a method to my madness. I didn’t want to show three sets of vs-V3 Mono photos, one set with each of the three PSUs I have in my possession: 32V/5A, 48V/5A and 48V/10A. Instead, here are the standalone undersides of the power supplies, capable of being used with any of the three Class D amplifiers I’m covering today. 32V/5A first:

Now 48V/5A:

And finally, the “Big Kahuna” 48V/10A version:

And here they are stacked on top of each other, with the 32V/5A one on top and (obviously) the 48V/10A one on the bottom:

The DC Power Filter

Independent reviews suggest that the incremental power output of the Fosi Audio V3 Mono tails off beyond the 48V/5A point…the Douk Audio units’ additive output performance increases more linearly at 48V/10A, but that’s to be expected as there are two audio power amplifiers—one for each channel—inside. But there’s another reason to run the Fosi Audio V3 Mono—two of them, actually—with a 48V/10A source. In such a configuration, the company also sells what it calls a “DC Power Filter”, which (in conjunction with an appropriate cable option) splits the 10A input current evenly among both of its outputs, enabling a single PSU to concurrently fuel two amplifiers:

I own two DC Power Filters. One came bundled with a two-amplifier set I bought off eBay. The other was a standalone purchase from Fosi’s online store, used with two other V3 Mono amps I got individually (for reasons I’ll explain further in a teardown to come!):

Schiit Gjallarhorn (AB) vs Douk Audio A5 (D)

What about those two Douk Audio units? Here’s the A5, alongside the 48V/5A power supply it came bundled with, on top of the Schiit Gjallarhorn:

Schiit Gjallarhorn (AB) vs Douk Audio A5 Pro (D)

Now for its A5 Pro “big brother”:

Schiit Gjallarhorn (AB) vs Douk Audio A5 + A5 Pro (D)

Because I couldn’t resist, the following shots prove that (after dispensing with the PSUs), I could fit both Douk Audio devices on top of a Gjallarhorn:

All the Class Ds stacked

And in closing, here are all three of today’s Class D devices stacked on top of each other, showcasing their form factor similarities:

Notable observations

Wrapping up, there are a couple of other points I wanted to note. First off, all three of the Class D amps support (believe it or not) user-accessible op-amp swapping:

analogous to the “tube rolling” of times past (and present, for some folks, and potentially others, too…there is, after all, a Vali 2++ now sitting at the top of the Schiit stack on my desk):

Finally, what’s with the “PFFB” promotion prominent on both manufacturers’ websites?

It stands for Post-Filter Feedback, and understanding what it is and does first requires a step (or few) back. Although, as the Wikipedia Class D definition I shared at the beginning of this piece noted, “A simple low-pass filter may be used to attenuate their high-frequency content to provide analog output current and voltage,” in practice the output filtering circuitry tends to be notably more complex than this; to render inaudible the otherwise distorting aforementioned “hiss like a demon cat”, for example, to suppress phase shift artifacts, etc. To reiterate on this circuit’s robustness importance, I’ll turn you over to Paul McGowan again for more on the topic:

PFFB, implemented in TI’s TPA3255 (and others), is the latest evolution in this output filtering scheme. Quoting from Google’s AI Overview summary of the search topic:

PFFB, or Post-Filter Feedback, is a secondary feedback loop in Class-D amplifiers that takes a portion of the signal after the LC output filter and feeds it back to the input to improve audio quality. This technique reduces distortion and improves the linearity of the power stage and output filter components, particularly the inductor, which is a primary source of distortion. PFFB also increases load independence, meaning the amplifier’s performance is less affected by the specific loudspeaker connected.

Here’s a visible example of PFFB’s benefits. First off, an output level-vs-frequency plot from Audio Science Review’s evaluation of the Fosi Audio V3 Mono:

Requoting highly recommended content expert (and long-time personal collaborator) Amir Majidimehr, “There is essentially no impact up to 20 kHz between the 4 and 8 ohm, indicating very low output impedance, albeit with a bit of peaking. Compare that to non-PFFB amps such as Fosi Audio V3 stereo amp:”

The TL;DR (at the end of another long writeup…sorry!) summary of Amir’s findings (and PFFB’s benefits): it suppresses an amplifier’s perceived loudness from otherwise varying with output frequency, not that this arguably was perceptible much if at all previously, candidly (specifically because the variability tended to occur at high frequencies, hard for all but the “golden ears” among us to discern, anyway). But since it now “comes along for the ride” with modern Class D amplifier designs anyway, at little if any incremental cost and with no sonic downside…

With that, encroaching on the 2,300-word threshold, I’m going to sign off for today. More on this topic, including the earlier-promised Class D amplifier teardown, to come soon. Until then, “sound off” with your thoughts in the comments!

—Brian Dipert is the Editor-in-Chief of the Edge AI and Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.

 Related Content

• Audio amplifiers: How much power (and at what tradeoffs) is really required?
• EDN Class D amplifier coverage
• Class D audio power amplifiers: Adding punch to your sound design
• The truth about Class D amplifiers

The post Class D: Audio amplifier ascendancy appeared first on EDN.

Source link