[SoundStage!]Home Audio
Equipment Review

September 2003

Orpheus Laboratories Three S Stereo Amplifier

by Doug Schneider

Click to view measurements of this product

Review Summary
Sound "Maintained a solid grip that resulted in deep, tight, and sometimes thunderous bass, along with rich, controlled, and textured midrange frequencies"; "squeaky clean to the max, transparent and detailed as anything, and most certainly never fatiguing."
Features Solid-state amp that is billed as delivering constant power via the Anagram Power Loop, "a feedback-type circuit that monitors the output of the amplifier against the input needs of the speaker."
Use Because of the Power Loop circuit, "the amplifier’s output impedance will vary," so auditioning with your speakers is mandatory; also, "when the amp plays, it gets hot," so you'll want to allow some space for ventilation, which is easy because of the amp's small footprint.
Value "Its $4600 price isn’t cheap, but this amplifier is ultra-refined-sounding, and that’s what you are paying for."

The people at Switzerland-based Orpheus Laboratories appear poised to make quite a splash in the high-end audio world. If they succeed, they’ll do it by not only making good-sounding electronics. They’ll also have bucked some trends and punched holes the size of Buicks into some conventional audiophile wisdom.

To say the least, Orpheus products are unique. The Orpheus One digital-to-analog converter features the Anagram Technologies ATF (Adaptive Time Filtering) 24/192 upsampling engine. Anagram, you see, owns Orpheus Laboratories and makes "high-performance digital signal processing solutions" -- not full components, but hardware and software like ATF that can be incorporated into other companies’ products. Indeed, that leg of the business has had considerable success. Today you can see the ATF upsampling module implemented in a number of high-end CD players and digital-to-analog converters, the most famous likely being Audio Aero’s Capitole 24/192 CD player, which uses a customized version of ATF that Audio Aero calls STARS. Of course, Anagram products get used in Orpheus products, too.

Then there’s the Two multichannel preamplifier -- a high-resolution preamplifier that's user-configurable between two and six channels. In November, 2002, Jeff Fritz praised it to the high heavens for how neutral it sounded, as did I, and it gained a well-deserved Reviewers’ Choice nod, along with the SoundStage! Network Edge of the Art award for 2002.

But what could Orpheus and Anagram do with a power amplifier -- a solid-state power amplifier at that? The Orpheus Laboratories Model Three M 150W monoblocks ($8400 USD per pair) and the Three S stereo amplifier ($4600), the subject of this review, include the Anagram Technologies Power Loop -- a unique circuit placed in the output stage of the amplifier that purports to turn what would be a normal solid-state amplifier into one that delivers constant power into any type of speaker load. The Three S is rated at 40Wpc into 16, 8, and 4 ohms. This is unusual for a solid-state amplifier, which will usually deliver less power into 16 ohms than into 8, and more power into 4 ohms.


The styling of the Three S (and Three M) is amazingly simple -- too simple, perhaps, for those who enjoy giant bulky devices with incredibly thick faceplates and massive heatsinks jutting from the sides. But I like the way this amp looks -- then again, I’m not one of those people who has to have something that looks like a car engine resting on my floor to bolster my audiophile status.

The faceplate is 19" wide, but only 1-3/4" high, and from front to back the amp is 15". There are no external heatsinks. The unobtrusive size and appearance, no doubt, is likely the reason that some audiophiles think that the Orpheus amps are digital switching designs. They’re not.

When you pick the Three S up you’ll notice how surprisingly heavy it is. It’s weighted mainly to the front, where the large but shallow transformers sit. And when the amp plays, it gets hot -- hot enough that I quizzed Orpheus Labs about this, wondering if the amp needed lots of "breathing room." It doesn't. In fact, Orpheus gave me the go-ahead to stack the other Orpheus components I had in-house on top of the amp -- exactly how they do it at their factory! I was surprised, but I did it, and I never encountered a problem. However, reviewer stress-testing isn’t necessarily something the user at home should do. If you buy a Three S, I recommend giving it some breathing room, or at the very least put the amplifier on top, or near the very top, of your stack of components. Why build up excess heat if you really don’t have to?

The backside of the amp is pretty conventional: single-ended and balanced inputs, a dual set of speaker binding posts, an IEC receptacle for a detachable power cord, and the on/off switch.

Now things get a little trickier

The Three S’s power rating is 40Wpc -- not all that great for a $4600 amp and not likely to impress those who buy watts by the dollar. But as I said, the interesting thing here is that Orpheus claims 40Wpc regardless of the load. Understanding this, though, is critical to successfully implementing the amp in your system. There are, after all, implications that consumers need to realize.

Constant voltage, constant power, and loudspeaker frequency response

Most speakers are designed for an amplifier that delivers constant voltage, not constant power. So what happens when a speaker is driven by an amplifier that delivers constant power? Frequency-response deviations that correlate with the speaker’s impedance curve. If the impedance is ruler-flat (i.e., the same regardless of frequency), there are no deviations. But virtually all loudspeaker impedance curves are far from flat and can vary greatly from their low to high points. Furthermore, amplifier damping factor is dependent on low output impedance. The lower the amplifier's output impedance, the better it can behave like a true constant-voltage source when driving a varying loudspeaker impedance. An "ideal" voltage source has no output impedance.

Take a typical speaker driven by an amplifier that provides constant voltage under varying load, and say that when measured from 20Hz to 20kHz it is perfectly "flat" (i.e., equal loudness at all frequency points). The manufacturer may specify the impedance to be a "nominal" 8 ohms, but in reality the impedance will likely vary widely depending on the frequency. For our example, we’ll say there is a 4-ohm low point at 1kHz, a 16-ohm high point at 100Hz, and most everything else is at about 8 ohms. As I mentioned, fluctuations can often be quite a bit greater.

First, let’s consider the amplifier delivering constant voltage. If we put in a constant voltage of 2.83V across the frequency band (typical for sensitivity-type measurements), the speaker will put out a "flat" frequency response -- every frequency point between 20Hz and 20kHz is equal in terms of how loud it is being played. But while the speaker’s acoustical output in SPLs is remaining the same at all frequencies, along with the voltage from the amplifier, the amplifier’s output is varying in terms of how many watts it is producing at a given frequency.

Power is measured in watts and the method to calculate is simple: voltage multiplied by current equals power output in watts; current equals voltage divided by resistance. So, at the point where the speaker is 8 ohms, that equates to 1 watt of power from the amplifier with a 2.83-volt input (2.83 volts x (2.83 volts / 8 ohms) = 1 watt). At the 4-ohm point (1kHz, in our example), the amplifier delivering constant voltage is now producing 2 watts (2.83 volts x (2.83 volts / 4 ohms) = 2 watts). At the 16-ohm point (100Hz, in our example), the amplifier is delivering half a watt (2.83 volts x (2.83 / 16 ohms) = 1/2 watt). So a normal amplifier is doubling its power output when the impedance goes from 8 ohms to 4 ohms, and halving its power output when the impedance goes from 8 ohms to 16 ohms. That’s precisely why you see some amplifiers rated at, say, 100 watts into 8 ohms, 200 watts into 4 ohms, and 50 watts into 16 ohms. All the while the SPL from the speaker -- the acoustical output -- remains constant.

On the other hand, the amplifier delivering constant power is putting out the same power regardless of impedance. At the 8-ohm point it is delivering 1 watt (the same as the amplifier delivering constant voltage), but at the 4-ohm point it is still delivering 1 watt (where the amplifier with constant voltage was putting out 2 watts), and at 16 ohms it is again delivering 1 watt (where the amplifier with constant voltage was delivering half a watt). The power output in watts from the amplifier delivering constant power compared to the amplifier delivering constant voltage is the same at 8 ohms, one-half at 4 ohms, and double at 16 ohms. But in this case, while the power is remaining constant, the SPL from the speaker is now fluctuating. Why?

To get an SPL increase of 3dB you need a doubling of amplifier power. Likewise, to reduce the SPL by 3dB, you need to halve the amplifier power. So at the point where the amplifier is delivering one-half the power (in our example, at 4 ohms, 1kHz), the speaker will be "down" (quieter) by about 3dB. At the point where the amplifier is delivering double the power (in our example, at 16 ohms, 100Hz), the speaker will be "up" (louder) by about 3dB. Therefore, the speaker that was "flat" when driven by an amplifier with constant voltage is no longer flat when you apply constant power.

...Doug Schneider

Delivering constant power is done using the Power Loop, a feedback-type circuit that monitors the output of the amplifier against the input needs of the speaker, and faster than Winona Ryder can stuff clothing into her purse, voltage and current get adjusted. As a result, the amplifier’s output impedance will vary, too. With a conventional solid-state amplifier, voltage remains constant.

This explanation is deliberately oversimplified (I recommend reading Orpheus Labs' literature or website to find out more), but it’s enough to raise a simple question: Why do things this way? Orpheus Labs says that adjusting the output voltage and current results in better control of the speaker’s drivers, and this, they say, results in better transient response. And transients, they point out, are what music is all about. So why doesn’t everyone do this? Because, as I said, Orpheus is bucking some trends -- most notably the way we’re led to believe an amplifier should work.

All other solid-state amplifiers I know of are designed to deliver constant voltage, with output impedance as close to zero as possible. And, not surprisingly, almost all speakers are designed to be driven by an amplifier that delivers constant voltage. Florian Cossy, head of research and development at Orpheus and Anagram, readily admits that the solid-state Three S and Three M behave more like tube-based amps that have a high output impedance. However, the similarity ends there because Florian doesn’t buy into the notion of actually using tubes, calling them "distortion devices."

But because most speakers are designed for amplifiers with constant voltage and a very low output impedance, they won’t react the same when fed constant power. In fact, their frequency-response curves will vary in a way that correlates with the speaker’s impedance curve -- just like those high-output-impedance tube amps. So even those who contend that all solid-state amplifiers sound the same might be forced to back off from that stance when they encounter the Three S and Three M because, if what Orpheus Labs says will happen does happen, it’s quite easy to demonstrate through measurements how the acoustical output of a speaker designed for an amplifier that delivers constant voltage will be altered using an amp like the Three S and Three M.

"So isn’t that bad?" That was the next question I posed to Florian, and he believes that the frequency-response deviations are minor compared to what a real room can do to the output of a speaker (room gain, for example, can bump a speaker’s output at various frequencies astronomically), and, he believes, the frequency-response compromise is offset by the improvement in transient response.

The only thing I could do at that point was listen to the Three S.

Before I get started…

Discerning this amplifier’s sound is a challenge because it is dependent on the amp's interaction with the speaker. As I noted, constant power will cause response variations in the speaker’s output that won’t be present in a conventionally designed solid-state power amplifier. Comparing two amplifiers is difficult enough. But comparing two amplifiers when there is a frequency-response variation with one that differs from speaker to speaker is close to impossible. Therefore, does it make sense to compare the Three S to a tube amp? Not really -- but for even more reasons. Even the all-amplifiers-sound-the-same group normally holds their arguments within the limits of solid-state designs, readily conceding that tubes have a flavor, or "sound," if you will. I love tube amps. I won’t dismiss them like some will, but I’ll be the first to admit tube-based amps vary sonically to a far greater degree than solid-state amps. The Three S sounds different from a tube amp, that’s for sure, but for a whole lot more reasons than just one.

But don’t get the idea that the Orpheus Three S’s sound varies so greatly that same music played through one speaker sounds completely different played through the next. The folks at Orpheus Labs are right: Frequency-response variations, when you factor in the room, are not quite as cut’n’dried as some may think. Still, it must be noted that I detected some variation with speakers that were audible enough that I couldn’t discount them. In the end, then, I have to confine my review to my experience with the amplifier, along with description of some basic sonic characteristics of the Three S that help to define its sound and were shared among the speakers I used.


I used the Three S with the Orpheus Two multichannel preamplifier and One DAC (with my Theta Data Basic as a transport). A variety of review speakers included Verity Audio Tamino, Song Audio Type II Silk DM, Ruark CL10, and Ascend Acoustics CBM-170. Analog cables were either from Nirvana (S-L) or Nordost (Valkyrja); the digital cable was i2Digital’s X-60.

The Tamino was quite telling about the amplifier’s output capability. The Power Loop module is one thing to consider, but the 40Wpc output was yet another. I feared the Three S would die trying to drive the Tamino, which has been a tough load for a few amplifiers I’ve had here. (I suspect quite low impedance, although I don’t have measurements for this speaker here to back that up.) As a result, the best partners to date have been beefy solid-state units: the Perreaux 200ip and Simaudio i-3 integrateds rated at 200Wpc and 100Wpc, respectively. Those amplifiers grabbed the Taminos like vices, turning them from the mushy mess that some amplifiers produced into quite an open, transparent, and lively sounding speaker.

And the Three S? It was able to do the same thing in a way that surprised me. There was no strain and certainly no sense of pain coming from it -- nor was there a hint that this is a low-powered amplifier. From weighty and gutsy orchestral pieces like Carmina Burana [Telarc 80056] to hard-edged and gutsy rock-type affairs like Steve Earle’s Jerusalem [Artemis 751147], the Three S maintained a solid grip that resulted in deep, tight, and sometimes thunderous bass, along with rich, controlled, and textured midrange frequencies. There was also loads of details amidst a top end that was airy, clean, and wonderfully extended. Voices and instruments hovered in space, carved from the mix, and the bass was so authoritative that I was impressed by how muscular this relatively small amp could sound. However, I’m apprehensive to say something like the Three S has stronger bass than some other amplifier. There is, after all, the Power Loop in the equation. Many speakers have rising impedances in the bass region that could account for a large part of what would be heard. Therefore, on that final note, I will conclude inconclusively, that the bass from the Taminos was impressive.

And while controlling the speaker is one thing, achieving high SPLs is another. In my smallish room, I did try to tax the Three S to its limits -- but surprisingly, I never succeeded. My room and I gave out before the Three S did. This isn’t to say that the Three S is some sort of magical powerhouse, mind you. Orpheus Labs does offer the 150W M version for a good reason -- some people will need more power. How loud you need a speaker to play will depend on your room size and speaker’s sensitivity. My room isn’t all that big, and the speakers I was using were of average sensitivity. What this does prove, though, is that 40 watts can get you quite far provided the amplifier is beefy enough to wrestle with any speaker load.

The Three S exhibited another strength that was consistent from speaker to speaker: astonishing transparency and cleanliness. Like its companion the Two preamplifier, there is starkness to the Three S's presentation that’s far more significant than, say, simply wiping a window clean, or lifting a "veil." It’s like taking a rock and breaking through the window to see a crystal-clear view of what’s outside. The Three S sounds so clean it’s as though it's not there.

Glenn Gould’s A State of Wonder: The Complete Goldberg Variations (1955 & 1981) [Sony Classical 87703] is a surprisingly pristine-sounding CD set -- great older recordings now polished up with state-of-the-art remastering. Here the Three S shows its abilities and lets you know why it costs what it does -- and why you shouldn’t necessarily buy your watts by the dollar. Gould’s piano hovered in space, nothing obscuring it, amidst a ton of detail. From the lows through the highs, the Three S sounds utterly clean -- crystalline is perhaps even better. There isn’t a hint of grain, hash, or grunge. What you get is an exceedingly pure, see-through rendering of the recording -- and the Three S does this without sounding fatiguing, as some solid-state amps can.

In fact, sometimes when I tire of the zippiness inherent in some solid-state designs, I turn to something like the all-tube Zanden Model 600 for its rather romantic take. The Zanden isn’t as airy as the better solid-state amps, sacrificing a bit up top for a less-fatiguing sound, but it can sound so much easier on the ears. Now I turn to the Three S, too. The Three S can’t match the 600’s richness, particularly through the midrange, but its top end is fully extended without going over the edge. And that’s the key to the Orpheus Three S. It’s not how it looks or how much power it delivers; it’s how refined it sounds -- or doesn’t sound, depending on how you look at it.

But that’s not to say that the Three S is without fault -- depending on your speakers it can be, well, a little odd. And although I can’t necessarily confirm it definitively, I suspect it has to do with the unique topology it employs, which includes the Power Loop circuit. That’s why you have to be careful with what speakers you match it with.

Whereas the Taminos and the Three S mated quite successfully -- as did the CBM-170 and Song Audio Type II Silk DM -- the results with Ruark’s CL10 were mixed. Bass and highs from the CL10s were fine: clean, tight, and articulate down low; clean, clear and infinitely extended up top. These things remained consistent from speaker to speaker. But still, the pairing wasn’t quite right in the midrange. I remember playing Norah Jones’ Come Away with Me [Blue Note 32088] and finding her voice and portions of her piano leaping out at me too much. The CL10s already have a front-row perspective, but this was putting me right up to the stage.

We’ve measured the CL10, and these measurements will perhaps help to explain things. Even driven by a normal solid-state amplifier, the CL10s have a peak, both on and off axis, starting at around 600Hz and extending a little past 1kHz. In my review of the speakers, I noted the speakers’ slightly forward balance. Surprisingly, though, it’s not really objectionable -- the CL10s sounded so clean that when they presented things like voices in a little more forward manner, the sound was still quite good and lively. However, with the Three S it was even more forward than that.

Strange? Not really. Looking at the impedance measurement for the speaker shows an additional rise starting in that same area, but extending upwards to about 2kHz. That means that harnessed to the Three S, you get the normal "bump" in the midrange that’s inherent in the CL10, and then, likely, an additional bump in that same region and beyond. As a result, the CL10 goes from a little forward, to even more forward. Too much of a good thing? With the Three S, yes.


As I mentioned, the unique operating nature of the Three S makes an apples-to-apples comparison with any solid-state amp almost impossible. However, there are some things that can be said about the Orpheus amp in relation to other amps, perhaps to put into perspective why someone might pay so much for a relatively low-powered amplifier.

The Perreaux 200iP I mentioned is priced about $2000 less than the Three S, and the i-3 almost $3000 less -- and they are integrated amps to boot! In terms of power output and features, the Three S can’t compete. The Three S’s strength, though, is its refinement. The 200iP and the i-3 sound great and are worth every cent of their asking prices -- and then some. But when you move to the Three S, you hear more detail and clarity in the midrange, along with a top end that is so squeaky clean and extended, without even the remotest hint of hash, that the differences hit you square in the face.

I played the soundtrack to The Mission [Virgin 90567-2], a long-time favorite of mine that features large-scale choral works. The speaker end of my room was flush with voices, each wonderfully delineated. The stage was precisely set, and the minutest details were easy to discern. The Three S gives a high-rez presentation, but without any sizzle or bite.


As Jeff Fritz has done, I recommended the Orpheus Labs Two multichannel preamplifier with plenty of enthusiasm. And I’m just as enthusiastic about the Three S in a number of ways. The Three S is only modestly powered, but it’s gutsy in a way that makes you forget its power rating. More importantly, though, it’s one of the most pristine-sounding amplifiers I’ve heard -- tube or solid state. It's squeaky clean to the max, transparent and detailed as anything, and most certainly never fatiguing. Its $4600 price isn’t cheap, but this amplifier is ultra-refined-sounding, and that’s what you are paying for.

However, I have to temper my enthusiasm some. The Three S, more than any other solid-state amplifier I’ve used, has to be matched carefully to the speaker with which you want to use it. Your experience will likely vary somewhat from speaker to speaker, as it did for me. The Three S is certainly unique, and the guys at Orpheus Laboratories are bucking some trends with their designs, but there’s a bit of a caveat to doing that. Still, don’t let reservations dissuade you from checking this amp out -- although you must try before you buy.

...Doug Schneider

Orpheus Laboratories Three S Stereo Amplifier
$4600 USD.
Warranty: Five years parts and labor.

Orpheus Laboratories SÓrl
Chemin du Saux 17
1131 Tolochenaz
Phone: +41 21 803 33 81
Fax: +41 21 803 33 82

E-mail: info@orpheuslab.com
Website: www.orpheuslab.com

North American distributor:
Globe Audio Marketing
1873 Rymal Rd. E.
Stoney Creek, ON
Canada L8J 2R6
Phone: (800) 330-3804 or (905) 522-5040
Fax: (905) 560-4850

E-mail: info@globeaudiomkt.com
Website: www.globeaudiomkt.com

[SoundStage!]All Contents
Copyright ę 2003 SoundStage!
All Rights Reserved