Bricasti Design M15 power amplifier

On the second morning of a recent audio show, I walked into the exhibit room of Bricasti Design.

"How about some Mahler?" asked cofounder and designer Brian Zolner.

"Oh God. Not at 10am!"

Listening to Mahler had become de rigueur for my visits to the exhibits of this Massachusetts-based company. From the very first time that Zolner had played part of Iván Fischer and the Budapest Festival Orchestra's recording of Mahler's Symphony 2 (SACD/CD, Channel Classics 23506) through a system that combined Bricasti's M28 monoblock power amplifiers, M1 DAC, and M12 Source Controller with Tidal Audio's Piano Diacera speakers, I'd been impressed by the layered darkness of the bass and midrange, and the system's ability to flesh out Mahler's despair. But beginning the day with Mahlerian doom? No one should have to suffer and be resurrected before noon.

After listening to a modest dose of alternate fare—it wasn't as mindless as "the flowers that bloom in spring, tra-la, nor as hackneyed as "Hotel California"—I learned that Bricasti's M15 fully balanced stereo power amplifier ($18,000), introduced in 2016 and offering 125Wpc into 8 ohms and 250Wpc into 4 ohms, had not been reviewed by Stereophile. After running through the components in my reference system with Zolner and Damon Gramont, Bricasti's head of sales, and noting that my speakers would soon change from Wilson Audio's Alexias to Alexia 2s, we felt that the M15 would be a good match for my system. John Atkinson signaled the okay, and a review sample was on its way to me.

History Lesson
Bricasti Design was founded in 2004 by two former employees of Lexicon, a division of the Harman Specialty Group: sales executive Brian Zolner and DSP software engineer Casey Dowdell. After Bricasti released the Model 7 reverberation unit in 2007, and it became what Zolner described to me as "a staple of the professional recording industry," the partners decided to enter the consumer market, in 2011, with the M1 DAC. In 2014 came the M28 monoblock power amplifier, and in 2016 the M12 Source Controller and matching M15 stereo power amplifier. These later products were engineered at AeVee Laboratories, in New Haven, Connecticut, by a small group of former employees of Madrigal Labs and manufactured at Bricasti's facilities in Shirley, Massachusetts.

"We like to make things that are very musical and natural," Zolner said. "I don't like things that are hyped-up in the top end. Everyone is mixing things with so much hype these days. I can't listen to that pop stuff that they're compressing and then pumping up, so we're careful to not go over the top end as well. We want to bring out what's in the recording while aiming for a relaxed sound that you can listen to for hours."

Description and Setup
Zolner, who calls himself "the mechanical guy," designed the M15's case, which comprises sections of milled and CNC-machined aluminum, assembled and given an anodized finish by Bricasti. Inside are a lot of custom mechanical components, and on the rear panel are proprietary, gold-plated speaker-cable binding posts, also designed by Zolner, that offer what he calls "exceptionally wide contact area," and open wide enough to accept two pairs of thick spade lugs. Inside, rather than using solder, Bricasti creates a superior molecular bond by high-pressure crimping the binding posts to double runs of fine-stranded (8 gauge) copper wire.


"There's nothing really esoteric or too crazy inside," Zolner continued. "There are 24 bipolar output devices per channel, bolted directly to the heatsink on the side of unit. The power supply has a very large capacitor bank—there are large, US-made electrolytics in the main cap bank—as well as lots of smaller, German-made, high-frequency WIMA film caps throughout the unit."

After consulting Bricasti's website and the M15's manual, and chatting further with Zolner, I learned that the M15 is, in his words, a "big, all-analog, high-current, fully balanced class-A/B power amplifier. Its bandwidth is to the megahertz. There's no filtering on the output, which is the case with analog amplifiers, so that the transients are not differently rendered, and it has very, very, extremely low output impedance and very low distortion. The thing is flat to 800kHz."

The M15's rear panel is in three horizontal sections. In the top section are widely spaced single-ended (RCA) and balanced (XLR) inputs. While shorting plugs (not provided) are urged for the balanced inputs if single-ended interconnects are used, I used only balanced cables between DAC and amp. In the middle section, below the inputs, are those two pairs of widely spaced speaker-cable binding posts.

A healthy distance below, in the bottom section, are, from left to right, trigger in/out, an RS-422 port currently used only for diagnostics, a detented trim control that can lower input sensitivity in four steps of 6dB each, to better match the output of a preamp or DAC, and a three-prong IEC connector. I never felt the need to change the amp's input sensitivity. Nor, thanks to the rear panel's intelligent design, did I ever need to worry about interconnects or speaker cables transmitting unwanted noise by coming in contact with the power cord.

On the M15's front panel are two vertically aligned buttons, Power and Standby, separated by a red LED power on/status indicator. Pressing both buttons generates a click, and triggers the LED to flash slowly as the amp enters Idle stage, in which it consumes only 2W. Only after the M15 has sensed that the voltage is within an acceptable range can it be fully powered up.

A quick press of Standby produces another click and faster flashing of the LED, to indicate that you've entered the pre-heating stage, in which 60W are consumed. A second quick press of Standby triggers a third click, the LED turns solid red, and the M15 unmutes its inputs to enable music playback.

Eager to hear how the solid-state M15 would sound with and without a tube preamp, I installed Lamm Industries' two-box L2.1 preamplifier ($22,790). With the bottom shelf of my five-shelf Grand Prix Monaco rack already filled with power products from Nordost and Isotek, I placed the dCS Paganini CD transport on the top shelf and the preamp's control module atop it. The remaining shelves held the Lamm preamp's power supply, the dCS Rossini DAC, and, sitting on a Grand Prix Formula platform, the Scarlatti clock.

I'd planned to place the Bricasti on one of my two Grand Prix Monaco amp stands, but when I looked at the M15's feet, I sensed in them a purposefulness meant to discourage the use of ancillary isolation products. Finding no information about them online or in the manual, I called Bricasti. It turns out that Zolner designed these feet, which house OEM Stillpoints decoupling modules and are intended to be placed directly on a hard surface. "I didn't want people to try balancing them on Stillpoints and having them fall over," Zolner said.

Goodbye amp stand, hello floor.

Bricasti Design, Ltd.
2 Shaker Road, Building J100
Shirley, MA 01464
(978) 425-5199

Bogolu Haranath's picture

"Cake by the Ocean" would be more appropriate for listening at 10am ........ or, I got the "power" ...........

hifiluver's picture

Could the anomalies in sound be caused by an impedance mismatch between power and tube pre amp?

Ortofan's picture

... the kid who got the new toy he desperately wanted for his birthday, only to open the package and find out that it's broken.
Despite Mr. Gramont's best efforts to provide an optimal set up, who would have guessed that the reviewer - an admitted "less-than-technical genius" - couldn't resist trying to "improve" it? Ah well, live by the subjective review, die by the subjective review.

Moving on, what does this $18K amp do that can't be done by a (much) less expensive Bryston 14B-cubed (let alone a 7B-cubed) or a McIntosh MC452 or a Parasound JC5 or a Rotel RB-1590?

spacehound's picture

And they don't even 'engineer' it themselves.
With a salesman and a software guy in charge I don't suppose they know how, as the result seems to demonstrate.
Bricasti has always struck me as 'Big hat, no cows'.

georgehifi's picture

Here again (not as bad as some) is another amp manufacturer that purposely understates the 8ohm wattage, to make the 4ohm wattage look like it's doubling, then the 2ohm doubling again, just to make the amp look like it's a real great current pusher for the evilest of speaker loads, when in fact it's not. When is this sort of deceit going to end?

Manufacturer spec:
Bricasti specifies the M15 as outputting
125Wpc into 8 ohms
250Wpc into 4 ohms
500Wpc into 2ohms

Stereophile tested results: Far from doubling fo each halving of impedance.!!!!!!!!!!!!
290Wpc into 8 ohms
550Wpc into 4 ohms
362Wpc into 2ohms, before switching off.

Cheers George

hifiluver's picture

in fairness to the poor manufacturer, it does push out quite a lot of watts but it looks like it might be expedient for them to change some of their promotional literature?

CG's picture

What makes you say deceit?

As you go lower in load impedance, eventually an amplifier's output power will be limited by the available current for a given output voltage. That's the doubling thing you refer to. It's based on Ohm's Law.

Higher load impedances require less current for a given output voltage. Again, Ohm's Law.

So, you end up with two forms of output power limiting.

One is the available current, which shows up at low impedances. This is usually due to either a control circuit or fuses limiting the output current to protect the devices or because of limited drive current for the output devices.

The other is the maximum output voltage, limited by the power supply rails and fixed voltage drops from various circuit elements, like the drivers. This shows up at higher load impedances where you normally aren't current limited.

For example, if you have bipolar power supplies of 56 Volts DC, that limits the maximum output voltage to 40 Volts RMS (.707 times the peak available voltage) in the case of a more-or-less conventional complementary push-pull transistor output stage that isn't of the "bridged-mode" variety. (That number gleefully presumes that there are no fixed drops due to base emitter voltages and that sort of thing.)

40 Vrms gives 200 watts into an 8 Ohm load, 400 into 4 Ohms, 800 into 2 Ohms, 1600 into 1 Ohm, 3200 into 0.5 Ohms, and so on. That's assuming there is adequate current available to get 40 Vrms into those load impedances. For example, 40 Vrms into 1 Ohm requires 40 Amperes of output current. For an 8 Ohm load, that's only 5 Amperes. Again, Ohm's Law.

So, at low impedances you're usually current limited for practical reasons and at higher impedances you're voltage limited.

But, all that plain ignores the realities of transistor behavior.

Very often, transistors display better linearity when run at higher voltages. There's a bunch of reasons for this, too involved to go into here - inspecting a few transistor data sheets would give you an idea as would doing some Spice simulations or device measurements. So, it's not at all uncommon to run devices above the absolute minimum required to attain the required rail voltages in order to get best linearity. A simple example of that is in preamps. Since the most output you ever need to clip almost every power amplifier in existence is about 2 Vrms, why not just run all preamps from 3 Volt rails? Performance!

Personally, I fail to see how an amplifier that operates to and above its specified output power is deceitful in any way. If anything at all, the manufacturer is being conservative and delivering more than they claim.

(No, I have no connection with Bricasti. I don't think I've even ever been in the same building as a Bricasti product, but I can't swear to the latter. I've certainly never seen or heard any of their products.)

prof's picture

I think you've made georgehifi's point.

There's a regime where the amplifier's power output is voltage-limited. That's the regime where halving the load impedance doubles the power output.

At low impedances, you leave that regime and enter a current-limited regime.

When driving a set of speakers, you'd like the amplifier to stay in the voltage-limited regime.

In reporting the power output as inversely proportional to the impedance, the manufacturer is implicitly claiming that their amp is in the voltage-limited regime from 8Ω all the way down to 2Ω. The actual measured power output belies that claim — with obvious implications for anyone contemplating using this amp to drive a "difficult" set of speakers.

CG's picture

We'll just have to disagree, since I think you have that backward.

The voltage output is just the voltage output, at least until something limits the current and the output no longer approximates a true voltage source. The voltage output should be a (linear!) function of the applied input voltage signal.

I fail to see how the manufacturer is implying anything other than what output power the amplifier will maximally produce at some distortion level into a couple specified resistive load impedances. That it will do more than the specification at some impedances is not a failing of the amplifier. Just as producing lower distortion or noise than specified is a failure.

From all these comments, I'd have to conclude that Bricasti would have been more honest and better received if they had reduced the amplifier voltage rails to just barely above the minimum required to get 150 watts into 8 Ohms. They would have saved on heat sinking and perhaps on some other components due to reduced voltage ratings. Silly them! They must've had some reason to run the rails higher, wouldn't you say? I don't know for sure, but I tend to doubt that it was to inflame people commenting on a review.

I'd also point out that even 100 plus watts into difficult loads is a really, really loud sound level in most listening rooms inside people's houses. Even with our own speakers, which measure on the lower side in terms of efficiency, 100 watts output is brutally uncomfortable and hardly pleasant to listen to. (I've made the measurements) However, YMMV as they say.

Ortofan's picture

... many of the speaker test reports, you'll note that at certain frequencies the impedance goes off the top of the chart above 20 ohms - let's suppose it may go as high as 24 ohms. In that case, your modified amp (with the power supply voltage reduced to the minimum needed to generate 150W into an 8 ohm load) will only be able output 50W.

To another comment, a continuous output of 100W may indeed result in a higher sound level than most can tolerate. However, if you listen to music with a sufficiently wide dynamic range then, even if the average sound level only requires just a few watts, well over 100W may be needed to avoid peak clipping. Also, you would need an SPL meter with a very fast response time and a peak hold function to accurately measure and display those peak levels.

CG's picture

Second thing first.

What I did was to play a couple pieces of music as you describe. There's even a listing somewhere on the internet that tabulates measured dynamic range for a zillion recordings. I chose the best examples I had based on that listing.

While listening, I turned the volume up to beyond what we'd ever listen to. The music was too loud to be enjoyable. I then left the volume control at that point.

Then, I played a track with a tone that was generated at -12 dBFS and measured the voltage right at the speaker terminals at that volume level. BTW, that tone was REALLY loud, I must say.

Anyway, the voltage at the speaker terminal was about 6 Vrms. (I can't recall, exactly). That would mean that a 0 dBfs tone - as loud as it could possibly go - would be 24 Vrms. That's equivalent to 72 Watts into 8 Ohms, and scaled accordingly for other load impedances.

This test was done at a level considerably above even what an "enthusiastic" listening session might be, so I think there's plenty of useful headroom there. Probably far more than needed. I'd bet that a 15 Vrms maximum output amplifier (roughly 30 watts into 8 Ohms) amplifier would be more than enough in our situation here. Emphasis on "our". (That's presuming a lot, though, since lower power amplifiers often have much lower capacitance power supply filters and maybe fewer output devices in parallel - both also affect actual performance.)

This pretty much correlated with another observation. On most normal recordings, our volume control is run at somewhere between a preamp gain of -20 and -10 dB. (Max gain is zero.) If you do the numbers, that suggests that the maximum ever coming out of the amplifier was about 15 Vrms. (The higher gain settings are usually used for the few recordings where the peaks don't hit 0 dBfs)

Note that with phono based systems, you might have different requirements. Pops and clicks might cause extreme output levels from a phono preamp well beyond the highest of high musical levels you'll ever get. This can be mitigated through phono preamp design to a large degree, but it's still something that might be an issue in some cases. With a digitally based system, full scale is full scale and that's it.

For the first, almost all modern speakers are designed to be driven with amplifiers that approximate pure Voltage sources. So, while your point is accurate that the power drops with greater load impedance for a fixed Voltage, that ends up being irrelevant. The output Voltage will be whatever the input signal is times the gain of the amplifier as long as the amplifier is not being current limited. Unless there is a design fault with the loudspeaker, that won't affect the volume level at any one frequency.

So, this becomes a spec thing. If you use certain headphones, the load impedance can be 300 Ohms. What then?

What this really amounts to, for me, is that manufacturers design amplifiers that are close approximation to Voltage sources. (Most manufacturers, anyway) That's to match up with the loudspeakers. (Interesting question - which came first??) The way that's traditionally been used to spec amplifiers is to describe the output power supplied into a couple mostly arbitrary load resistances. That's what people more or less understand.

A better description might be, "Our UltraGreat (tm) amplifier acts as a Voltage source with an output impedance of no more than 0.2 Ohms from 20 Hz to 20 KHz. See the attached plot. The specified maximum output Voltage is 47 Volts rms, up to a maximum specified output current of 8.7 Amperes rms. At higher current demands, the output Voltage drops accordingly. Every amplifier will meet or exceed those performance parameters at an AC mains supply of 118 VAC or 236 VAC, depending on the input Voltage setting for the market.."

Personally, I think that would be better than using terms like damping factor and the rest. If individual examples beat those numbers, great!

This is a very funny discussion. It's like people getting worked up because the car they bought was only specified to attain 38 mpg on the highway and they're all getting between 48 and 56 mpg. Clear fraud on the manufacturers part.

Ortofan's picture

... your amplifier, with a 56V supply, can produce into a 24 ohm load?

Suppose your program material has a dynamic range of 22dB, instead of only 12dB as in your example. Then you would need at least 720W to avoid clipping with a 0dBFS signal. Even if the average level was 8dB lower, you'd still need over 100W for a 0dBFS input.

CG's picture

No, no, no, no, no....

I played music that many people measured having a high dynamic range.

I turned up the volume to way beyond the normal operating level. Just to be sure.

The -12 dBfs signal could have been anything whatsoever, as long as its level relative to 0 dBfs was known. This was just a simple value that wouldn't be unbearably loud when I crouched down next to the loudspeaker to make the Voltage measurement.

Knowing that -12 dBfs is 1/4 the voltage of full scale - by definition - that says the output Voltage level at 0 dBfs would be 4 times what I measured at -12 dBfs. Make sense?

Ortofan's picture

... "way beyond" (in dB)?
What value of power output corresponds to your average "normal operating level"?

You haven't answered the question about the maximum power that your amplifier, with a 56V supply, can produce into a 24 ohm load.

CG's picture

Jeez, you got me.

I just reread your post and realized that I didn't answer your question. My apologies.

The answer is 66 watts.

The point is that it doesn't matter. My loudspeakers, like many (most?), are designed and built to be driven by Voltage sources. The acoustic output is a function of the Voltage applied at any frequency. Sure, the applied electrical power at any frequency is the square of the Voltage divided by the load impedance. All that means is that less electrical power is required at some frequencies to maintain the acoustic output level than at others. The same is true for very low impedances, but in the opposite direction. Just as it's not a given that loudspeaker impedance is uniform across the audio band, it's not a given that loudspeaker efficiency is uniform across the audio band with respect to applied electrical power.

Beyond that, I truly don't get what you're getting at here.

The original suggestion, since repeated, was that Bricasti and all these other manufacturers are a pack of deceptive liars because their amplifiers will produce higher output powers at 8 Ohm load impedances than they specify and that this behavior doesn't extend down to a very low load impedance. Hence, the amplifiers do not do what they claim or people infer that they claim. I didn't and don't support that position (which I now realize was a really stupid thing for me to do) because the amplifiers do what they say. (The jury is out on 2 Ohm load performance in the M15 due to the amplifier shutting down when driven to maximum power into that low an impedance. If that's a design problem, then the manufacturer's specifications are inaccurate.)

So, I suggested a reason why many amplifiers are able to produce greater than their specified output into an 8 Ohm load. My contention is that it's not due to some conspiracy, but because the amplifier linearity and/or noise performance may be improved when the Voltage rails are higher than absolutely required to meet the specified power output into 8 Ohms.

How this got twisted around into the need or desire for 720 Watts output into 24 Ohm loads is beyond me. So, I'm afraid I can't answer your question on the subject to your satisfaction.

As for what's "far above a 'normal operating level'", you'll just have to take my word for the idea that I turned up the volume control until a number of tracks, all measured to have very high dynamic range (as high as I could find), were painfully loud. Like so loud that the dogs ran out of the room, my wife came in to see what had happened, and I couldn't stand it much more. Then, I applied a track with a known level tone into the system and measured the Voltage at the loudspeakers while keeping the volume control unchanged. From that, I was able to calculate what power level is needed to blast our faces off on high dynamic range music material. Normally, we don't use the system in full face blast mode. In fact, this was the first time.

The full scale output of our DAC is 2 Vrms. The maximum gain of the preamp is unity (0 dB.) The power amplifier gain is 17 (just under 25 dB) as measured. That gives a maximum of just under 34 Vrms at the speaker terminals. (2 X 1 X 17). That's it - there ain't no more. During the test, the volume control ended up way below maximum. How much below maximum? Based on previous measurements, about 6 dB, but I admit that I did not measure this during this particular test. 17 Vrms into 8 Ohms is 36 Watts. And that was too loud for me.

I was curious what power we really needed and I answered the question to my own satisfaction. YMMV, as they say.

Jason P Jackson's picture

200+ watts into 8ohms, 400+ into 4 with opamp-like distortion, I bet the Bricasti would sound fantastic with most speakers.

hifiluver's picture

It would and you would do just as well with a textbook topology amp costing a tenth of this one...

Jason P Jackson's picture

True. Point taken. I'll be sticking to diy..

ok's picture

..Lamm L2.1 is a solid state preamp.

PTman's picture

At AXPONA in 2016 and this year, I felt the Bricasti room provided one of the best sounds at show. The pairing of Bricasti amp with Bricasti M1 (or the other DAC/preamp products) into Tidal speakers was exceptional.

Prior to the 2016 AXPONA, the Bricasti DAC/Amp combo (into Tidal speakers) at my dealer is what drove my decision to purchase the M1 Limited Gold (which I use with BMC monos driving YG Acoustics Carmel), that I've never regretted. It bested the dealer's prior MSB/D'Agostino combo (driving Raidho) that had been a sonic benchmark for me.

Additionally, I've always received stellar support from Brian Zolner, answering any questions promptly and professionally.

Not meant to be a defensive post; just trying to illustrate that different component combos can have different results, and (in my opinion) the quality of Bricasti is worth the price.

ok's picture

Years ago I happened to confront my own precious player/integrated combo along with some pricey speakers unknown at a local audio show and it sounded as awful as any system can be; even most shocking to me was the fact that the man in charge didn’t seem to take the slightest notice of the monstrosity orchestrated. I left the room totally confused.

Jason Victor Serinus's picture

It is a hybrid preamplifier, with tube voltage regulation.

For more on its power supply, see:

Ortofan's picture

... a hybrid (pre-)amplifier. More typically, the term 'hybrid' would imply the use of both tubes and transistors directly in the signal path.
Of course, some will argue that the power supply is in the signal path.

georgehifi's picture

Don't get me wrong the Bricasti probably sounds great.
There has been a few that can do this doubling act, which usually means they can stay flat in frequency response no matter what the load even down to 1ohm, the old ML2 monoblocks come to mind at close to 25w 50w 100w and yes 200w at 1ohm, but it's becoming a bit of a hoax these days to understate the 8ohm wattage to make the amp look as though it can do the doubling act.

Manufacturer spec:
Bricasti specifies the M15 as outputting
125Wpc into 8 ohms
250Wpc into 4 ohms
500Wpc into 2ohms

Stereophile tested results: Far from doubling fo each halving of impedance.!!!!!!!!!!!!
290Wpc into 8 ohms
550Wpc into 4 ohms
362Wpc into 2ohms, before switching off.

Cheers George

CG's picture

I guess you're not swayed by the engineering argument presented above.

OK by me.

Gerardus's picture

The symphonies of Mahler are not all about doom. In fact they are quite romantic for their time. Compare them to contemporary composers like Bruckner. I have no problem with Mahler at 10.00am.

georgehifi's picture

If you can't see what they're trying to "imply" what their amps can do, then that's OK by me too.
As too many now are using this "understating the 8ohm wattage" now to present a falsehood of the amp being able to double with each halving of impedance load. This has only crept in, in the last few years but it's growing, as it looks good on paper/ad specs

Cheers George

CG's picture

Well, I just looked at the published review of a Levinson Number 29 amplifier from April 1991. Guess what? It doesn't double its output power every time you halve the load resistance either. And since the measured output power into 8 Ohms was above what their specified value is (66 Watts versus 50 claimed), they must be (or were) promoting falsehoods, too.

So, this isn't new by any means.

My point is that that by the letter of what both manufacturers claimed, they met their specification. (With the exception of the M15 shutting down at very high power into 2 Ohms.) Both of these amplifiers do double output power every time the load impedance is halved as long as you operate them below their specified maximum (again, the M15 2 Ohm load thing applies here).

Here's the thing - neither manufacturer implied anything beyond what they said. I could be wrong, but in a quick search of the Bricasti literature, I saw nothing that said the M15 would produce even 126 Watts into 8 Ohms. Please correct me if I'm wrong about that.

So, if anybody is inferring anything, it's the readers of the review.

Bogolu Haranath's picture

Also, look into the recent measurements (and review) of Mark Levinson No.536 mono blocks ......... Somewhat similar type of measurement behavior like you describe ........... Unfortunately, measurements for Mark Levinson No.534 are not provided (to date) ............. Also, look into measurements of Progression mono blocks ................

georgehifi's picture

Here as quoted by Stereophile for the manufacturer:
Bricasti specifies the M15 as outputting 125Wpc into 8 ohms (21dBW), 250Wpc into 4 ohms (21dBW), and 500Wpc into 2 ohms (21dBW).

And again in the Manufacturer Specs page:
Sidebar 1: Specifications
Description: Solid-state stereo power amplifier. Inputs: 1 unbalanced (RCA), 1 balanced (XLR). Outputs: 1 pair speaker binding posts. Power output: 125Wpc into 8 ohms, 250Wpc into 4 ohms, 500Wpc into 2 ohms (all 21dBW).

Cheers George

13DoW's picture

Great discussion on power output specs. Putting on my blue helmet, everyone is right! CG is absolutely correct in his analysis but George has a valid point because "doubling power into halving loads" is something that audiophiles think important, whether or not it is. So, a company massaging their numbers to imply it does fall somewhere in the continuum between "accentuating the positive" and "deceit".

Switching to the frequency domain. Brian Zolner told JVS that the amplifier is "flat to 800kHz". JA's Fig.1 shows that it is actually 200kHz (-3dB) and Fig.6 shows the open loop gain starts to roll off from 1kHz. I would be more impressed if the designer was more aware of the electrical performance than the casework.


CG's picture

A considerable source of distortion with transistors is capacitance that varies with applied signal voltage. Being capacitance, this non-linearity becomes more severe as you go higher in frequency. This alone can cause the distortion curve shape shown in Figure 6, even in amplifier designs that don't use overall loop feedback.

Your conclusion may well be accurate, but it's hard to say for sure.

As for the doubling thing, I think that's hard to pin on the manufacturer. They say that the M15 produces 125 Watts into 8 Ohms. It does that and more, but they only specify 125 Watts. (Not even 126 Watts.) At the power levels they specify, the M15 does double output power as the load halves - with the exception of the 2 Ohm thing. People can infer whatever they want into the spec, but that's on them, not the manufacturer. I don't think that's massaging of any sort.

You can certainly criticize Bricasti over the 2 Ohm issue as well as their flatness claim. The rest falls on the reader(s).

The whole doubling of power into halved loads is one of those urban legend things, at least to me. There's just so many reasons why an amplifier may have less apparent bass than another that you can't put it down to one measure. In many cases, what people call "greater dynamics" is actually just the opposite - the circuit design allows for signal compression, which makes it all sound louder. That approach has been used by guitar peddle builders since they started building pedals. The amplifier power supply often generates much higher noise levels when providing higher current, which can give the illusion of more bass as well.

Are either of these accurate reproduction? Nope. But, if it's pleasing to somebody's ears, it works.

Lots of people genuinely love the sound of SET amplifiers, which tend to compress and distort in a way similar to the way human auditory systems respond to ever louder sounds. There's something to be said for that approach, even if traditional measurements suggest that these amplifiers may not be so great.

The real unfortunate thing is that measurements as done today don't tell the potential customer (or designer!) enough. That's not because audio has some heretofore undiscovered physics behind it all. Instead, it's because the measurement techniques are really not very sophisticated, at least compared to other fields within electronics systems. In many cases, it's not even because the test gear is inadequate. It's because it's used in a way that's misleading or oversimplified. But, that's just the way it's done and has been done.

At this point, my bleating here is adding nothing to anybody's thinking and is just a waste of bits floating about the internet. So, I'm done. Good listening!

Bogolu Haranath's picture

The associated equipment used with the amplifier(s) also makes a difference in the perceived sound quality ....... So, the moral of the story - "listen for yourself" ......... Of course, the measurements are very important ..............

hifiluver's picture

Measurements must surely be sophisticated enough in this day and age to determine if a good and true amplifier is doing its job i.e. amplify with little noise, alteration to source and do it with sufficient watts within a bandwidth. The basic audio amplifier has hit peak technology and done so for decades. What we’re seeing are some manufacturers trying very hard to sell what is basically un-differentiable technology to a not so well informed public. I can sense the confusion and palpable frustration here. Confused customers hold back or they get burnt and it is so strange that those who lament impending doom to this industry are the ones who chart this path.

dalethorn's picture

We have sites on the Web run by people who tear new phones apart and examine every piece, and they tell us as much as they can possibly glean from those observations, as well as any measurements they make. Wouldn't it be great if someone were taking apart expensive amps and speakers etc., and reporting on the quality of the parts and assembly, and whatever else they can discover?

I'll offer just one example of my thinking from the computer business. In 1985, HP introduced the second-generation laptop computer - the first edition being the "110" and the second gen being the "Portable Plus". It cost $3000 in 1985 dollars, or $5000 with 2 mb (not gb) of memory. Disk drives were separate. Mine failed, and spent nearly 3 months in HP's repair shop, while they "couldn't determine" what caused the failure.

We sold the same model to Steve Martin as he was just beginning to start heavy script writing. His also failed. Somehow we found one that worked to keep him up and running. I had good sources inside of HP in those days, and one particular engineer who worked on the project explained: "We had our design, and the bean counters decided that we could "tin-plate" a particular board rather than gold-plate it, and the voltages were low enough that the micro-corrosion was killing them."

So there's a lot that can be learned from tear-downs, especially if someone is using cheaper parts than need be in a high-priced component, on the premise that "it doesn't matter". We have enough good websites to report these teardowns too. But $18k amps, sold in handful quantities, aren't $900 cellphones sold by the millions, so we might never see this happen.