On October 20, 2021, Stereophile published a review of the Dan D’Agostino Master Audio Systems Progression M550 monoblock power amplifiers ($44,950 USD per pair) by Jason Victor Serinus.
The article began with a discussion of the design with the man himself, Dan D’Agostino (D’Agostino’s senior engineer, Burhan Coskun, also sat in). The 550’s basic specs are as follows: it’s a 115-pound monoblock rated to produce 550W into 8 ohms, 1100W into 4 ohms, or 2200W into 2 ohms. The signal-to-noise ratio is said to be 105dB unweighted and 75dB A weighted (maybe a typo?), with .15% distortion at 1kHz at full rated output into 8 ohms.
Serinus included excerpts of his fact-finding discussion with D’Agostino, and this was interesting. Highlights included D’Agostino dissing Krell: “With Krell, I never achieved sonically what I’ve achieved with Dan D’Agostino.” Dan also discussed listening versus measuring and asserted that, now that he is master of his own company, he’s doing it his way: “I think that at Krell, we got involved in a numbers game—we can do this better because it measures better, and it does this better. It was more about technology than listening. But at D’Agostino, we listen a lot. Listening is the most important thing, not how it measures.” Remember that last part.
Serinus then gave a summary of the bits and parts that make up the amp, along with some manufacturer-supplied specifications. After this, D’Agostino told readers about some of the differences between his three lines of products: Progression, Momentum, and Relentless. Then he dissed Krell again: “I had 11 engineers at Krell, but none was as keen on circuits as Burhan.”
Next, Serinus did a bit of foreshadowing: “While the D’Agostino website claims that the Progression’s ‘distortion, signal/noise ratio, channel separation, and bandwidth measurements have all improved,’ that doesn’t guarantee that JA [John Atkinson, technical editor at Stereophile] will find the M550’s measurements even close to superb.”
There’s a problem with the sound!
But not to worry: apparently, this all came down to the rubber feet on the M550s. Once Serinus installed some new supports under the amps, everything was peachy, as they “did more than resolve all lower midrange mush; they also helped smooth out some longstanding rough edges in my system without adding brightness, glare, or other coloration.” Glad we got that sorted.
Serinus then listened to the amps and loved them, impressed with “. . . the Progression M550s’ ability to convey the inner warmth and depth of someone transported. I was equally struck by the clarity of a piano recorded in a surprisingly quiet and exceptionally dimensional acoustic. The silence was as astounding as the heart that shone through this great artist’s voice. When an amp can deliver all that from a simple performance with voice and piano, it’s a great amp.” Serinus went on to compare the new Progressions with the old and found the new ones much better. There is certainly no surprise there. I would expect a new version of most anything to be better than the previous one. But, of course, that’s not always the case.
Next, Serinus received a set of Wilson Audio Acoustic Diode spikes ($3200 for a set of eight) and these, along with his recent electronics changes, necessitated a reassessment of the physical locations of his speakers in his room. He subsequently moved his Alexia 2s around and, with the improvements wrought by the newfound speaker locations, discovered that the M550s sounded different from each other. He next looked into the cause of the channel-to-channel mismatch. Long story short: there was an issue with the amps’ biasing. A tech from a D’Agostino dealer then came by and rebiased the amps, and all was well in the world . . . or at least Serinus’s audio system. In the end, Serinus concluded that the Progression M550s “are wonderful amps.”
Then John Atkinson went and measured a sample of the Progression M550. Cool. We at SoundStage! like measurements along with our listening tests. So much so that we bought an Audio Precision APx555 B Series audio analyzer ($30,000) and hired Diego Estan as SoundStage!’s electronics measurement specialist. Take a look at Diego’s comprehensive suite of measurements of the Simaudio Moon 860A v2. Diego’s handiwork is as complete as we can make it, and it gives those inclined to learn about and interpret electronics measurements microscopic-level insight into the product being tested.
Since Diego is so used to looking at measurements of audio components, I asked him to give us a quick summary of what he noticed in Atkinson’s M550 measurements. There were some problems.
In general terms, given the price tag of this solid-state amplifier, a few of the Stereophile measured parameters are quite poor by modern hi-fi standards. The measured power output into 8 ohms does meet Dan D’Agostino’s claim of 550W into 8 ohms, and with higher AC line voltage, no doubt the measured 4-ohm power at clipping would have been close to the 1.1kW claimed. What is noteworthy to mention here is that, while it’s possible for a manufacturer to have a specified power output into 4 ohms twice that of the 8-ohm specified power output, it is not possible for any amp to truly double its power when the load is halved for a constant input signal amplitude. This would imply an output impedance of zero, which is not possible.
On the subject of output impedance, in terms of the measured parameters that stood out as poor by modern hi-fi standards, the 0.37 ohms at 1kHz, which equates to a damping factor of 21.6 (8 ohms), is surprising. One would expect a huge amp like this one to have a litany of paralleled output transistors, which should yield an output impedance at least an order of magnitude lower than what we see here. Another example of a poor measurement is the high THD+N figures at various power output levels. In particular, at 1-20W, where I suspect many users with reasonably sensitive speakers will end up operating this amp, THD+N varied between 0.15 and 0.25%. Is this below the threshold of audibility? Perhaps just, but it’s still quite poor, and at least an order of magnitude worse than what the competition would typically offer. Yet another example of a poor measurement is presented by the multiple signal harmonics in the 50Hz 100W FFT, which are in the -70dBrA (0.03%) to -80dBrA (0.01%) range, and JA’s description of the signal’s second harmonic for a 1kHz input signal at 1W of -60dBrA (0.1%). These measurements indicate that it’s mostly distortion that is responsible for the high THD+N figures noted above. JA goes one step further and shows a scope capture of the distortion pattern with the signal notched out, in which the peaks line-up where the signal crosses the 0V threshold. JA concludes that the high crossover distortion, which he suggests may be due to insufficient output transistor biasing, is puzzling. Whether this is caused by insufficient output-stage biasing or perhaps insufficient negative feedback, and whether it’s intentional or not, I agree that both the high levels of distortion (especially at lower power) and high output impedance are puzzling for a solid-state amp of this perceived pedigree.
Stop the presses!
If you guys think this is a hit piece on Stereophile, it’s not. Stereophile seemed—from what I could see—to be transparent in their reporting of the events surrounding this review, and that is indeed admirable. They could have easily concealed the whole kit and caboodle from the reader, and yet, they chose not to. Kudos to Stereophile.
I’m also not intentionally disrespecting Dan D’Agostino (however, I don’t think the Krell remarks were necessary; it’s been a long time, move on). Dan is making products that he likes and that his many customers around the world seem to like, too. And let’s face it: Dan is a legend in the industry. Despite what he now thinks of the Krell products produced under his tutelage in decades past, some of those components—like the legendary KSA-250—are classics. And those D’Agostino products are simply gorgeous.
No, it’s the readers—you guys—that I’m concerned about. There are many great power amplifiers out there that measure superbly and sound amazing and, in many cases, cost less than a pair of Progression M550s. Take the previously mentioned Simaudio Moon 860A v2 stereo/mono amplifier ($19,500) that Jason Thorpe reviewed for SoundStage! Hi-Fi on November 1. Jason loved that amplifier (a set of mono Moon 860A v2s would come in about $6000 shy of the M550s’ price). When we look at Diego’s measurements, the Simaudio/D’Agostino amplifiers are a study in contrasts. Diego summarizes this as follows:
There are myriad measurements that can be made with several different reference points, many of which, unfortunately, do not align between SoundStage! and Stereophile. Nonetheless, it is possible to draw a few important comparison points. For example, if we look at THD+N vs output power into 8 (blue/red) and 4 ohms (purple/green) below for the Simaudio Moon 860A v2, we not only find that the Moon is nearly load impervious (i.e., it shows the same results into both 4 and 8 ohms), but also yields a THD+N between 1 and 20W that is more than ten times lower than the Progression amp, at 0.005%-0.008%.
Diego: Our 50Hz FFT measurement below for the Moon amp, while at 10W into 8 ohms compared to JA’s 100W, does show worse-case signal harmonics in the -90dBrA (0.003%) range, 20dB better than what the Progression yielded.
Diego: Next, if we look at the Moon’s IMD FFT result (10W, 8 ohms, 18+19kHz) below, we find the second-order difference signal (1kHz) at around -100dBrA; this is about 30dB better than the result JA measured at 100W (-73dBrA). The third-order distortion products (17/20kHz) for the Moon amp are just above -90dBrA, while JA measured -75dBrA for the Progression amp, a difference of 15dB in favor of the Canadian amp.
Diego: Finally, if we look at the Moon’s damping factor as a function of frequency below, we find a damping factor of 1000 (1kHz, right channel), which is nearly 50 times higher than what JA measured with the Progression amp!
Diego: This very high damping factor, coupled with a very large and robust power supply, makes the Moon 860A v2 nearly impervious to changes in THD when the load is reduced. This is evidenced by the data I collected below, which shows THD vs frequency for 8/4/2 ohms at roughly 20/40/80W. What is impressive here is that THD is constant and unchanging regardless of load from 20Hz to about 500Hz.
When I picked my reference amplifier—an amplifier I’d heard many times before and therefore knew very well in terms of its sound—I still insisted on seeing the actual measurements of the unit I was receiving. I wrote about my Boulder Amplifiers 2060 stereo amplifier back in April of 2018, after it made a side trip to Colorado-based Boulder Amplifiers for a check-up before making its way to me in North Carolina (I purchased this amplifier used). Although Diego wasn’t measuring gear for us back in 2018, Boulder provided me with a standard suite of their own tests. And I published the measurements they supplied. Diego had a look at them recently and described what he saw:
With the Boulder, we find THD+N figures that are squeaky clean: at 53W into 8 ohms, 0.00053% at 2kHz and 0.00082% at 20kHz (10Hz to 30/80kHz measurement bandwidth). That’s better than -100dB below the signal, distortion and noise combined. For a large solid-state amp with a huge linear power supply, it doesn’t get much better. As a point of comparison, the Moon 860a v2 measured just above 0.01% THD+N at the same power level at 1kHz, while the M550 measured 0.06%.
I could probably write a book on my takeaways from the Stereophile review, the D’Agostino measurements, and the ensuing kerfuffle (Stereophile turned off the comments section in this review posthaste), but I’ll make this next part brief.
Audiophiles tend to fall into either the measurements camp or the listening camp. Not unlike politics, the debates that ensue get downright contentious. Personally, I’ve never felt the urge to define myself in this way. I like more data, not less, and therefore, when I’m compiling info to base a buying decision on, I prefer to consider my own listening tests in conjunction with a set of comprehensive measurements. Where the positive results from both tests converge, I tend to find my contenders.
Rather than adhering to a diehard belief in the benefits of one criterion to the exclusion of the other, my perspective aligns more towards being skeptical of both: 1) I don’t believe measurements tell us everything there is to know about a product; 2) I also know that my hearing—and my assessment of the results of my listening tests—are fallible. To me, having both is really a system of checks and balances.
Bottom line: I’d pass on a product that didn’t produce excellence in both the measured and the listening test results.
. . . Jeff Fritz