ATI AT522NC Stereo Amplifier Review

[pma]

We need much faster square or step rise time to investigate stability of the amplifier that has to be unconditional.

And even digital sound sources are not restricted to Fs = 44.1kHz. Further, you spoke about 40us. Even for 44.1 kHz sampling the fastest rise time is about or below 20us, depending on digital reconstruction filter response. For 96kHz it is below 10 us. It is not possible to soften test conditions just for the reason to get nicer plots.

 

[SIY]

Square Wave...

I'd like to see the AP's harmonic display on one of those as far as it will go, just for fun.

This (10Hz square) was one posted here earlier, with lesser precision, and less bandwidth:

(click and double-click to expand the image)

View attachment 39847

I estimated 1,150 harmonics displayed, much to my surprise, both in them being generated, and in the ability of the software to see them all at once.

S'prise me mo'.

 

[RayDunzl]

Wow...

 

[March Audio]

You didn't address my question. I asked what you learned from the graph that I provided from stereophile review.

On the rise time, you didn't address that either. 99% of our audience here uses digital as a source and 44.1 kHz sampling. There, you can NOT have a squarewave at 10 kHz. Squarewave can be decomposed into its frequency and odd harmonics. A 10 kHz will have a sine wave at 10 kHz. Its next component will be at 30 kHz which the DAC will NOT output. So all you get is a sine wave if you tried to play a 10 kHz sine wave!

Even if you upped the sample rate, you still cant' get anything comes close to square wave with a digital source. But let's say you did. Then your amplifier such as the one I have tested band limits it. And so does the speaker.

So I ask again, what will you learn from a high speed measurement of an amplifier when it doesn't resemble what you can create in your audio system?

My analyzer can go up to 1 MHz. I showed you those measurements. Did you think that was a limit and what did you learn from them?

This is the point. A square wave is helpful as a design evaluation tool to investigate stability, peaking etc, but as anything representative of real world encountered signals, it is not. If you want to investigate the above you also need to test the gamut of reactive loads.

The question is where do you stop. Amir could spend literally weeks testing every aspect of an amplifier performance and produce a 100 pages of test results, but is that useful to the wider audience?

Is this a magazine for electronics engineers?

 

[pma]

99% of our audience here uses digital as a source and 44.1 kHz sampling. There, you can NOT have a squarewave at 10 kHz.

Uh - huh, these are real banalities and complete ignorance. Only ignorant would use 44.1kHz sampling with a brickwall to generate the square. I thought this was intended to be a scientific forum, not a kindergarten.

 

[March Audio]

It looks to be .25 volt swings on the waveform that should not be there.
How should I inerpret this?

- Rich

It's the amplifier switching signal at around 450kHz in Hypex designs. This is filtered out by the speaker. Obviously tweeters have no response at this high a frequency so it literally isn't there as an audio signal. It's simply that a scope has high bandwidth and can see it.

So if unfiltered, as you found, to the uninitiated the plot is a little confusing and misleading.

 

[March Audio]

OK. Looking at this individual measurement, I do not like the result. Oscillations on top and bottom of the squares may indicate to instability or to output filter resonances.

Ok,so if Amir does this test in isolation what firm conclusions can you draw?

 

[March Audio]

@restorer-john, @SIY, @pma

Question, what's the most capacitively reactive speaker load you have seen? Biggest phase angle?

 

[DonH56]

Plenty of speakers dip into the 2-ohm range, ESLs often <0.5 ohms, with phase angles of up to around 60 degrees (positive and negative). A quick look through some Stereophile reviews provides lots of data supporting that range. Some amps, tube in particular, also struggle on the other end when impedance goes way high (20~40 ohms or more) even with phase angles below 45 degrees.

Probably not the place to delve into Nyquist stability criteria, root locus plots, and all that jazz... IME/IMO a square wave with a fast edge can be useful to show if an amplifier may have stability issues. In the past I used 1 us edges, so about 350 kHz equivalent (first-order) bandwidth, to look for instability into speaker loads. Most were fine, but occasionally one would exhibit large poorly-damped ringing or outright oscillation.

Analog systems still have wide bandwidth even if digital systems are more band-limited. But higher sampling rates provide the potential for >100 kHz signals intentional or not. This, for me, is less about hearing or musical content (not even a consideration unless the amp is really bad) and just another basic performance test. I have had amplifiers that oscillated ultrasonically and destroyed speakers in the past. Pathological? Maybe, but I don't want to have it happen again... I don't really care to look for music with 100 kHz+ content nor loudspeakers with 0.2-ohm 60 degree impedance spikes; to me it is another of the numerous sort of boundary condition tests I routinely perform on products (RF, mW, mmW, audio...) Steady-state sweeps of sine waves limited to 20 kHz do not represent a worst-case condition to me. As a designer, I am always working to test the worst case and beyond, to make sure the amplifier (or whatever) is stable for any reasonable (and most unreasonable) loads.

 

[miero]

Infinity Kappa 9 is also a nice example "the impedance dips to 0.8 ohm at 33 hz and 0.9 ohms at 7000 hz".
Source: http://www.audioreview.com/product/speakers/floorstanding-speakers/infinity-systems/kappa-9.html?p=2

 

[March Audio]

Plenty of speakers dip into the 2-ohm range, ESLs often <0.5 ohms, with phase angles of up to around 60 degrees (positive and negative). I quick look through some Stereophile reviews provides lots of data supporting that range. Some amps, tube in particular, also struggle on the other end when impedance goes way high (20~40 ohms or more) even with phase angles below 45 degrees.

Probably not the place to delve into Nyquist stability criteria, root locus plots, and all that jazz... IME/IMO a square wave with a fast edge can be useful to show if an amplifier may have stability issues. In the past I used 1 us edges, so about 350 kHz equivalent (first-order) bandwidth, to look for instability into speaker loads. Most were fine, but occasionally one would exhibit large poorly-damped ringing or outright oscillation.

Analog systems still have wide bandwidth even if digital systems are more band-limited. But higher sampling rates provide the potential for >100 kHz signals intentional or not. This, for me, is less about hearing or musical content (not even a consideration unless the amp is really bad) and just another basic performance test. I have had amplifiers that oscillated ultrasonically and destroyed speakers in the past. Pathological? Maybe, but I don't want to have it happen again... I don't really care to look for music with 100 kHz+ content nor loudspeakers with 0.2-ohm 60 degree impedance spikes; to me it is another of the numerous sort of boundary condition tests I routinely perform on products (RF, mW, mmW, audio...) Steady-state sweeps of sine waves limited to 20 kHz do not represent a worst-case condition to me. As a designer, I am always working to test the worst case and beyond, to make sure the amplifier (or whatever) is stable for any reasonable (and most unreasonable) loads.

Yes the question wasn't based around audibility but the potential for creating instability.

Whilst a designer wants to define the limit conditions, is that our purpose here? I don't think so. Demonstrate to me real world conditions and signals that its not going to melt into destructive oscillation, yes. I don't mean benign but equally not pathological. Just representative.

 

[pma]

Yes the question wasn't based around audibility but the potential for creating instability.

Whilst a designer wants to define the limit conditions, is that our purpose here? I don't think so. Demonstrate to me real world conditions and signals that its not going to melt into destructive oscillation, yes. I don't mean benign but equally not pathological. Just representative.

This is a measurement of one of my power amplifiers into 0R22 resistive load (0.22 ohm). What do you think?

 

[AudioTodd]

It's the amplifier switching signal at around 450kHz in Hypex designs. This is filtered out by the speaker. Obviously tweeters have no response at this high a frequency so it literally isn't there as an audio signal. It's simply that a scope has high bandwidth and can see it.

So if unfiltered, as you found, to the uninitiated the plot is a little confusing and misleading.

I get that tweeters - as far as I am aware - have no response at that frequency, but isn’t there energy there that must be dissipated somewhere, like in the tweeter? Would this cause some heating of the voicecoil or some component in the system?

 

[SIY]

Plenty of speakers dip into the 2-ohm range, ESLs often <0.5 ohms, with phase angles of up to around 60 degrees (positive and negative). A quick look through some Stereophile reviews provides lots of data supporting that range.

I've seen nasty impedance magnitude dips like that in a few incompetently engineered high end speakers (like some of the Infinitys of yesteryear as well as Wilsons). Never seen that in any of the ESLs I've measured (Quads, Acoustats)- typically, the impedance falls with increasing frequency, but then starts rising again because of the interaction between the panels and the transformer. Do you remember an example offhand?

 

[pma]

It might be "incompetently engineered", as you are saying, but once you bring your amp into the comparative test you do not like to end up with saying a phoney express like "the speaker used is incompetently engineered". You want to show that you are able to drive it regardless its challenging complex impedance. This is the case for the amp designer.

 

[SIY]

It might be "incompetently engineered", as you are saying, but once you bring your amp into the comparative test you do not like to end up with saying a phoney express like "the speaker used is incompetently engineered". You want to show that you are able to drive it regardless its challenging complex impedance. This is the case for the amp designer.

Well, as with everything else, it depends. If I am advertising an amp as capable of driving every single commercial speaker out there, then, sure, how it handles the load of incompetently designed speakers is important.

But that comes with a significant cost penalty, and 100% of the users would be paying significantly more for the capability of driving that last 0.1% of the speakers out there. For 99.9% of speakers, it's just not necessary.

 

[DonH56]

I've seen nasty impedance magnitude dips like that in a few incompetently engineered high end speakers (like some of the Infinitys of yesteryear as well as Wilsons). Never seen that in any of the ESLs I've measured (Quads, Acoustats)- typically, the impedance falls with increasing frequency, but then starts rising again because of the interaction between the panels and the transformer. Do you remember an example offhand?

It was one of the recent Martin Logan ESLs, maybe the new 15A? I'd have to dig... And some of theirs dip to ~0.2 ohms at 20 kHz or so.
<pause> https://www.stereophile.com/content...-renaissance-esl-15a-loudspeaker-measurements states 0.52 ohms at 20 kHz from manuf data but does not show the curves.

IME, which is dated now, the impedance magnitude for ESLs gets low at HF but phase angle is not horrible because the transformer drive compensates much of the panel's capacitance.

Big Maggies (2.xR, 3.x, 20.x, and now 30.x) dip below 3 ohms due to the ribbon tweeter but phase angle is near 0 degrees (almost purely resistive).

I'm just going to ignore the early Apogees -- can't imagine too many of them around now, and no new ribbons as pathologically low AFAIK.

 

[SIY]

It was one of the recent Martin Logan ESLs, maybe the new 15A? I'd have to dig... And some of theirs dip to ~0.2 ohms at 20 kHz or so.
<pause> https://www.stereophile.com/content...-renaissance-esl-15a-loudspeaker-measurements states 0.52 ohms at 20 kHz from manuf data but does not show the curves.

Gracias! I've added a couple speakers to my list of "stuff I never want in my house." It's odd not to show the impedance in the measurements- I wonder why they didn't do it in this case?

 

[pma]

Well, as with everything else, it depends. If I am advertising an amp as capable of driving every single commercial speaker out there, then, sure, how it handles the load of incompetently designed speakers is important.

But that comes with a significant cost penalty, and 100% of the users would be paying significantly more for the capability of driving that last 0.1% of the speakers out there. For 99.9% of speakers, it's just not necessary.

Yes you are right, taking into account strictly reasonable engineering, cost-effective approach. However, I have to substitute JC's attitude a bit, here , and I assume you understand. To me, audio is a hobby and finding the ways of no-cost-related best solution. Not a business approach, you know, just a challenge.

 

[SIY]

Yes you are right, taking into account strictly reasonable engineering, cost-effective approach. However, I have to substitute JC's attitude a bit, here , and I assume you understand.

Oh, I absolutely did understand your point. But for reviewing purposes, I use a Pareto approach.

That said, if I had a PowerCube, I'd be using the hell out of it.