ATI AT522NC Stereo Amplifier Review

[SegaCD]

Many? At what level?

"Many" is a relative term. Like how "many" people can't afford even the "cheap" products reviewed on here! If I may keep riding the Pioneer super tweet train, the HPM-100 was one of the best selling hi-fi speakers in the 1970's (up there with the JBL L100) so "many" may own a super tweet capable of 100kHz.

In any case the low levels of distortion in the ultrasonic band still won't be of any issue.

This is a fairly correct answer. The only issue I see is possible intermodulation distortion with high frequency data in audio. For example, on spectrum analysis of many hi-res audio files sourced from analog tape, I frequently see rather hot 40kHz+ carrier-like signals that are a result of tape biasing during the recording process. If you have a hot 45kHz bias signal with a hot switch-mode PSU at 100kHz, the 3rd order intermodulation distortion harmonic is at 10kHz which is in the audible band. I'd like to assume that this 3rd order harmonic will be at extremely low levels but, you know, its there & should be pointed out.

 

[AudioTodd]

Hey what’s with the “ce” logo? It doesn’t appear to meet the actual drawing specs for it (spacing isn’t right) but looks like what people on this forum have claimed is a “close but no cigar” fake Chinese logo made to deceive the consumer into assuming it is the real logo. Can anyone speak definitively about the logo and its verifiability? I believe Amir or someone else has said it is self-certified, but I’m curious.

 

[Rusty Shackleford]

JA doesn't mention the bandwidth for the measurement which makes a big difference. Here it is at 1 Mhz bandwidth (same 1 kHz tone/amplitude he uses):

View attachment 39664

What now?

Thank you. I believe JA usually specifies 10kHz squarewave into 8 ohms below most graphs.

 

[March Audio]

"Many" is a relative term. Like how "many" people can't afford even the "cheap" products reviewed on here! If I may keep riding the Pioneer super tweet train, the HPM-100 was one of the best selling hi-fi speakers in the 1970's (up there with the JBL L100) so "many" may own a super tweet capable of 100kHz.



This is a fairly correct answer. The only issue I see is possible intermodulation distortion with high frequency data in audio. For example, on spectrum analysis of many hi-res audio files sourced from analog tape, I frequently see rather hot 40kHz+ carrier-like signals that are a result of tape biasing during the recording process. If you have a hot 45kHz bias signal with a hot switch-mode PSU at 100kHz, the 3rd order intermodulation distortion harmonic is at 10kHz which is in the audible band. I'd like to assume that this 3rd order harmonic will be at extremely low levels but, you know, its there & should be pointed out.

So the answer is hardly any and again how much has the response dropped off by 100kHz?

We discussed with plots of ultrasonic distortion and content earlier from a hi res recording.

https://www.audiosciencereview.com/...22nc-stereo-amplifier-review.9999/post-271023

I have looked at a great many hi res recordings and have never seen tape bias signals at high level. The music content and noise signal levels are almost exclusively extremely low, we are talking lower than - 60 or 70 dB above 40kHz. Even a very poorly performing amp will keep that potential IM below the noise floor in the audible band and the Hypex measured above clearly has very low IM from ultrasonic signals.

What you are really saying is that high bandwidth audio is actually a pretty dumb idea. It adds nothing useful to what is audible and has the potential to cause problems.

Edit

Oh of course another thing to consider here is that if you do have a tape with a 45kHz bias signal, that means that the tapes audio recording bandwidth was much lower than this. Most multi track analogue studio reel to reels wouldn't go past 35kHz. So if you are making a digital transfer at higher bandwidth you are wasting your time. You will be just recording noise. It's just dumb audiophilia thinking (by some) that it's beneficial to record tape at 192kHz.

 

[amirm]

I believe JA usually specifies 10kHz squarewave into 8 ohms below most graphs.

I searched and the first one I found had this:

Is he changing it in different reviews?

 

[RichB]

Here is the Stereophile AHB2 small scale 10 kHz signal.

- Rich

 

[amirm]

So he has changed levels and frequency. With 10 kHz square wave, it is essential to know the bandwidth of the measurement system. Without that, you could very well be seeing the rise time of the instrument/its ringing thereof than the amp under test.

 

[DonH56]

DC Blocks: DC blocking circuits are simple but guaranteeing their performance over many years is a challenge and good capacitors to handle high currents are not cheap. Liability is a concern since it is on the power line. DC became a much larger problem when toroidal transformers became the standard (no air gap, which means they saturate easier). They allow smaller designs by placing components nearer the transformers but in many (perhaps most) ways perform worse than standard EI transformers.

Square waves: I like seeing them but more to check for slew rate and stability issues. The frequency matters much less (to me) than the edge rate (rise/fall time). A 1 kHz square wave will allow you to see the harmonic series to see how the amp handles it and if the edge is very fast (maybe 5-10 us? could filter to 100 kHz or something) it will show any transient instability. Using a lower-frequency square wave also allows some idea of settling tails and such that a HF square wave might not show.

Always trades.

 

[restorer-john]

So he has changed levels and frequency. With 10 kHz square wave, it is essential to know the bandwidth of the measurement system. Without that, you could very well be seeing the rise time of the instrument/its ringing thereof than the amp under test.

The square wave shots are more useful if the generator direct is on one trace and looped through the DUT on the other trace.

 

[pma]

So he has changed levels and frequency. With 10 kHz square wave, it is essential to know the bandwidth of the measurement system. Without that, you could very well be seeing the rise time of the instrument/its ringing thereof than the amp under test.

Because it is not the task for AP. It is a task for generator + oscilloscope. Even cheap 25MHz oscilloscope has rise time ≤15ns. And square generator has rise time 10-100ns or less, depending on type. Below my square gen + Pico 2204 USB scope.

 

[pma]

So he has changed levels and frequency.

No. He has always measured both with 1kHz and 10kHz squares, if the component under test was able to do it. BTW, makes not much difference, one may switch time base to read the rising rate and use lower frequency square, rise time is not changed. I am much more demanding to my power amp designs and want for them to pass full swing square wave test, not some small signal. Yes, care about Zobel then.

 

[amirm]

No. He has always measured both with 1kHz and 10kHz squares, if the component under test was able to do it.

Not that I can see. I just looked at the last 10 or so reviews there and only one had both. The rest were 10 kHz. Here is one of them:

What do you learn from this that you want me to lug an entirely different setup to measure it? What source and speakers do you have which supports nanosecond rise time?

 

[pma]

What do you learn from this that you want me to lug an entirely different setup to measure it? What source and speakers do you have which supports nanosecond rise time?

Sorry, this is a misinterpretation, no one wants for you to measure nanoseconds, but it was you who said

With 10 kHz square wave, it is essential to know the bandwidth of the measurement system. Without that, you could very well be seeing the rise time of the instrument/its ringing thereof than the amp under test.

and I am replying that even affordable or cheap systems used primarily for time domain measurements will have no problem with 10kHz square response and will reliably measure rise time or rising rate of any audio component.

I do not need nanosecond systems for audio, but I do need to know impulse response and/or step response. As you know, impulse response completely describes behaviour of a linear system and transfer function is directly derived from impulse response. Impulse response is a derivative of step response, so if I know step response, I know very much about behaviour of any linear system, including stability and hidden oscillations. I need to know both time domain and frequency domain behaviour, even though they are dual. There are issues better seen in frequency domain and other issues better seen in time domain. I cannot emphasize one and not doing the other - that is my design philosophy supported by almost 40 years spent as an EE designer and consultant.

 

[restorer-john]

What do you learn from this that you want me to lug an entirely different setup to measure it? What source and speakers do you have which supports nanosecond rise time?

It's just a function generator and a DSO, you have them both anyway- what's the problem?

 

[pma]

Not that I can see. I just looked at the last 10 or so reviews there and only one had both. The rest were 10 kHz. Here is one of them:

Talking about amplifiers, from solid state power amp reviews
https://www.stereophile.com/category/solid-state-power-amp-reviews

4 of first 5 have 10kHz square measurements and the 5th is 23 years old, so better look to newer reviews - of amplifiers.

https://www.stereophile.com/content/dartzeel-nhb-468-monoblock-power-amplifier-measurements

https://www.stereophile.com/content...ono-ii-monoblock-power-amplifier-measurements

https://www.stereophile.com/content/schiit-audio-aegir-power-amplifier-measurements

https://www.stereophile.com/content/ch-precision-m11-power-amplifier-measurements

 

[restorer-john]

Talking about amplifiers, from solid state power amp reviews
https://www.stereophile.com/category/solid-state-power-amp-reviews

4 of first 5 have 10kHz square measurements and the 5th is 23 years old, so better look to newer reviews - of amplifiers.

https://www.stereophile.com/content/dartzeel-nhb-468-monoblock-power-amplifier-measurements

https://www.stereophile.com/content...ono-ii-monoblock-power-amplifier-measurements

https://www.stereophile.com/content/schiit-audio-aegir-power-amplifier-measurements

https://www.stereophile.com/content/ch-precision-m11-power-amplifier-measurements

Every single magazine reviewer of note, both in the 1970s, 80s and even into the 90s had small signal (and sometimes large) square wave oscillograms. They also depicted overload/recovery behaviour into capacitive loads.

Rise times/slew rates were very much front and centre, once DC coupling and high speed output devices (RETs/diffused emitters etc) became freely available from the likes of Sanken and Fujitsu and to a lesser extent Toshiba. It was a golden age of speed, bandwidth and low THD.

It certainly sorts the sheep from the goats, but certain people will cry from the rooftops about it not being 'fair',' it's inaudible' or it's 'unnecessary' and yet they will trumpet THD figures which are well below anything they can hear as differentiators of superiority.

 

[pma]

It certainly sorts the sheep from the goats, but certain people will cry from the rooftops about it not being 'fair',' it's inaudible' or it's 'unnecessary' and yet they will trumpet THD figures which are well below anything they can hear as differentiators of superiority.

Yeah, however SINAD (numbers!) of -120dB are important and audible . I was hoping this forum is strictly neutral and may accept a balanced view .

 

[amirm]

Sorry, this is a misinterpretation, no one wants for you to measure nanoseconds, but it was you who said

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?

 

[amirm]

Every single magazine reviewer of note, both in the 1970s, 80s and even into the 90s had small signal (and sometimes large) square wave oscillograms. They also depicted overload/recovery behaviour into capacitive loads.

So? They performed a single review every few weeks. And got paid to do it. Neither applies to me. I have to get out a review every day and as it is, I am way behind.

As I just explained, those measurements are academic anyway. We are not trying to design amplifiers. We are trying to evaluate their performance. You are welcome to answer what you get out of the squarewaves I have posted from stereophile and my own.

I ask that because the moment I run a measurement, people will ask me what they mean. If I can't clearly articulate it, then it has no business being run.

As you say, people have performed such measurements since beginning of time. Why is that that my measurements have become so popular? Answer: I keep them short and to the point. Throwing dozens of graphs at viewers is a guarantee that their eyes will roll and they will ignore it all. For this reason, I have a very high bar for addition of any new measurements. People having done them in 1970s is definitely no reason I am going to add them.

 

[amirm]

It's just a function generator and a DSO, you have them both anyway- what's the problem?

I also have a car. It doesn't mean I am going to drive the equipment around the block if you ask me to do so. Everything takes time to do and the worst are the ones that require different gear, different way of capturing the instrument and documenting it. In addition, you need differential probes or dual channel to avoid shorting out the output of the amps that have floating negative. And if you want a load attached at the same time, I would need yet another adapter cable to make it so.