Topping PA5 Review (Amplifier)

[J-Sine]

Because distortion above 10K is supersonic ultrasonic . What the multitone test shows is (at least at 5W) that higher frequency distortion is largely above 20KHz, even including IMD products folding back into the audible spectrum.

Actually this plot shows otherwise. That there is Frequency distortion as low as 5khz to 15khz. Granted there is very little granularity.

As I understand it, with a 40 khz filter as used by Amirm when measuring class D, everything IMD above about 13.khz are filtered out.

This brings partial light as to why but then it begs to ask why do Class-D amps get special testing?
Asking this can only create an inflammatory disposition with the community. Not my objective.
This however leads to further scrutiny of Class-D designs with their known Phase Linearity issues and more specifically this Frequency Distortion above 5k. Generally it likely comes from the LC output filter(coil/choke capacitor) integral at the core of its design.. Since it shows it has frequency reactivity then it likely also has phase issues. I believe currently Harman Citation and Purifi are trying to address this in Class-D foremost at the moment.

Your reference is a person who just happens to sell equipment marketed as having low phase distortion. At no point did you or him reference a study backing the claims on the audabilty of phase distortion at high frequencies.

Tons of AES papers on the properties of phase linearity in audio. Many companies extolled its value from the 70's on wards. Harman Citation designs called for High Bandwidth to 300khz or higher to address this so that the 1khz to 20khz were not phase shifted. Mark Levinson employs this discipline in design as well. Phase linear(the company) which became Carver I believe were clearly inspired. Benchmark isn't the only company that acknowledges it's importance in Amp Design Discipline.

Here's an example of a poor performing class a/b at Sinad 65db https://www.audiosciencereview.com/...on-2004-mk-ii-review-vintage-amplifier.29248/

This matches its frequency distortion plot and actually shows mostly linear frequency distortion. Even in poor performing designs Frequency distortion is mostly Linear.

NOTE:The Topping PA5 almost matches this level of frequency distortion plot at 10khz and 15khz and it clear that as frequency increases above 5k so does distortion in a frequency specific way.

Arguably the IMD is superior on the PA5 but we are told it these are special "Class D" filtered plots. More important is that Phase Linearity isn't tested/plotted and we can only extrapolate from the Frequency Distortion that it may lack linearity.

We all want to believe Topping made audio Nirvana for $350 but Topping did not design the TI TPA32XX chip with its potential SNR 111db THD they just implemented it well with its known inherent Phase frequency flaws and frequency distortion. Check TI TPA32xx Data Sheets.

I cannot dismiss what I see with my eyes nor hear with my ears.

What is most disconcerting is that many new people to audio will buy a PA5 as the new Budget "benchmark" reference thanks to ASR SINAD and become biased that this is the way Audio "transparency" should sound. Yet if the PA5 was implemented in a Multichannel ATMOS configuration with its potential phase linearity flaws it would become apparent. Especially if mixing amps.

Thanks J-Sine for this well written post. That Benchmark AHB2 would be nice, one day it will be mine. This PA5 has a lot going for it, most of the time it sounds really great. Sometimes, something sounds off to me but I am also comparing it to more expensive amps, Class AB designs...

I know this a long thread but a lot of your previous posts regarding sonic impressions and a few others correlate to exactly the issues I am addressing.
For a stereo setup the PA5s flaws will likely bring a new clarity to certain instruments. Yet knowing that the sound stage is altered as an artifact the naturalness of many tones will be skewed. Other Class A/B designs with lower ranking SINAD with phase linearity will sound more natural but likely more generic. Knowing ones gear and its limitations allows you to use it more effectively. Much like home theater systems vs a 2 channel playback setups.

The PA5 exposes the "One SINAD to Rule them all" as potentially flawed and Class-D amps need further phase tests included.

Despite my aging ears I still listen above 1KHZ and the harmonics to at least 15khz and relative phase coherency matters. Speakers are still the greatest offenders and constant directivity design is trying to address them.

As a whole phase linearity does matter but to what degree is currently a hot debated topic. Class-D has brought it to the forefront again.

 

[J-Sine]

If you are worried about phase, have you seen the step response for your speakers?

I agree, Speakers designs are full of phase coherency issues and compromises. Then the listening room environment factors heavily. Even Dr Floyd Toole has a terrible room but makes the best of it. See: https://www.thescreeningroomav.com/...te-real-world-home-theater-and-listening-room

If the PA5 was mid tier testing amp, I wouldn't even bother. It isn't and is testing 2nd best in SINAD with all to date amplifiers on ASR.
It merits Higher Scrutiny.

In short: The theory is that above 96db SINAD it is difficult to hear a Distortion difference. So the SINAD 105db PA5 should sound indistinguishable from the Benchmark AHB2 is quite possibly flawed.

Subjective listening tests for some are showing the PA5 to be questionable.

I'm certainly not biased to any topology. Class-D has size ,cost,component and efficiency advantages which is great. Yet SINAD isn't fully addressing its flaws.

 

[AudioArchitech]

I also have a hard time believing that the Yamaha R-N803 sounds better than Audiolab 6000A. Both Class AB amps and the R-N803 ranks much higher in Sinad according to ASR. But it’s possible, would love to hear both.

 

[MaxBuck]

I also have a hard time believing that the Yamaha R-N803 sounds better than Audiolab 6000A. Both Class AB amps and the R-N803 ranks much higher in Sinad according to ASR. But it’s possible, would love to hear both.

Much higher SINAD doesn't necessarily translate to better sound.

 

[Andrej]

Much higher SINAD doesn't necessarily translate to better sound.

What does "better" mean in this context, to you? Higher fidelity? Personal preference? Which aspect of fidelity? Unbiassed personal preference (blind evaluation) or biassed preference? SINAD only claims to be a measure of higher fidelity along some dimensions. And it does it extremely well. It may not correlate well with other versions of "better", and it does not claim to do so. However when claiming that something is "better", specifying in what way it is better is extremely important, or it becomes a meaningless statement.

 

[anotherhobby]

Did you get your Flex yet? If so, how is working out for your. I got a Minidsp 2x4HD and use it as the DAC and crossover. I’m not really happy with it. It doesn’t sound as clean as my other setup which was DAC to old preamp to PA5.

I did get it! It came 2 days ago. For me it was a big improvement having better and properly matched crossovers, plus being able to EQ the subs independently and in a mono configuration. I still have more fiddling to do yet, but so far I'm super happy. Also, I have more gain on the PA5 (as well as my Crown) since I'm feeding balanced signals now.

 

[antcollinet]

Actually this plot shows otherwise. That there is Frequency distortion as low as 5khz to 15khz. Granted there is very little granularity.

View attachment 182613

You misunderstood my statement. I said the distortion is ultrasonic above 10Khz fundamental.

10Khz harmonics will be 20K, 30K 40K etc - all ultrasonic. The test plot you reference is tested with a 45kHz bandwidth, and shows THD+N up to around -75dB for 20KHz, -87 for 10KHz, and -93 or lower for 5KHz and below.

Yet the multitone (showing only distortion products in the audio band up to 20KHz) has most distortion products below about -120 until you get to higher frequency distortion products, where they just peak above -110. And SINAD (also audio band) is -110,

So that tells us that much of the distortion shown in the power/distortion chart must be products above 20KHz and hence ultrasonic - the answer to your question of "why SINAD and distortion tests do not reveal this distortion"

 

[antcollinet]

Arguably the IMD is superior on the PA5 but we are told it these are special "Class D" filtered plots. ...

Where do you get that from. The power distortion chart has the same 45KHz bandwidth, and the multitone test shows the same 20KHz band, done at the same power of 5W

 

[georgeT]

You'd think this claim would have been addressed on this site already. Oh wait, it has been:

 

[antcollinet]

Subjective listening tests for some are showing the PA5 to be questionable.

Those tests are meaningless in terms of an assessment of the PA5 unless they are controlled tests (level matched/blind/statistically significant).

If someone wants to listen and say "this one is not for me" then that is fine - but it doesn't say anything about the actual performance of the amp for anyone else, nor anything about a false correlation between objective measurements and subjective listening.

I'm certainly not biased to any topology. Class-D has size ,cost,component and efficiency advantages which is great. Yet SINAD isn't fully addressing its flaws.

What is your evidence for these flaws?

 

[gfinlays]

I think people are splitting hairs here. Looking at the graphs above, phase shift ("distortion") at 5kHz starts to increase when the PA5 is pushed above about 5 Watts output.

A 5kHz tone cycles through a full 360° 5000 times a second, or in 0.2 of a millisecond. Even a 90° phase rotation at 5kHz gives a time delay of just 0.05 milliseconds! Those amounts of delay are completely inaudible in a normal room, whether in 2-channel or multi-channel listening.

I'm sure there will be a lot of "noise" from Benchmark for example regarding Class D. I posit that the newest Class D designs Topping PA5, Purifi, Hypex NC400 etc. will be audibly indistinguishable from the AHB2 in properly controlled listening tests.

A competitive product offering audibly equivalent performance at a significantly lower price point has the potential to adversely affect your sales volume.

 

[Actungz]

Those tests are meaningless in terms of an assessment of the PA5 unless they are controlled tests (level matched/blind/statistically significant).

If someone wants to listen and say "this one is not for me" then that is fine - but it doesn't say anything about the actual performance of the amp for anyone else, nor anything about a false correlation between objective measurements and subjective listening.



What is your evidence for these flaws?

So I only have limited experience with AVR’s and stereo receivers as amplifiers, but from your comment (and many others here) I’m getting the idea that there are no ways to compare amplifiers. If there is no major flaws in the measurements (high noise flower, low SINAD, high THD), then any amplifier should sound the same?

Does that mean if I never listen above 85 decibels with a 86dB sensitivity pair of speakers at less than 1m away, then any amplifier would reasonably work and sound the same? Since I would never be pushing the power supply at all, and any impedance dip wouldn’t really matter since I’m using so little power?

Whenever I see words such as, “better woofer control”, “better dynamics”, “better at lower volume” I just get super confused. Is this a higher volume thing that I don’t understand?

Thanks

 

[antcollinet]

So I only have limited experience with AVR’s and stereo receivers as amplifiers, but from your comment (and many others here) I’m getting the idea that there are no ways to compare amplifiers. If there is no major flaws in the measurements (high noise flower, low SINAD, high THD), then any amplifier should sound the same?

Does that mean if I never listen above 85 decibels with a 86dB sensitivity pair of speakers at less than 1m away, then any amplifier would reasonably work and sound the same? Since I would never be pushing the power supply at all, and any impedance dip wouldn’t really matter since I’m using so little power?

Whenever I see words such as, “better woofer control”, “better dynamics”, “better at lower volume” I just get super confused. Is this a higher volume thing that I don’t understand?

Thanks

No

What I wrote means you cant trust your subjective judgement that one sounds significantly different from another unless you do it blind. You may be hearing an actual difference in the soundwaves reaching your ears, or you may just be perceiving a difference based on bias/expectation etc. Or just because of the mood you're in - or that you are feeling unwell, or have had a drink. There is no way to tell.

If you listen to two well designed amplifiers both with very good SINAD - it is likely that any difference you hear will not be audible when you don't know which one you are listening to. That doen't mean certain it will disappear - there may be reasons for a difference, for example hard to drive speakers, but you simply don't know until you try.

 

[Xulonn]

What is your evidence for these flaws?

Tony, the mere fact that you are knowledgeable enough about science to ask for "evidence" rather than "proof" makes you far more credible than many who post here at ASR and claim to be science literate. I appreciate your contributions.

 

[antcollinet]

Tony, the mere fact that you are knowledgeable enough about science to ask for "evidence" rather than "proof" makes you far more credible than many who post here at ASR and claim to be science literate. I appreciate your contributions.

Thanks.

 

[mocenigo]

What does "better" mean in this context, to you? Higher fidelity?

I do not know what @MaxBuck meant, but higher SINAD does not translate to higher fidelity sound overall. It only means lower noise and distortion (overall) ad 1Khz. It is, frankly, an absolutely ridiculous metric. It is ALMOST a pure gimmick to attract most of the traffic to this site, in other words not to attract the (many) knowledgeable forum participants, but to attract the masses that contribute to traffic. I say "ALMOST" because it helps separate the wheat from the chaff. But Amir provider many other measurements for those that can understand them (and according to these the PA5 is a good amplifier, but nothing to scream about if it were not for the low price).

Roberto

Personal preference? Which aspect of fidelity? Unbiassed personal preference (blind evaluation) or biassed preference? SINAD only claims to be a measure of higher fidelity along some dimensions. And it does it extremely well. It may not correlate well with other versions of "better", and it does not claim to do so. However when claiming that something is "better", specifying in what way it is better is extremely important, or it becomes a meaningless statement.

 

[maty]

I think for many people, this PA5 will be great. I have high end, revealing speakers (maybe too revealing) and they show this amp's weaknesses at that level. I was hoping for a $450 (CAD) giant killer and it didn't work out for me but this is still a great little amp. I just did another A/B comparison with Aiyima A08 and the PA5 is certainly better than A08 in my system. I'll sell it and get something else. Some ideas: Parasound 275V2, Emotiva BasX A2, Buckeye or Audiophonics NC252MP, IOTAVX PA3. Or maybe just get another Audiolab 6000A? Any other ideas? looking for best bang for buck.

Yamaha S-701 or S-501 are better option than cheap class D amps to listen to good recordings, I suspect. Jazz, orchestral, folk... you know, with real instruments and voices without autotune. With a good external DAC. And Pure Direct.

YAMAHA A-S701 - tear-down, thoughts about the internals & few measurements

I bought this amplifier with built-in DAC over 1 year ago from Amazon and had no issue using it several hours per day. My kid is using it daily on TV and XBOX over the SPDIF input, I'm using it couple of times per week as well and it works as expected, as in day one. It's barely warm during...

www.audiosciencereview.com

Yamaha N-803D but with external DAC. And Pure Direct.

Yamaha R-N803 Smart Receiver Review

This is a review and detailed measurements of the Yamaha R-N803 Receiver, amplifier, and streamer. It was purchased by a member (b-stock?) and drop shipped to me. The R-N803 costs US $750 from amazon including Prime shipping. The R-N803 has AV-style features such as Room EQ but with only two...

www.audiosciencereview.com

 

[Vini darko]

No

What I wrote means you cant trust your subjective judgement that one sounds significantly different from another unless you do it blind. You may be hearing an actual difference in the soundwaves reaching your ears, or you may just be perceiving a difference based on bias/expectation etc. Or just because of the mood you're in - or that you are feeling unwell, or have had a drink. There is no way to tell.

If you listen to two well designed amplifiers both with very good SINAD - it is likely that any difference you hear will not be audible when you don't know which one you are listening to. That doen't mean certain it will disappear - there may be reasons for a difference, for example hard to drive speakers, but you simply don't know until you try.

I disagree that 1khz sinad would determine if two things sound the same.

 

[Andrej]

I do not know what @MaxBuck meant, but higher SINAD does not translate to higher fidelity sound overall. It only means lower noise and distortion (overall) ad 1Khz. It is, frankly, an absolutely ridiculous metric. It is ALMOST a pure gimmick to attract most of the traffic to this site, in other words not to attract the (many) knowledgeable forum participants, but to attract the masses that contribute to traffic. I say "ALMOST" because it helps separate the wheat from the chaff. But Amir provider many other measurements for those that can understand them (and according to these the PA5 is a good amplifier, but nothing to scream about if it were not for the low price).

Roberto

I claim that it does, by it's very definition. In terms of the metrics it reports. However, I am very much interested in learning, one of the reasons I spend too much time reading posts on this site. Can you please specify in what other ways (other metrics) you measured fidelity to determine that SINAD (alone?) does not translate into a correlate of fidelity?

 

[mocenigo]

I claim that it does, by it's very definition. In terms of the metrics it reports. However, I am very much interested in learning, one of the reasons I spend too much time reading posts on this site. Can you please specify in what other ways (other metrics) you measured fidelity to determine that SINAD (alone?) does not translate into a correlate of fidelity?

Yes, higher SINAD means it reproduces a single sinusoid at 1Khz better than a lower SINAD device. It is also strongly correlated to other metrics, and a very high SINAD is difficult rot achieve without being very good in many other areas. I am not denying it. But there are examples *here* (one discussed in this very thread).