Amplifier measurements may require improvement

[peng]

Which kind of music did you use? -Which tracks?

At that high level I used the following album, didn't pay attention to which track, but a few tracks anyway.

 

[peng]

Brilliant example for test conditions coming from one of our own:

View attachment 315917

(more as thread continuous)

And many more here same as @VintageFlanker .
Stating the test conditions is paramount.As an example take my own measurements few posts above.
I could narrow the view to 20-20Khz,scrap the screenshot and only show the graph and present it either as 105.7db SINAD, 102.7db SINAD or 101.5db SINAD,in all three the visuals would be the same.

I don't question Hypex's measurements,not at all (although some hard-core objectivists would as lacking info).

(and no,I would not test other peoples stuff,I'm too cannibal and newbie for this,that takes decades of experience)

Edit:unitless numbers corrected

Different reviewers use different set of tests and measurements. I think it is time some of use propose some sort of a standard that includes only tests that are of practical value. That would mean putting emphasis/priority on, but not limited to the following:

a) Frequency response 20-20,000 Hz, at output level 0.1 V, 2.83 V, 50 V. Or sweep, in practical to do.
b) THD+N (SINAD) from 20-20,000 Hz, at output level as above, can use power instead of V, such as from 0.1 W to rated output.
c) SNR/DR at 0.1 V to 50 V, or again if W is used, from 0.1 W to rated output.
d) Power cube test, not so important for class D amp, I guess, and I prefer using a simulated load with 45 deg inductive and 45 deg capacitive, just to save time. If the amp does well at 45 deg., that's good enough.
e) "Continuous" power output, for 20 Hz, 60 Hz, 120 Hz, 600 Hz, 1000 Hz, duration: 1 minute, forget the sweep test, just do it for 0.1 W to rated output at steps of 20 W.
f) As e) but the regular sweep that Amir has been doing.
g) Burst power output, 20-20,000 Hz, , for duration, I think the 20 ms on, 480 ms off standard is not going to satisfy audiophiles or many ASR members so I would suggest going with 5s on, 1 s off for maybe 1 minute.
h) Cross talk
i) IMD
j) Maximum voltage at <-60 THD into 4 ohm simulated load with phase angle 45 degrees. Maximum current not needed because it can be calculated.

Many of the above are of course already done in Amir's, though not always consistent.

To be continued..

 

[Holdt]

Different reviewers use different set of tests and measurements. I think it is time some of use propose some sort of a standard that includes only tests that are of practical value. That would mean putting emphasis/priority on, but not limited to the following:

a) Frequency response 20-20,000 Hz, at output level 0.1 V, 2.83 V, 50 V. Or sweep, in practical to do.
b) THD+N (SINAD) from 20-20,000 Hz, at output level as above, can use power instead of V, such as from 0.1 W to rated output.
c) SNR/DR at 0.1 V to 50 V, or again if W is used, from 0.1 W to rated output.
d) Power cube test, not so important for class D amp, I guess, and I prefer using a simulated load with 45 deg inductive and 45 deg capacitive, just to save time. If the amp does well at 45 deg., that's good enough.
e) "Continuous" power output, for 20 Hz, 60 Hz, 120 Hz, 600 Hz, 1000 Hz, duration: 1 minute, forget the sweep test, just do it for 0.1 W to rated output at steps of 20 W.
f) As e) but the regular sweep that Amir has been doing.
g) Burst power output, 20-20,000 Hz, , for duration, I think the 20 ms on, 480 ms off standard is not going to satisfy audiophiles or many ASR members so I would suggest going with 5s on, 1 s off for maybe 1 minute.
h) Cross talk
i) IMD
j) Maximum voltage at <-60 THD into 4 ohm simulated load with phase angle 45 degrees. Maximum current not needed because it can be calculated.

Many of the above are of course already done in Amir's, though not always consistent.

To be continued..

Fully on board with your proposal. The points are, well, exactly on point. IR to point g) This would be akin to bursts, AKA transient capability so I wouldn't stretch it as far as 5 sec.

 

[Holdt]

At that high level I used the following album, didn't pay attention to which track, but a few tracks anyway.

View attachment 315922

So am I right in assuming not much energy in the bass area with this kind of music? I think this cabaret jazz genre is kind of easy on the amplifier and speakers in any case. It doesn't surprise me you could achieve high SPL without much power.

Try different genres like rock or EDM, orchestral, to explore where your systems' limits are.

 

[burkm]

The US's OSHA exposure guidelines (PEL = permissible exposure limits), straight from the web to you!
source (a third party): https://www.jespear.com/preventing-noise-induced-hearing-loss/
OSHA regulations: https://www.osha.gov/noise/standards

... and don't forget that the effect is cumulative

A very strong emphasis should be placed on "cumulative", because many people don't think about the follow ups like hearing damage. which then stays *forever*...

 

[peng]

So am I right in assuming not much energy in the bass area with this kind of music? I think this cabaret jazz genre is kind of easy on the amplifier and speakers in any case. It doesn't surprise me you could achieve high SPL without much power.

Try different genres like rock or EDM, orchestral, to explore where your systems' limits are.

I can't stand rock or EDM, I couldn't even stand Neil Diamond Hot August Night anymore let alone the modern rock stuff, now that I am old. Can try orchestral, I doubt it would make a difference though especially classical ones, average and max spl will be roughly in the same ball park as the likes of Particia Barber, Diano Krall kind of Jazz, that the amp can handle. Still, will try those, but need to find the "alone" moment lol.. So, stay tune.

By the way, I would suggest for rock or EDM, people should just use pro amps and not worry about SINAD, 60 dB or even 50 dB SINAD will be fine for that kind of music that are heavily amplified by pro amps anyway to begin with so assuming the best recording/mastering, pro amps will be good for those.

The amp that I can compared to now is the buckeye NC502MP, that I have already compared to my Parasound A21 and Bryston 4 B SST before already, so I have my reference, to some extent.

 

[Holdt]

I can't stand rock or EDM, I couldn't even stand Neil Diamond Hot August Night anymore let alone the modern rock stuff, now that I am old. Can try orchestral, I doubt it would make a difference though especially classical ones, average and max spl will be roughly in the same ball park as the likes of Particia Barber, Diano Krall kind of Jazz, that the amp can handle. Still, will try those, but need to find the "alone" moment lol.. So, stay tune.

By the way, I would suggest for rock or EDM, people should just use pro amps and not worry about SINAD, 60 dB or even 50 dB SINAD will be fine for that kind of music that are heavily amplified by pro amps anyway to begin with so assuming the best recording/mastering, pro amps will be good for those.

The amp that I can compared to now is the buckeye NC502MP, that I have already compared to my Parasound A21 and Bryston 4 B SST before already, so I have my reference, to some extent.

That's okay. The reason I asked for the music used was that I suspected it was the low energy kind. And that's fine. It's just a necessity to have that parameter when assessing needed amplifier power. Not everyones taste and thus need is the same. I'm sure you'll get the same result if you choose classical music and you don't have to listen to rock or EDM for me at all.

I don't share your view that EDM or (certain) rock music don't benefit from low noise and distortion. Why do you have that opinion? EDM can be just as high fidelity as classical music. Maybe some rock music with distorted guitars and such the noise and distortion may not be as audible but not all rock music is loud and distorted.

 

[peng]

Why do you have that opinion? EDM can be just as high fidelity as classical music. Maybe some rock music with distorted guitars and such the noise and distortion may not be as audible but not all rock music is loud and distorted.

If you can give me examples of those high quality rock recordings I would like to listen to them. My opinion was based on my limited live concert experience, rock or not, if they are highly amplified, I won't enjoy listening to them. I suppose studio recorded rock music could be of high quality.

 

[peng]

That's okay. The reason I asked for the music used was that I suspected it was the low energy kind. And that's fine. It's just a necessity to have that parameter when assessing needed amplifier power. Not everyones taste and thus need is the same. I'm sure you'll get the same result if you choose classical music and you don't have to listen to rock or EDM for me at all.

I listened to the 1812 overture, just the CD quality one as I was too lazy to do anything this morning after running a bunch of DL calibration runs, so I just listen to Qobuz and the 1812 overture they have if of CD quality. I am very familiar with that track, so I am quite sure the bass drum sounded real. AB compared briefly with my buckeye NC502MP, and just by level matching, sighted, I would not say one sounded better than the other.

Also listened to Hotel California, at V3's volume at 0, was getting 96 dB peak from 9 ft, must have gone higher but it was hard to catch the highest peak moment. Ran a few REW sweeps, will post later.

 

[peng]

Here's one of the REW sweeps I took today, will post on the V3 thread too:

Mic was 106 inches from the speakers, the spl is for left and right channels, a few dB lower for one speaker obviously. I had to wear earplugs to do it, it was scarely loud, but I know the relative big R900 can handle the current for such sweep. Basically, from 2 to 2.5 meter, one could push the V3 to get close to reference level for 8 ohm nominal with dip to 3.7 ohm such as the R900, and sensitivity of 90 dB/2.83 V/m. It is best to keep the input voltage no lower than 2 V, I guess 1 to 1.2 V maximum would be optimal as the graph below shows. It can go much higher, but THD is going to rise, based on comparing to the NC502MP. Response is great down to 30 Hz, below that it is hard to tell whether the speakers were the issue, I believe both were.

 

[restorer-john]

Different reviewers use different set of tests and measurements. I think it is time some of use propose some sort of a standard that includes only tests that are of practical value. That would mean putting emphasis/priority on, but not limited to the following:

a) Frequency response 20-20,000 Hz, at output level 0.1 V, 2.83 V, 50 V. Or sweep, in practical to do.
b) THD+N (SINAD) from 20-20,000 Hz, at output level as above, can use power instead of V, such as from 0.1 W to rated output.
c) SNR/DR at 0.1 V to 50 V, or again if W is used, from 0.1 W to rated output.
d) Power cube test, not so important for class D amp, I guess, and I prefer using a simulated load with 45 deg inductive and 45 deg capacitive, just to save time. If the amp does well at 45 deg., that's good enough.
e) "Continuous" power output, for 20 Hz, 60 Hz, 120 Hz, 600 Hz, 1000 Hz, duration: 1 minute, forget the sweep test, just do it for 0.1 W to rated output at steps of 20 W.
f) As e) but the regular sweep that Amir has been doing.
g) Burst power output, 20-20,000 Hz, , for duration, I think the 20 ms on, 480 ms off standard is not going to satisfy audiophiles or many ASR members so I would suggest going with 5s on, 1 s off for maybe 1 minute.
h) Cross talk
i) IMD
j) Maximum voltage at <-60 THD into 4 ohm simulated load with phase angle 45 degrees. Maximum current not needed because it can be calculated.

Many of the above are of course already done in Amir's, though not always consistent.

To be continued..

Frequency response at 1W or 5W doesn't make any difference. I prefer 1W but whatever. You do realize 50V is 625W@4R? That will rule out pretty much every amp except a few.

If you are going to load the amplifier with its rated impedance and then wind up the output level to do a FR test, you are simply straying into the area of the power bandwidth test which looks for the -3dB points from the reference level (usually 1kHz). So, nothing to be gained there if the PBW test is performed vs Freq.

Continuous power has to take into account the PBW or FR deviations or else the input level will be all over the shop attempting to obtain spot frequency max power outputs.

The CEA/EIA burst is fine. Your proposal of 5s on 1s off for 1 minute is not a burst test. That's a continuous test with an 83% duty cycle. Nothing to be gained there as the analyzer won't be able to unlock and lock reliably in that 1 second gap.

Crosstalk is already done.

IMD is already covered in 19/20kHz or 250Hz/8kHz or can be gleaned from the multitone with reasonable confidence.

 

[Rick Sykora]

Different reviewers use different set of tests and measurements. I think it is time some of use propose some sort of a standard that includes only tests that are of practical value. That would mean putting emphasis/priority on, but not limited to the following:

a) Frequency response 20-20,000 Hz, at output level 0.1 V, 2.83 V, 50 V. Or sweep, in practical to do.
b) THD+N (SINAD) from 20-20,000 Hz, at output level as above, can use power instead of V, such as from 0.1 W to rated output.
c) SNR/DR at 0.1 V to 50 V, or again if W is used, from 0.1 W to rated output.
d) Power cube test, not so important for class D amp, I guess, and I prefer using a simulated load with 45 deg inductive and 45 deg capacitive, just to save time. If the amp does well at 45 deg., that's good enough.
e) "Continuous" power output, for 20 Hz, 60 Hz, 120 Hz, 600 Hz, 1000 Hz, duration: 1 minute, forget the sweep test, just do it for 0.1 W to rated output at steps of 20 W.
f) As e) but the regular sweep that Amir has been doing.
g) Burst power output, 20-20,000 Hz, , for duration, I think the 20 ms on, 480 ms off standard is not going to satisfy audiophiles or many ASR members so I would suggest going with 5s on, 1 s off for maybe 1 minute.
h) Cross talk
i) IMD
j) Maximum voltage at <-60 THD into 4 ohm simulated load with phase angle 45 degrees. Maximum current not needed because it can be calculated.

Many of the above are of course already done in Amir's, though not always consistent.

To be continued..

Could this be continued on another (non-review) thread please?

At least one exists already...

Amplifier measurements may require improvement

Which kind of music did you use? -Which tracks? At that high level I used the following album, didn't pay attention to which track, but a few tracks anyway.

www.audiosciencereview.com

 

[peng]

Frequency response at 1W or 5W doesn't make any difference. I prefer 1W but whatever. You do realize 50V is 625W@4R? That will rule out pretty much every amp except a few.

I just pick a number as example, though 50 V seems reasonable because you need that voltage for some applications (higher nominal, peak impedance, larger room, distance, listening loud etc.), but I agree, like you said, could rule out most but a few such situations.

If you are going to load the amplifier with its rated impedance and then wind up the output level to do a FR test, you are simply straying into the area of the power bandwidth test which looks for the -3dB points from the reference level (usually 1kHz). So, nothing to be gained there if the PBW test is performed vs Freq.

Sure, that's why I am looking for inputs, so that can be substituted by saying "power bandwidth test.... -3 dB points.....etc.

Continuous power has to take into account the PBW or FR deviations or else the input level will be all over the shop attempting to obtain spot frequency max power outputs.

okay

The CEA/EIA burst is fine. Your proposal of 5s on 1s off for 1 minute is not a burst test. That's a continuous test with an 83% duty cycle. Nothing to be gained there as the analyzer won't be able to unlock and lock reliably in that 1 second gap.

I am familiar with the CEA thing, read the standard but in my opinion, it's bs at the time promoted by the likes of NAD, Yamaha who wanted to convince people that burst output is important for real world music listening. While that is true in many cases, but we all know it depends on the type of music, in many cases, including classical music, 20 ms is just not long enough. 5s I proposed, may be too long, but it is a starting point, it needs further study to find out what duration may be more appropriate, in terms of being more useful/meaningful.

Crosstalk is already done.

IMD is already covered in 19/20kHz or 250Hz/8kHz or can be gleaned from the multitone with reasonable confidence.

As I said, many of what I listed were already included in Amir's typical battery of tests, just that he has not been consistent, in some tests he did certain things while in others he did not. Some obviously were for good reason, but sometimes it might have been due to time constraints at the time, or whatever.

Thank you for responding. I have no idea what this might lead to, but I really do hope one day there will be a more formal and widely accepted list of home audio devices bench test guide, if not standard that cover the specific tests and methodologies.

 

[Rick Sykora]

That would be great, but don't we need our moderator to move the posts over?

Yes, but they are usually helpful. Will request assistance. Thanks!

 

[AdamG247]

Yes, but they are usually helpful. Will request assistance. Thanks!

Posts relocated as requested. Many thanks to @Rick Sykora for helping us keep our Review Threads on topic and working with the Moderator Staff to relocate off topic posts rather than them being deleted. This works fairly well for everyone involved and is a method we should use when we can.

Thanks for the help Rick!