• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Need Power amplifier Recomendations 1000.00 or less 500 watt minimum @4ohm.

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,406
The argument is that the measurements don't show the whole picture, because class d amps are usually measured with a filter which filters out the noise in the ultra-high frequencies, which could make an difference, by modulating the audible part of the high frequencies. Benchmark makes this claim for example: https://benchmarkmedia.com/blogs/ap...ory-use-of-the-benchmark-ahb2-power-amplifier

For the record, I don't have any skin in this game! Have heard many setups with class d amps I've really liked.

Hmmm I hadn't thought of this. It's a legitimate concern in theory at least. Not sure it's of any concern in practice.

Does anyone know of any experimental evidence of audible IM resulting from this within the audio band?

And does anyone with more technical expertise than me in this area have an opinion on this question?
 

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,721
Likes
1,939
Location
Oslo, Norway
Does anyone know of any experimental evidence of audible IM resulting from this within the audio band?

For me, the closest thing I know of in this regard is things I have read in a Swedish forum from guys who are involved with the Swedish Audio Society (Ljudtekniska sällskapet). It's an enthusiast club of audiophiles of an objectivist bent who have been testing audio gear for transparency for many years, using rather elaborate test methods. They publish an excellent magazine, which is one of the few audio publications I bother reading, except for ASR and a couple of German audio magazines.

Lately they've mostly measured and listened to dacs. Results are pretty consistent with Amir's findings here. Quite a lot of the dacs go un-detected in their tests, including some rather cheap ones. But they have also been able to detect subtle audible colorations in some of the dacs they have tested, which suprised me at first. Some of the guys have stated online that they have done basic blind listening tests on "well-regarded class d amps", without specifying which, and that they have always detected some coloration, and have therefore not found it worthwhile to do further testing.

But this is very anecdotal of course, and until they publish any full tests in their magazine I'm not sure how much I trust those reports. My impression is that audio objectivists all over the world have jumped on the class d train by now, and that it's mostly the Swedish objectivists and a few engineers like Benchmark who claim that class AB or class A still remains superior.
 

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,406
For me, the closest thing I know of in this regard is things I have read in a Swedish forum from guys who are involved with the Swedish Audio Society (Ljudtekniska sällskapet). It's an enthusiast club of audiophiles of an objectivist bent who have been testing audio gear for transparency for many years, using rather elaborate test methods. They publish an excellent magazine, which is one of the few audio publications I bother reading, except for ASR and a couple of German audio magazines.

Lately they've mostly measured and listened to dacs. Results are pretty consistent with Amir's findings here. Quite a lot of the dacs go un-detected in their tests, including some rather cheap ones. But they have also been able to detect subtle audible colorations in some of the dacs they have tested, which suprised me at first. Some of the guys have stated online that they have done basic blind listening tests on "well-regarded class d amps", without specifying which, and that they have always detected some coloration, and have therefore not found it worthwhile to do further testing.

But this is very anecdotal of course, and until they publish any full tests in their magazine I'm not sure how much I trust those reports. My impression is that audio objectivists all over the world have jumped on the class d train by now, and that it's mostly the Swedish objectivists and a few engineers like Benchmark who claim that class AB or class A still remains superior.

Nice summary :) I guess a big question is which "well-regarded class D amps" they are talking about.. Is their magazine available online or only in print form? And which German mags/sites do you recommend? I live here but have never looked into it much.
 

oivavoi

Major Contributor
Forum Donor
Joined
Jan 12, 2017
Messages
1,721
Likes
1,939
Location
Oslo, Norway
Nice summary :) I guess a big question is which "well-regarded class D amps" they are talking about.. Is their magazine available online or only in print form? And which German mags/sites do you recommend? I live here but have never looked into it much.

Yeah, I don't know which amps they meant either - only that it wasn't cheap china stuff. No, the magazine is only in print, and only in Swedish: http://www.lts.a.se/lts/senaste-molt

German magazines: I really like Fidelity. German version of Stereophile, kind of, just less fluff (though no escaping it completely). Anselm Goertz often does measurements for them, which is cool. Sound & Recording is also excellent IMO, but geared more towards pro sound and studio etc.
 
Last edited:

restorer-john

Grand Contributor
Joined
Mar 1, 2018
Messages
12,706
Likes
38,863
Location
Gold Coast, Queensland, Australia
My issue with Class D and a bunch of D/As is the HF rubbish they produce outside the audible bands as well. It's simply not good design for an audio bandwidth device acting as a low power transmitter that obliterates AM radio, some DTV reception and also affects FM radio signal quality.

Have you ever tried to listen to Shortwave with a Class D amp on? It's not pretty.

The AES 17 S/N testing conveniently stops at 20KHz and IMO, particularly with D/As, they should be measuring to at least 1/2fs and up to the sampling rate.

Basically, the S/N ratio has been gamed over and over, firstly by muting zero level, then by raising the output level to silly numbers and now by shifting all the noise up above the LPF in the AES filter specs.
 

stunta

Major Contributor
Forum Donor
Joined
Jan 1, 2018
Messages
1,156
Likes
1,403
Location
Boston, MA
obliterates AM radio, some DTV reception and also affects FM radio signal quality

If I don't care about any of this, the smaller footprint and lower idle power consumption are key benefits, aren't they? Given that, should one generally dismiss class D amps?
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,321
Location
Albany Western Australia
The argument is that the measurements don't show the whole picture, because class d amps are usually measured with a filter which filters out the noise in the ultra-high frequencies, which could make a difference, by modulating the audible part of the high frequencies. Benchmark makes this claim for example: https://benchmarkmedia.com/blogs/ap...ory-use-of-the-benchmark-ahb2-power-amplifier

For the record, I don't have any skin in this game! Have heard many setups with class d amps I've really liked.

"Class-D amplifier switching noise can create problems in a loudspeaker measurement system. More importantly, this ultrasonic noise may become audible when it is demodulated by the non-linearities that occur in all loudspeakers."

I know this is going to look biased :) but I struggle with the second part of the above assertion. Measurement system bit is correct, the signal can cause issues in the front end.

The switching frequency of say Hypex amps is IIRC 480kHz. A speaker driver is going to have to actually move to create IM based on this signal. Can someone show me a speaker that has any response at 480kHz?

The speaker is simply an electrical (inductor + R if directly connected and + C if with a crossover) and a mechanical filter to that signal, it has no effect.

I can post some scope plots of the output signal of a hypex later
 
Last edited:

Ron Texas

Master Contributor
Forum Donor
Joined
Jun 10, 2018
Messages
6,233
Likes
9,361
I doubt that even bats can hear 480 khz.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,321
Location
Albany Western Australia
Just through sheer cusriousity I have perfomed a calculation. Assuming a tweeter with resistance of 8ohms and inductance of 0.1mH and switching frequency of 480kHz impedance will be 2*pi*f*L + R
=310 ohms assuming no crossover in the way increasing it.
 
Last edited:

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,321
Location
Albany Western Australia
Just measured a Hypex 502 output connected to a B&W something or other.. Its actually 430kHz switching frequency 1.27 V pk-pk (0.45 V rms)

IMG_20180927_110525.jpg


scope_0.png


so using the tweeter example above, 0.72 mW into a driver that wont move at those frequencies anyway :)
With a crossover in the way it will be less, probably significantly so.

The inductance of a woofer is much higher, maybe 0.2 mH as an example, so maybe 0.45mW if directly connected and far less with a crossover.

I just cant see this as being any real world issue.
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,406
Just through sheer cusriousity I have perfomed a calculation. Assuming a tweeter with resistance of 8ohms and inductance of 0.1mH and switching frequency of 480kHz impedance will be 2*pi*f*L + R
=310 ohms assuming no crossover in the way increasing it.

That all makes sense, but wouldn't the far more likely source of IMD be in the amplifier itself? The speaker would need only be capable of reproducing the IM products, not the switching noise per se.
 

Wombat

Master Contributor
Joined
Nov 5, 2017
Messages
6,722
Likes
6,464
Location
Australia
Just through sheer cusriousity I have perfomed a calculation. Assuming a tweeter with resistance of 8ohms and inductance of 0.1mH and switching frequency of 480kHz impedance will be 2*pi*f*L + R
=310 ohms assuming no crossover in the way increasing it.

Z= R+jxL = sqrt[(2.pi.f.L)^2 + R^2} = 302ohm ;)
 
Last edited:

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,321
Location
Albany Western Australia
That all makes sense, but wouldn't the far more likely source of IMD be in the amplifier itself? The speaker would need only be capable of reproducing the IM products, not the switching noise per se.
Well I cant speak for other class D amplifiers but for the ones I am using, here are some IMD plots. Do you think its a problem more than in any other type of amp? :) Both shown at about 100watts output. Looks pretty good to me ;)

NC500

1538032362430.png


NC252
1538032470550.png
 
Last edited:

andreasmaaan

Master Contributor
Forum Donor
Joined
Jun 19, 2018
Messages
6,652
Likes
9,406
Well I cant speak for other class D amplifiers but for the ones I am using, here are some IMD plots. Do you think its a problem more than in any other type of amp? :) Both shown at about 100watts output.

NC500

View attachment 15935

NC252
View attachment 15936

Absolutely. The original claim though was that the use of a low-pass filter for the purposes of testing results in the in-band IMD caused by out-of-band noise not showing up in the standard tests.

I think it's a major long shot personally, although would be interested to know if anyone had studied the effects (if any) of out-of-band noise on in-band IMD, if only to rule it out.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,321
Location
Albany Western Australia
Absolutely. The original claim though was that the use of a low-pass filter for the purposes of testing results in the in-band IMD caused by out-of-band noise not showing up in the standard tests.

I think it's a major long shot personally, although would be interested to know if anyone had studied the effects (if any) of out-of-band noise on in-band IMD, if only to rule it out.
Im afraid that claim cant be correct. The filter only filters out of band signals. Intermodulation created in the amplifier circuit in band would not be filtered and show up.

The filter is there to stop problems (like IM) with the input stage of the measurement system, which would be erroneous.
 
Last edited:
Top Bottom